A lot of economics bloggers have been discussing the merits and flaws of the paradigm of aggregate supply and demand. Mark Thoma linked to a couple of the more recent examples.
I think that, viewed on its own terms, AS-AD is confusing (or confused). That is, mainstream economists do not know what to put on the vertical axis. If you put the price level on the vertical axis, you get a nice textbook model, but no connection to the real world, where economists talk in terms of inflation rather than the price level. On the other hand, if you put inflation on the vertical axis, that does not work very well, either, as I explain in this 8-minute video. (Comments, other than about my handwriting, are welcome.)
I developed PSST as an alternative to AS-AD. PSST does not appear to explain nearly as much as AS-AD. That may look like a feature in AS-AD, but on closer inspection it is a bug. The explanatory “power” of AS-AD is a delusion.
1. The AS-AD paradigm is invoked to explain changes in the combination of inflation and unemployment. If one goes up while the other goes down, we call this an aggregate demand shift. If both move in the same direction, we call this an aggregate supply shift. Thus, we can explain anything. But there is a lot of hand-waving involved. The stories we tell about AD and AS are always post hoc, just-so stories.
2. In the 1970s, we had a huge rise in the combination of inflation and unemployment. What caused this? The rise in the price of imported oil is often cited, but it cannot possibly account for the large acceleration in inflation. At most, it would cause a temporary increase in inflation, followed by a decrease. Many (most?) macroeconomists attribute the 1970s disaster to a “rise in inflation expectations.” Prices rose because they were expected to rise. Through habit and repetition, we have come to accept this story. But how intellectually satisfying is that?
3. Turning to events since 2008, the typical macroeconomist describes the rise in unemployment as the result of a “demand shock.” However, they do not all agree on what the shock was. Scott Sumner says that it was a contraction by the Fed. Others say that it was the wealth effect of a decline in house prices. Others say that it was a credit crunch. Others say that it was the effect of inflation dropping too close to zero. All of these are just-so stories.
From April of 2003 through April of 2008, the rate of growth of the CPI averaged 3.2 percent. From April of 2008 through April of 2013, it averaged 1.6 percent. If in 2007 you had asked macroeconomists to predict the consequences of a decline in the inflation rate of that magnitude, how many would have told you to expect unemployment to rise above 7 percent? None of them would have foreseen it. My guess is that many of the macroeconomists would have regarded a drop in inflation of 1.6 percentage points as close to a non-event for unemployment. (Note to Scott Sumner: yes, the decline in nominal GDP growth was much bigger than the decline in inflation. But that only restates the mystery–it doesn’t solve it.)
The terms “aggregate demand” and “aggregate supply” are highly loaded. That is, they lead economists to imagine something analogous to supply and demand in microeconomics. But the analogy is mostly misleading, and economists who invoke AS and AD are the ones who are most misled.
Amen.
You could have gone on and on, but a couple twists of the knife into AD-AS is good work for one day.
Are you familiar with Cowen and Tabarrok’s AD/AS model in their textbook? They use inflation and GDP growth, and the AD curve is made up of combinations of inflation and GDP growth for a given nominal GDP growth.
PSST to me seems to map more closely to a nonlinear macroeconomics where there are multiple equilibrium points with multiple levels of stability against perturbation. Inflation, by definition, produces noise in the system. Conversely, deflation, by definition, produces noise in the system. Fixed predictable prices, by definition, have the lowest randomness.
With an aging population, increased centralized governance, and a more specialized supply chain, does it become more difficult to solve the hill climbing problem? Do we need a historically high rate of inflation in order to produce enough randomness to break local minima in PSST and transition to a more productive global minima. Lots of good mathematics on this issue which I never see mentioned in these sorts of discussions…
I like your PSST paper, thanks for posting it.
You write in the paper that “the phenomenon of financial euphoria followed by financial crisis is important to address in either AS-AD or PSST.” IMO, AS-AD won’t get there because it’s fundamentally incompatible with the “Kindlebergian” view. I think of the broad economy as a never-ending sequence of virtuous and vicious loops involving credit and asset price cycles, the housing cycle and what I call the core business cycle. The loops don’t obey general equilibrium “stop signs.” They continue to exhaustion (or until a shock of some type) and then reverse like a pendulum. On the upswing, stability breeds instability. On the downswing, vicious circles breed creative destruction that eventually turns the cycle around again. (Or maybe I should say vicious circles lead to changing PSST after reading your paper.) In other words, there’s no such thing as an equilibrium “stopping point” where Kindlebergian forces die out.
I mention this because it may be another point of differentiation between PSST and AS-AD in the box at the start of the paper. It seems to me that it may be easier to reconcile PSST with financial cycles because 1) you’re not proposing mathematically-derived equilibrium solutions, and 2) you’re recognizing that pretty much everything in the economy – including job definitions, structural unemployment, etc. – is a moving part.
I’m a little confused by the end of the post. A large sustained decline in NGDP sounds like recession, which sounds like unemployment, to me.
The problems of the AS-AD model are resolved in the AS-ED model. ED = Effective demand.
The AS-ED model has new equations, new insights into spare capacity, LRAS curve, Phillips curve, NAIRU and more… I won’t spend a lot of time writing about it. Just let you check it out…
http://effectivedemand.typepad.com/ed/as-ed-model/
Don’t bad-mouth the AS-AD model, until you are able to bad-mouth its improved version.