A new model of the macroeconomy
By Scott Sumner
As you will see, the title of this post is half joking and half serious. I’ll try to explain both halves, starting with the joking part.
My brand new model of the macroeconomy is called the NS/RO model, which stands for nominal spending/real output. The model shows how equilibrium output and the price level are determined. The key insight is that (positive) changes in nominal spending have a short run positive effect on real output, but no long run effect. The following graph illustrates the basic idea:
NS is the nominal spending curve. It is a rectangular hyperbola, where each point along a given line represents a given amount of nominal spending, aka NGDP. SRRO is the short run real output function. LRRO is the long run real output function, reflecting the natural rate of output (Yn.)
Those of you who are familiar with macro probably now see the joke. This isn’t actually a brand new macro model; it’s one version of the old AS/AD model that has been dressed up with new terminology. I say “one version”, because in the most popular version of AS/AD, the one used in most textbooks, the AD curve (NS in this case) is not a rectangular hyperbola. The graph presented here is the less frequently used (monetarist) version of AS/AD. Still, it is an AS/AD model.
Now for the half serious part of the title of the post. I’d argue that this should be viewed as a brand new model. Deirdre McCloskey has shown that there’s more to economics that mathematical equations and graphs. Much of what we do is persuasion, and that requires effective methods of communicating our ideas. Historical examples, stories, metaphors, and many other forms of rhetoric all play a role in making an argument persuasive. Even the terminology that we use matters. Consider these two equations:
MV = PY
M = kPY
To an unimaginative, mathematically oriented economist the equations might seem identical, as V = 1/k. But these two equations trigger different ideas in the reader’s mind. The first equation makes one think of money as a medium of exchange. How much money is there in the economy (M) and how fast is it spent (V)? The second equation makes one think of money as a store of value. How much money is there in the economy and how much money do people choose to hold as a share of their gross income (k)?
My “brand new model” of the macroeconomy is intended to dissuade people from thinking of this as a supply and demand model. When I used to teach macro, students had all kinds of trouble understanding AS/AD, partly because they assumed it was basically a supply and demand model. Thus if I asked them to show the effect of population growth, they might shift the AD curve to the right. In fact, population growth shifts the AS curve to the right, causing deflation (as we saw during 1870-95).
Why were students confused? Probably because in microeconomics an increase in population really does shift demand to the right. More consumers out shopping, etc., etc. But aggregate demand isn’t really about demand at all; it’s about money. More specifically, it’s about the medium of account. In the late 1800s, gold was the medium of account, the thing in terms of which prices were measured. Population growth didn’t cause more gold to circulate.
[Yes, population growth might lead to new discoveries of gold, but I view that as a second order effect—the main impact is on supply. And yes, short run effects are more complex, perhaps influencing velocity. But for any given money supply, a doubling of the population will cause prices to fall by roughly 50% in the long run.]
By removing the terms ‘demand’ and ‘supply’ from the AS/AD model, I hope to shake students out of their complacency, to show them how truly strange this model actually is. It’s not a supply and demand model at all; it’s a model of how nominal shocks interact with sticky wages and prices to produce short run (but not long run) effects on real output.
PS. In an actual supply and demand model the long run supply curve is usually almost horizontal. In AS/AD it is vertical. Sorry, but AS/AD is not a supply and demand model.