Macro is basically about sticky wages and prices, and why nominal shocks have real effects. Macroeconomists also talk about other issues, but it’s the nominal shock/sticky price problem that requires a separate field. Otherwise macro is just applied price theory, i.e. classical economics.
Arnold Kling recently discussed the AS/AD model and suggested:
This aggregate supply mechanism assumes that wages stay fixed while prices move. This sticky wage hypothesis is at the center of the whole mechanism.
. . .
I want to particularly exempt Scott Sumner from this. He understands aggregate demand and supply exactly as I understand it.
I pretty much agree with the first part, although I’d add that sticky prices can also be an issue. As for the second two sentences, I hope for Arnold’s sake it’s not true, as I don’t understand AS/AD very well at all. And each time I read Nick Rowe I feel like I understand it a bit less well. (That’s a dig at myself, not Nick.) BTW, Arnold is clearly talking about long run equilibrium.
Here’s a Nick Rowe post entitled “What equilibrates AD and AS?”
If a binding minimum wage law sets W/P too high, there will be excess supply of labour. People can’t sell as much labour as they want to. If a binding pro-usury law sets r too high there will be excess demand for bonds. People can’t buy as many bonds as they want to. But if M/P is right, AS will still equal AD. The output market clears, even if the labour market and bond market don’t.
Arnold’s still right in a way. If M/P adjustment can be taken for granted, then if people want less labour and less output, or firms want more output and more labour, then W/P will adjust to coordinate their conflicting desires.
And those Keynesians are still right in a way. If M/P adjustment can be taken for granted, then if people want less consumption today and more consumption tomorrow, or firms want less investment today and more consumption today, then r will adjust to coordinate their conflicting desires.
And in the long run M/P adjustment can be taken for granted. And W/P adjusts if we decide we want to consume more leisure and buy fewer Ipads and BMWs. And r adjusts if we want more consumption tomorrow and less consumption today. But we can’t take M/P adjustment for granted in the short run.
After reading all of this, I’m not even sure what the phrase “equilibrate AD and AS” means. In the comment section Nick points out that the SRAS curve isn’t really even a supply curve. (He uses the sticky price Keynesian model, which assumes monopolistic competition.) I’m actually fine with that model, although for some reason people seem to think I believe in flexible prices, maybe because I talk about wage stickiness as the biggest problem. But if AS isn’t really a supply curve, but rather a sort of loci of equilibria for various settings of AD, then why even talk about the need to equilibrate AS and AD?
I’ll even go further, since prices are such a distraction, let’s entirely remove the price level from macroeconomics. While we are at it, let’s dump interest rates as well. BTW, by “macro” I mean business cycle theory, what I was discussing in the intro. You need interest rates to explain saving and investment, and you need prices to estimate real economic growth; although as we see from the comments to Tyler Cowen’s new book, nobody has a clue as to what “real economic growth” is supposed to mean. More stuff? More enjoyment? Your guess is as good as mine.
Since macro is about nominal shocks having real effects, we need a nominal variable that actually, you know . . . “shocks.” Inflation won’t do, as the Keynesians tell us that prices are fixed, and only move in response to an overheating real economy. So I nominate NGDP. Even the old Keynesians accept that NGDP can move immediately (or at any rate just as fast as RGDP) in response to factors such as monetary and fiscal policy, as well as financial crises.
So just do AS/AD with NGDP on the vertical axis. Everything is the same, except the AD curve is flat. People often assume this requires some sort of assumption about the quantity theory of money being true. No, 100% of AD shocks might be velocity. Nor does it assume the Fed targets NGDP. It would work equally well under a free market gold standard. It’s true that I often equate monetary policy with expected future NGDP, but that’s not required. You can use NGDP as AD and define monetary policy any arbitrary way you please. (To me the only definition that makes sense is easy or tight relative to what would be expected to hit the target.)
It’s tempting to make the model even simpler by replacing SRAS and LRAS with a single AS, which crosses the horizontal axis at the natural rate. No need for an AD line. Just put NGDP minus expected NGDP on the vertical axis. But I’ve given up on that idea; there is too much misunderstanding about what the term “expected NGDP” really means. (It means different things in different contexts.)
Obviously if there is no price level and inflation, there is no RGDP, so put hours worked on the horizontal axis.
The real wage is now W/NGDP, (per capita NGDP if you wish, but it doesn’t really matter.) This is actually a sort of “relative wage,” the share of national income going to an hour’s labor. The late 2008 crash raised that relative wage rate, and caused mass unemployment.
This model can explain why nominal shocks have real effects in the short run, but hours worked return to the natural rate in the long run. If you wish you can add the price level, the nominal or real interest rate, or the price of a pound of pistachio nuts, but none of them will add anything useful to the model.
So how about it Arnold, do you still want to claim you understand AS/AD the same way I do, or do you wish to remain a highly respected economist?