I've been reading three things this morning:
Roger Farmer, showing that lagged stock prices plus lagged unemployment have really good out-of-sample forecasting properties for unemployment.
Paul Krugman, saying:
"As I read it, price and wage setting — as opposed to bond market behavior — are mainly backward-looking, responding to inflation experience rather than attempts to read the central bank’s mind."
I read it roughly the same way as Paul.
There's a lot of inertia in nominal wages and prices. (I mean "goods" prices, like the CPI and GDP deflator, which I think is what Paul means too.) It takes something quite big and noticeable to happen, like a currency reform, or abandoning the gold standard, to get wage and price setters to coordinate quickly on a new equilibrium; otherwise it takes a longer time, while each one watches to see what the others are doing. Saying wages and prices are "backward-looking" is a rough and ready way to describe that inertia and inability of wage and price setters to coordinate quickly on a new equilibrium.
But asset prices (I mean things like stock and bond prices) can and do move very quickly, and are forward-looking. They respond quickly to any clues about expected future policy, including monetary policy.
So some prices (wages and goods prices) are "backward-looking"; and other prices (asset prices) are "forward-looking".
That's prices. What about quantities? What about real output (real income), employment, and unemployment? Are those quantities backward-looking or forward-looking?
If Roger's empirical work is right (regardless of whether he is right about the exact causal mechanism), and stock prices help forecast unemployment, that does suggest (it does not prove) that quantities are at least partly forward-looking too.
The whole idea behind Keynesian macroeconomics (and monetarist too) is that prices (I mean wages and goods prices) adjust slowly, and quantities will adjust instead. Quantities adjust more quickly than prices. Even though we know it's a simplification, we do short-run analysis where we assume prices and/or wages do not adjust at all, and all the adjustment is in quantities, which adjust immediately to the new fix-price equilibrium.
Is it too much of a stretch to say that while wage and (goods) price setting is mainly "backward-looking", quantity setting, like asset pricing, is much more "forward-looking"?
Suppose we get a new clue about what's on the central bank's mind, and what future monetary policy is likely to be. Asset prices respond immediately. Quantities respond a little more slowly (because it takes time to hire), both directly to the clue, and also indirectly to the clue via asset prices. And wages and goods prices eventually follow along in the rear.
The fact that wages and (goods) prices are mainly "backward-looking" does not mean that quantities are mainly "backward-looking" too. Quantities could be "forward-looking", even if wages and prices are "backward-looking".
If I'm right about this, a credible commitment by the central bank to a higher future NGDP level-path target could have a bigger and quicker effect on quantities than on prices. Which is what we (or rather, countries like the US and UK) want.
(I don't think Paul Krugman would necessarily disagree with this).