"But what concrete steps will the Fed actually take to raise Nominal GDP? Can anyone tell me that?"
I must have heard that question a hundred times over the last couple of years. And a dozen more times in the last day, ever since Paul Krugman endorsed the proposal to target NGDP. I always imagine it being asked in a gruff Yorkshire accent, by some middle-aged no-nonsense practical man of business with a background in mechanical engineering.
So this is written for the people of the concrete steppes.
First off, you aren't thinking about this right.
Sure, I've used mechanical metaphors for monetary policy in the past. But those metaphors only take you so far. Machines don't have expectations; people do. The actions that people take now depend very much on their expectations of what other people will do, and on their expectations of what the central bank will do.
You want me to tell you a story in which the central bank pulls a lever, and that lever causes another lever to move next, followed by another lever, then another, spelling out a causal chain from beginning to end, where the end is a higher level of NGDP.
But economics isn't like that. Because people aren't like that.
Sometimes the future causes the present, because people's expectations of the future affect what they do in the present.
Sometimes it's not even what will happen in the future that causes the present. It's people's expectations of what would happen in future if they behaved differently today that causes them to behave the way they do today. I switch to snow tires in the Fall because I expect I would have an accident in the Winter if I didn't. It's the threat of an accident that makes me put on snow tires. I don't actually expect to have an accident.
Next week, millions of Canadians will get up and go to work about one hour later than they did this week, if we measure time by the sun. "What concrete steps will the government take to get Canadians to do this? Can anybody tell me that?"
No, I can't tell you that. I don't have a clue what concrete steps, if any, the government will take. But I fully expect it will work. All the government does is announce that it wants us all to do this, and to put all our clocks back one hour. Maybe the government has the power to force government clocks back one hour, and force some government workers to start and leave work one hour later by the sun. But the rest of us just follow along, simply because we expect everyone else to follow along.
There will be two groups of people who will continue to get up at the same time by the sun. People who didn't get the memo (and most of them will follow along with a short lag). And people who live very isolated lives, whose timing doesn't depend on what everyone else is doing.
Driving on the right side of the road is an equilibrium. If you expect everyone else to drive right, you too will drive right. Driving on the left side of the road is also an equilibrium. If you announce that beginning Sunday everybody will switch to driving on the left, and if people believe you, or simply believe that other people believe you...everybody will switch to driving on the left. You don't actually have to pull any levers. All you need is credibility. Or people to believe you have credibility. Or believe that others believe you have credibility.
The US economy is currently in equilibrium. It's not a market-clearing equilibrium. It's not a very good equilibrium. But it is an equilibrium. If it wasn't an equilibrium, it would be somewhere else. But it isn't somewhere else, so it must be. Given what people expect other people to do, both now and in the future, each person is choosing to do what he is currently doing.
But this isn't the only possible equilibrium. I can imagine a better equilibrium, in which Nominal GDP is higher and growing faster, and expected to remain higher and growing faster. NGPG is higher and growing faster both because real GDP is higher and growing faster and because prices are higher and growing faster. It's a better equilibrium. And those of us who advocate E(NGDP) level-path targeting want the US economy to move to that better equilibrium.
What would the Fed be doing differently, in that other, better equilibrium? The Fed will be smaller than it is today, and the Fed's interest rate will be higher than it is today. Real interest rates will need to be higher, because consumption and investment demand will be higher, because consumers and investors will have higher expectations of future real income and real expenditure. Nominal interest rates will be higher because prices are expected to be growing faster. The Fed will be smaller, because people won't want to hold as much money, and banks won't want to hold as many reserves at the Fed, now the economy is growing and interest rates are higher.
So, all the Fed needs to do to get the economy to that new, better equilibrium is to pull the lever in the right direction, right? Raise interest rates, and reduce the money supply, right?
Of course not. If the Fed did that, without changing expectations, the result would be a a move even further away from the better equilibrium, as demand fell even further.
The Fed needs to change expectations. Get people to expect that NGDP will follow the higher path. That's what the "E" in "E(NGDP)" stands for.
"Right!" the people from the concrete steppes exclaim gruffly "and how exactly will the Fed do that?!"
1. The Fed clearly announces its target path for NGDP. That's by far the most important bit. Everything else is secondary. And if the Fed had credibility, that would be enough.
"Why should anyone believe the Fed can hit that path?"
2. The Fed makes a threat. On the first day the Fed will print $1 billion and use it to buy assets. On the second day the Fed will print $2 billion and use it to buy assets. On the third day the Fed will print $4 billion and use it to buy assets. And the Fed will keep on doubling the amount it prints and buys daily, forever and ever, until E(NGDP) rises to the target path. (And will go into reverse and sells assets if E(NGDP) rises above the target path).
"What assets will the Fed buy?"
3. The Fed puts on its best James Dean (oops, Marlon Brando, thanks Andy) voice and replies: "What have you got?"
There are two rooms at a party. The first room is nearly empty. The second room is nearly full. Because everyone wants to be where everyone else is. Then Chuck Norris enters the second room. He threatens to beat up 1 person at random in the first minute, 2 people in the second minute, 4 people in the third minute, and so on, until the room is empty. This is no longer an equilibrium.
A few people were nearly indifferent to being in the second room. So they leave even if the chance of them getting beaten up is tiny. That means there are fewer people left in the second room. This makes the second room slightly less attractive for those who want to be where everyone else is. And it slightly raises the probability of being beaten up by Chuck Norris. So more leave. Which repeats the process, so still more leave. And if you and I can see what's coming, so can the people in the room, who don't want to be the last to leave. There's a rush for the exits, and Chuck doesn't even have to lift a finger. OK, if someone didn't hear the threat, or doesn't recognise Chuck Norris, he might actually have to carry out his threat for a few minutes. But simply seeing all the others leave the room will be enough to induce most to leave the room very quickly.
Chuck Norris doesn't have to beat up everyone in the room. He just has to threaten to beat up as many as it takes to clear the room. The number of people he will actually beat up is a lot less than the number he threatens to beat up. If his threat is credible, and everyone hears it, he doesn't need to beat up anyone.
Eventually, if the Fed bought up every single asset in the economy, and swapped it for cash, NGDP would rise to the Fed's target path. Prices would rise without limit as the Fed bought up the last remaining assets because the sellers could name their price. And people would hire the unemployed to build factories which they could float on the stock and bond markets and sell to the Fed at any price they liked. Or sell to the people who had already sold all their assets to the Fed.
But there is no way it would ever get that far. That's like saying that Chuck Norris will eventually beat up everyone in the room. That's not an equilibrium.
Some people are just barely willing to hold cash in the current equilibrium. If they expect even the slightest rise in NGDP in even the distant future, they will get out of cash, and into real assets, or claims on real assets like commercial stocks and bonds. And this will increase the demand for goods today, either directly, or because firms find it easier to issue new stocks and bonds to finance investment. Which raises NGDP, and expected future NGDP, even if just a little. Which encourages additional people to exit cash too, and buy real assets and claims to real assets. Which raises NGDP and expected future NGDP still further. And so on. As soon as people figure out what's going on, and what's going to happen, expected NGDP rises to the target path. The Fed only has to carry out its threat until people catch on to what's happening. Then it has to reverse course and sell all the assets it bought, and then some more, to prevent the economy overshooting the new equilibrium.
Want me to make it more concrete still? OK. Here's the New Keynesian version:
Eventually, if it carries out its threat for long enough, NGDP will eventually rise. Some people figure this out. Maybe only a few. They expect either a rise in the future price level or a rise in future real income, or a bit of both. If they expect a rise in the future price level, that means lower real interest rates for given nominal interest rates. That encourages current consumption and investment demand. If they expect a rise in future real income, that also encourages current consumption demand and current investment demand. (People consume more today if they expect to have higher real income in future; firms invest more today if they think there will be bigger demand for the extra goods that will help them produce in future.) So current consumption and investment demand rises, which increases current NGDP.
The more slow-witted people and firms see that rise in current NGDP, and spend more on consumption and investment, and also revise upwards their expectations of future NGDP, which causes an additional rise in consumption and investment and NGDP.
More people begin to figure out that maybe the Fed's target is credible after all.
And so on.
Oh, and if you believe in the debt-deleveraging hypothesis, there's an additional channel. Higher expected future NGDP makes it easier to handle a given debt load, by reducing future expected debt/income ratios.
Update: I really want people to read a triad of very good posts by Bill Woolsey: his response to thoughts on NGDP targeting by Paul Krugman; Brad DeLong; and Steve Williamson. Bill lays out clearly some of the issues about how NGDP level path targeting works.