My colleague and co-blogger Frances Woolley doesn't do macroeconomics and monetary theory. It's not her cup of tea at all. But when I started to write a response to Steve Waldman's very good post on NGDP targeting, I found myself writing about Frances' also very good post on demographics and saving for retirement.
Steve: meet Frances. Macro/money: meet micro (or rather intertemporal general equilibrium theory, minus the usual math).
Suppose, for example, that there are large numbers of ageing baby boomers, and very few young people who will be working when the boomers all retire. The boomers all want to save for their old age, so they will be still able to eat when they have stopped producing goods themselves. And there just aren't enough profitable investment projects to match desired saving even at a zero real rate of interest. The boomers would want to store consumption goods for their old age, if they could. But as Frances says, if you bury a loaf of bread in the ground it probably won't be good to eat in 20 years.
Some newly-produced goods are storable at low cost, and that storage demand might be enough to match the supply of desired saving. The boomers store all the goods that can be stored, and then dig them up 20 years later and swap them for newly-produced bread. But maybe there just aren't enough goods that are cheaply storable and that the young will want for that to work.
Could fiscal policy solve the problem, and raise the natural rate of interest back up to zero? Well, yes and no. Yes, the government can borrow all the boomers' excess savings, raising the natural rate back up to zero. But does that really solve the problem? No, not really. Not unless the government has figured out a way to bury loaves of bread in the ground and dig it up so it tastes freshly baked 20 years later.
The only way the government can get bread into the mouths of the retired boomers is to tax the young so they are forced to bake bread, sell it to the boomers, hand the proceeds over to the government in taxes, which the government uses to buy back the bonds they sold to the boomers 20 years ago, so the boomers can use the proceeds from selling their bonds to buy bread baked by the young. Force the next generation to bake bread and give it to the retired boomers for free, in other words.
Either the government makes sufficiently large intergenerational transfers to push the natural rate back up to zero. Or else you have to let the natural rate stay negative, and let the actual rate of interest fall to the natural rate. Which one's more fair? You tell me. The second is tough on the boomers, who get a negative rate of return on their saving. The first is tough on their kids, who get to pay the high taxes.
As Frances says, "Quit whining and deal with it". You can't solve this problem; you just have to choose some mix of the two bads.
Monetary policy certainly can't solve this problem. Forgive the cliche, but you can't eat money, whether it's paper or gold. But monetary policy would have to adjust to this problem. If the natural rate of interest is minus 3%, you can't target 2% inflation. If you did, then the boomers would want to hold a lot of their savings in cash. And the central bank would have to sell them as much cash as they wanted to hoard to prevent a recession. And either invest the proceeds at a loss, which the next generation would have to cover, or else force the next generation to pay taxes to buy back the ageing boomers' cash to prevent high inflation when they spend it. Which is just an intergenerational fiscal transfer done by the central bank.
If the central bank could observe the natural rate of interest in real time, this would all be rather easy. Just make sure the inflation target is more positive than the natural rate is negative. If the natural rate is minus 3%, then target 4% inflation. So nominal interest rates stay at 1%. And if you see the natural rate fall lower, immediately raise the inflation target in response. But, of course, we don't observe the natural rate in real time.
So, what's the next best thing? Well, you could argue that expected real GDP growth is a fairly good rough proxy for the (real) natural rate of interest. Low expected real GDP growth will be associated with a low natural rate. And so targeting a level growth path of Nominal GDP would give you higher inflation during times when the natural rate and real GDP growth are both lower than normal. That's how I interpret Scott Sumner's response to Steve. You just need to set the target growth path of NGDP high enough to give the central bank a margin for error, in case the proxy isn't perfect. Or else, be prepared to have a central bank with a rather large balance sheet if you target too low a growth rate of NGDP and don't leave a large enough margin for error.