Decades ago, George Selgin used the term ‘productivity norm’ to describe the intuition behind NGDP targeting (although his proposal was slightly different than simple NGDP targeting.) The basic idea was that the price level should move inversely to productivity shocks, keeping the dollar flow of spending (per capita) relatively stable.
I’d like to illustrate the intuition with a simple (and admittedly unrealistic) example, and then you tell me whether it has any relevance for today.
Consider the restaurant industry. Suppose a sudden shock cuts restaurant productivity in half. What might that shock be? Suppose people suddenly become very anti-social and prefer to sit at least 6 feet away from other restaurant patrons. Restaurant capacity falls sharply. What is the new long run equilibrium?
Restaurants are a competitive industry, so in the long run we can expect the relative price of restaurant meals to double. Because of 50% lower productivity, it’s now more costly to produce meals. And in competitive industries, price equals marginal cost in the long run.
What about the nominal price of meals in the short run? That depends on monetary policy. So what sort of nominal price change might lead to the least short run disruption?
In my view, there’s a lot to be said for having the nominal price of restaurant meals double at the same time that the long run relative price is doubling. After all, we are assuming that because of the public’s anti-social behavior, the number of patrons per restaurant will fall in half. If the nominal price doubles, then restaurant revenue will remain roughly stable in nominal terms.
In that case, the average citizen will spend exactly the same amount on restaurant meals as before. They’d buy half as many meals, at twice the price. That means their non-restaurant spending would not need to change.
In addition, the average restaurant owner would earn the same nominal income, and thus have the same ability to repay their nominal debts. The average restaurant would not go bankrupt, and thus they would not have to shut down.
Of course things might not work out so smoothly, it depends on the elasticity of demand for restaurant meals. But the basic idea is that if you keep the flow of nominal spending stable, or along a stable growth path, then productivity shocks do not by themselves lead to debt crises. In aggregate, firms earn just as much revenue as before the productivity shock. So while some individual firms will have more trouble repaying debts (think cruise ships), the average firm will not have more difficulty repaying nominal debts.
In contrast, if the central bank responds to the negative productivity shock by allowing NGDP to fall sharply, you will have a severe debt crisis. Most debts are nominal.
Former Fed governor Jeremy Stein once argued that financial excesses could be better addressed with monetary policy than regulation, because “monetary policy gets in all the cracks.” Tight money discourages everyone from borrowing, whereas regulation will miss many excesses.
Stein was wrong; tight money cannot address financial excesses without tanking the economy, as the Fed learned in 1929-30. But his money “gets in all the cracks” argument actually does apply to monetary policy during a productivity crisis. Right now, the government seems to have given up on monetary stimulus, and instead plans to bail out a bunch of individual firms. But that will never work; they cannot possibly address a problem this massive on a case-by-case basis. Many smaller firms will go bankrupt. And if they were all bailed out, the fiscal cost would put an enormous burden on future taxpayers.
In contrast, monetary stimulus is (as a first approximation) costless. There is no burden on future taxpayers. And it gets in all of the cracks. Monetary stimulus is like a rising tide that lifts all boats. Some boats (i.e. firms) won’t be lifted far enough to avoid bankruptcy, but with stable NGDP growth the average firm will be just as able to repay nominal debts as before the productivity shock.
Right now, NGDP level targeting has two huge advantages over the current bailout policies being discussed:
1. It’s way cheaper.
2. It’s way more effective.
So why not adopt NGDPLT? It’s probably too late to prevent a big drop in Q2 NGDP, but there’s still time to boost longer-term prospects.
PS. I know that my example is oversimplified, as I ignored multiple costs such as labor, etc. I was trying to get at the intuition with a simple example, as the argument for NGDP targeting still holds if we make restaurant cost functions more complex.
READER COMMENTS
Michael Sandifer
Mar 19 2020 at 4:55pm
I’m curious as to the intuition behind not wanting to target NGDP during the current year in our current crisis. Yes, most of the monetary stimulus needed is for future years, but why not level target NGDP over shorter periods, in some cases?
I understand that potential RGDP has recently fallen due to the nature of the crisis, such that it’s easier to overshoot on monetary stimulus in the short-run. But, it’s also fair to expect that potential RGDP will only be down in the short-run, so I don’t see a reason to favor temporary overheating, wage adjustment, and then a fall in employment and output. This isn’t the 1970s.
Am I wrong not to be concerned about pushing the economy a bit over capacity, if even possible right now, and instead focus on disallowing a rise in wages/NGDP?
Scott Sumner
Mar 19 2020 at 5:51pm
I think that would be hard to do. Suppose you peg the price of 12 month forward NGDP futures contracts. NGDP would probably still fall in Q2. Doing enough to stabilize Q2 might destabilize longer term NGDP.
Michael Sandifer
Mar 19 2020 at 7:13pm
What about in a scenario in which the Fed level targets NGDP by making daily adjustments in response to seasonally-adjusted NGDP-related final sales data conveyed electronically? This would be as opposed to targeting NGDP in future years
Scott Sumner
Mar 20 2020 at 1:26pm
That might work, but I don’t think we know much about that sort of policy. I’m trying to stick with more mainstream approaches for now.
Thaomas
Mar 21 2020 at 1:33pm
So make a guess of how much supply constraints will reduce real GDP and announce a 1 year PL target (using TIPS as adjusted for CPI/PCE) to achieve estimated NGDP growth of 4-5%. As estimates of the effects of supply constraints change, the PL target could change too.
Michael Sandifer
Mar 19 2020 at 4:57pm
To correct one sentence above, “so I don’t see a reason to FEAR temporary overheating, wage adjustment, and then a fall in employment and output.
Thaomas
Mar 19 2020 at 5:42pm
NGDP targeting seems to best carry out the central bank’s obligation to enforce Says Law. If supply goes down in one industry relative price of its output should go up. But with sticky prices higher prices of most everything will be needed to keep some prices being constrained at the existing lower bound of non-negative change.
What this means for monetary policy right now is that the Fed should announce a higher inflation target that is compatible with what it calculates would keep NGDP growing at the current 4-5& rate. When the crisis has passed they should create that NGDP futures market and intervene to keep IT growing at 4-5%.
Matthias Görgens
Mar 20 2020 at 2:51am
The Fed could just switch to level targeting right away with their current transmission tools. They have already made some noises about average inflation targeting; and with a long enough period to average over, that’s pretty close to price level targeting.
A higher temporary inflation target would be better than nothing. But committing to an averaging or level target would be better and actually less adventurous.
NGDP level targeting is also possible. But a bigger jump politically.
The NGDP futures market would be nifty, but no need to wait for them. After all, the Fed doesn’t use TIPS spreads as an instrument at the moment either.
Benoit Essiambre
Mar 19 2020 at 9:04pm
I’ve been reading your blogs for years and I learned an incredible amount. Despite finding that market monetarists laid out macroeconomics better than anyone else, the one thing I couldn’t fully buy into was NGDP targeting. I find this odd since NGDP seems so central to the MaMo gospel. I always thought that even if NGDP targeting was optimal in some theoretical sense, there seemed to be practical hurdles making it a subpar approach. Something like PLT made more practical sense to me.
What got me closest to accepting NGDPLT was reflections along the lines of the scenario described in this post. I was imaging an asteroid hitting earth and wiping out 20% of production capacity and I thought to myself, I guess in a situation that drastic, NGDPLT might be more suitable because targeting prices would just not be enough for the credit markets to clear, for wages to adjust fast enough or for nominal income to be shared widely enough across a post apocalyptic economy. I did not think I would witness anything close to it in my lifetime.
Comments are closed.