What is the optimal NGDP growth rate?
By Scott Sumner
I often get commenters asking my opinion on the optimal NGDP growth rate. This is a difficult question, which can be addressed from a number of perspectives.
Josh Hendrickson has a new policy brief at Mercatus, which advocates the “Friedman rule” for NGDP targeting. This would involve setting the growth rate for NGDP at a point where risk free short term interest rates would be close to zero and there would be no opportunity cost of holding currency. Because currency can be produced at near zero cost, economic efficiency suggests that nominal interest rates should be low enough that the public doesn’t waste resources trying to economize on the use of currency, which doesn’t pay interest.
I’d encourage people to look at Hendrickson’s policy brief, which is very clear and well written. For those who want a more technical analysis, Josh also has a Mercatus working paper.
I see two arguments for keeping NGDP growth low enough to insure near-zero nominal interest rates:
1. A low trend rate of NGDP growth minimizes the opportunity cost of holding currency, which (as mentioned above) is the rationale behind the Friedman rule.
2. A low rate of NGDP growth lowers the effective tax rate on capital income (interest, dividends, rents, capital gains, etc.) because in the US and most other countries the income tax applies to nominal capital income, not real capital income. (Ask me what happened when I sold my 2-family house in Newton MA, back in 2017.) Because capital income taxes are inefficient, faster NGDP growth discourages capital formation and slows economic growth.
I can see three arguments for slightly faster NGDP growth:
1. Faster NGDP growth reduces the risk of the zero bound problem with interest rates. I actually don’t think the zero bound is a problem, but as long as central banks are obsessed with interest rate targeting, the zero bound does make monetary policy a bit less effective.
2. Faster NGDP growth reduces the problem caused by money illusion, the reluctance of workers in declining industries to accept nominal wage cuts, even though they would be perfectly willing to accept the exact same real wage cut achieved via inflation.
3. Faster NGDP growth imposes an implicit tax on currency hoarding balances, the large stock of $100 bills that are held to avoid taxes.
I have advocated an NGDP growth rate of 5%/year, and more recently a rate of 4%. In practice, the target should probably be in per capita (or worker) terms (say 3.5% or 4.5%) but I’ve ignored that minor complication in this post.
My current advocacy of 4% is based on two factors:
1. I believe that a 4% growth rate does a decent job of balancing the costs and benefits discussed above. Not perfectly, but well enough that any welfare loses are probably small.
2. It’s easier to sell the idea if you are not simultaneously attacking current monetary policy in two directions. Thus a proposal for 2% or 6% NGDP growth would be seen as an implied criticism of both inflation targeting in general and the Fed’s specific 2% inflation target.
In contrast, given the Fed’s current view that trend RGDP growth is roughly 2%, proposing a 4% NGDP will correctly be seen as roughly consistent with the Fed’s 2% inflation target in the long run. Instead, the goal is to make inflation more countercyclical, which is of course consistent with the Fed’s dual mandate. Thus proponents of a 4% NGDP target don’t have to apologize for ignoring the Fed’s Congressional mandate, or for abandoning the Fed’s goal of 2% inflation, on average, in the long run.
But I’d also be thrilled if we went to a 3% or 5% NGDP target, as long as it was a level targeting regime. Instability of NGDP is a far bigger problem that getting the average slightly wrong.
PS. People should check out Josh Hendrickson’s excellent twitter account. I particularly enjoyed his tweets on Marco Rubio’s strange decision to put out a policy paper based on MMT ideas. In fairness, I don’t expect politicians to be particularly well versed in macro, just as I know little about bankruptcy law. But I do recommend they not put out papers on subjects in which they are not well versed. Thus in my view Congress was correct to give the Fed a broad mandate, and let the Fed fill in the details. Congress’s mistake was to not ask the Fed for a high degree of specificity as to how they interpret the law, so that it would be crystal clear when the Fed had failed to achieve its Congressional mandate (as during 2008-13). That’s the flaw that lies behind my various accountability proposals.
PPS. Off topic, but this FT story relates to my recent post on the dangers of libertarians allying themselves with rightwing authoritarian parties.