Imagine that you have a company pension plan that invested in a set of ultra-low cost stock and bond index funds.  Then someone suggested that the company should change the plan by hiring a manager to try to select stocks and bonds that would beat the market.  Each year that stock picker would be paid a commission equal to 1.5% of your retirement assets.  How would you feel about that decision?  I don’t know about you, but I’d be pretty unhappy.

Now imagine that the US already had an NGDP futures targeting regime in place—something similar to the “guardrails approach” that I advocated in chapter 5 of my new book.  Market forces nudged policymakers until the policy instruments were set at a level consistent with 4% expected nominal GDP growth, including reversion to trend after a deviation.  Then someone suggested getting rid of the program, and hiring Jay Powell to set the Fed’s policy instruments at a level that he thought was appropriate.  How would you feel about that decision?

One problem with my thought experiment is that we tend to have status quo bias.  Right now, the US has a discretionary policy regime.  In my thought experiment, we start with a rules based regime that uses market forces, and moves to a discretionary regime.  There are good reasons to be cautious when abandoning a system and adopting a new approach.  (Think about examples such as “Chesterton’s Fence”.)

If we currently had my preferred system in place, I don’t believe we would blow it up and move to a discretionary regime.  But how can we overcome status quo bias and get to this sort of regime?  That’s not obvious.

In my view, the best option is to move gradually to a market-oriented rules-based regime.  Thus the central bank might begin by creating a NGDP futures market and taking a short position on contracts linked to 6% NGDP growth and a long position on contracts linked to 2% NGDP growth.  If that went well, the following year the range could be reduced to 5.9% and 2.1%.  Each year, the guardrails would get a bit closer together.  Through trial and error, you could eventually determine what sort of band is optimal.

PS.  I know nothing about highway engineering, but I assume that something similar must have occurred with actual roadside guardrails.  If the guardrail is set 20 feet from the edge of the road, it’s too far away to do much good.  If it’s set one foot from the edge of the road, then even a momentary lapse in concentration from a driver could cause a costly scraping of paint from the passenger side of the car.  Most guardrails that I’ve seen are about 6 feet from the edge of the road.