LIKE other areas of public policy, central banking is prone to fads and fashions. From limits on money-supply growth to pegging exchange rates, orthodoxies wax and wane. Yet the practice of inflation-targeting has proved remarkably long-lived. For almost three decades, central bankers have agreed that their best route to stabilising an economy is to aim for a specific target for inflation, usually 2% in advanced economies and a little higher in emerging ones.

This orthodoxy is still intact in many emerging economies where inflation is yet to be tamed.

But in the rich world the consensus is beginning to fracture. As central bankers gather this weekend for their annual shindig in Jackson Hole, Wyoming, John Williams, head of the San Francisco Fed, has caused a stir by suggesting it is time for a rethink on what central banks should aim for. He is right.

The reason is that the rich world’s central banks are working in a different context from the 1990s, when today’s inflation-targeting doctrine was formed. Then, it seemed that inflation would spend as much time above target as below it. And the “natural real rate of interest”—the inflation-adjusted price that balances the supply of, and demand for, savings in a full-strength economy—was as high as 3.5%. But inflation has been below the central bankers’ target for years. And the underlying real natural rate of interest has fallen to 1% or lower, probably because population ageing has boosted saving even as lower expectations of growth have cut investment.

This matters because low inflation and a low natural interest rate limit the effectiveness of central bankers’ traditional policy lever: setting short-term interest rates. Since nominal interest rates are the sum of real rates and inflation, the rich-world central banks cannot, under today’s regime, expect their policy rates to rise much higher than 3% (the 2% inflation target plus a 1% real rate). That leaves very little room to cut when the next recession strikes. In the three most recent recessions the Fed slashed rates by 675 basis points (hundredths of a percentage point), 550 basis points, and 512 basis points.

Fear of future impotence is the main cause of today’s misgivings over a low inflation target. But there are other drawbacks with the current regime. First, a target for annual inflation gives the central bank no leeway to make up for periods during which inflation has been too high or too low. If central bankers could credibly promise that they would allow a burst of catch-up inflation, they might be more successful at boosting too-low inflation today. Second, when supply shocks such as a sudden rise or fall in the oil price send inflation and economic growth in opposing directions, central bankers face a tricky choice of which to respond to.

How might these problems be fixed? One possibility is simply to raise the inflation target to, say, 4%.

Credibly enacted, that ought to alleviate the risk of impotence. If investors and consumers believe inflation will reach 4%, nominal interest rates should eventually rise to 5% or so even if real rates stay low. But rich-world central banks have undershot their targets for so long they may struggle to persuade the public to expect higher inflation. And a higher target would still leave central banks with a dilemma when economic growth and inflation diverge. Neither would it make up for big misses.

Who ate all the pi?
A more radical option is to move away from targeting inflation altogether. Many economists (and this newspaper) see advantages in targeting the level of nominal GDP, the total amount of spending in the economy before adjusting for inflation. A nominal-GDP target would allow for temporary variations in inflation. Downturns would be tempered by an expectation of protracted stimulus later on to make up lost ground. In better times, a rise in real GDP would provide the lion’s share of the required nominal-GDP growth and inflation could drift lower.

Changing targets is not something policymakers should do lightly; their credibility depends on stability. And, like every regime, a nominal-GDP target has its drawbacks, not least that few non-economists have ever heard of the concept. It will not be easy to build a consensus for it.

But it is right to start doing so. A 2% inflation target is ill-suited to the rich world today.

Doubling it would be an improvement, but targeting nominal GDP would be better still. Time for a new era.