SOMETIMES doing nothing really is better than doing something. On September 17th the Federal Reserve made the right decision to leave its benchmark interest rate, unchanged since 2008, near zero. With inflation sitting well below the Fed’s 2% target and doubts about China’s economy prevalent, a rise would have been an unnecessary risk.

Yet Janet Yellen, chair of the Fed, will face a similarly tough choice in October—and possibly for many months thereafter. And whenever “lift-off” occurs, financial markets expect rates to stay historically low for years to come. The era of unconventional monetary conditions shows no sign of ending. If the rich world’s central banks are to get back to the normality they crave, their standard toolkit may not suffice. It is time to think more boldly, especially about the idea of inflation targeting.

That is because the usual relationship between inflation and unemployment appears to have broken down. In the short run, economists think these two variables ought to move in opposite directions. High joblessness should weigh on prices; low unemployment ought to push inflation up, by raising wages.

Unfortunately, in many rich countries this standard inflation thermostat is on the blink. In 2008 economic growth collapsed and unemployment soared, but inflation only gradually sank below target. Now, by contrast, unemployment has fallen to remarkably low levels, but inflation remains anaemic. This has wrong-footed central banks. Assuming that rising prices would follow hard on the heels of a jobs boom, both the Fed and the Bank of England ended stimulative bond-buying programmes and prepped markets for looming rate rises. Their recoveries have instead proved nearly inflation-free. Worse, with interest rates close to 0%, central bankers have less room to respond if they misread inflation risks and tighten too soon.

Given this double bind, it makes sense to look beyond inflation—and to consider targeting nominal GDP (NGDP) instead.

Nominal but substantive
 
A target for nominal GDP (or the sum of all money earned in an economy each year, before accounting for inflation) is less radical than it sounds. It was a plausible alternative when inflation targets became common in the 1990s. A target for NGDP growth (ie, growth in cash income) copes better with cheap imports, which boost growth, but depress prices, pulling today’s central banks in two directions at once. Nominal income is also more important to debtors’ economic health than either inflation or growth, because debts are fixed in cash terms.

Critics fret that NGDP is hard to measure, subject to revision, and mind-bogglingly unfamiliar to the public. Yet if NGDP sounds off-putting, growth in income does not. And although inflation can be measured easily enough, central banks now rely nearly as much on estimates of labour-market “slack”, an impossibly hazy number.

Most important, an NGDP target would free central banks from the confusion caused by the broken inflation gauge. To set policy today central banks must work out how they think inflation will respond to falling unemployment, and markets must guess at their thinking. An NGDP target would not require the distinction between forecasts for growth (and hence employment) and forecasts for inflation.

What might an NGDP target mean in practice? Most economies have fallen well short of their pre-recession trend in nominal-income growth. Before the financial crisis, nominal GDP growth of 5% was considered normal in America. Yet the economy is 16% below the income threshold it would have reached had it grown at that pace since 2006. In Britain, too, NGDP is 15% short of where it could have been. The euro zone and Japan are even worse. Such shortfalls are too great to make up quickly; doing so would imply dangerously high inflation rates. Yet even relative to recent trends, rich economies are coming up short; American NGDP is 5% below what you might have anticipated in 2010. Faster NGDP growth could come from better productivity, more hiring or faster inflation; all of which rich economies could use a bit more of.

Setting a different target does not mean central banks will automatically reach it. And their unconventional toolkit looks depleted. Quantitative easing, which is still in use in Europe and Japan, is falling out of favour because of worries about asset prices. Interest rates cannot be cut far below zero without radical changes in the nature of money (the Bank of England’s chief economist recently suggested eliminating cash). But getting the target right is an important start. Patiently waiting for inflation to turn up is no longer good enough.