Back in the 1990s inflation targeting was all the rage. I was a sceptic. I recall asking a senior central banker at the time what he would do if faced with stagflation — high inflation, low growth. Would he raise interest rates and force the economy into recession just to meet the target? He said the situation would never arise.

He was right. It did not. Inflation targeting became an improbable success. But it is failing now for reasons different from those I feared.

Under inflation targeting, a central bank sets itself a target inflation rate, usually about 2pc. Armed with new-generation economic models, the bank produces a forecast for inflation and gross domestic product, and fine-tunes its interest rate policies in a way that is calculated to meet the inflation target, usually within two years. A central banker following this regime is supposed to let bygones be bygones. If you miss the target, for whatever reason, you do not have to make up for it next year. You just try to hit it next time.

In an age when official interest rates are close to zero, this particular characteristic of inflation targeting has become a problem. Narayana Kocherlakota, president of the Minneapolis Federal Reserve, last week estimated that US inflation would stay below the Fed’s target for another four years. In the eurozone, the situation is worse. Core inflation — excluding food and energy — has been below 2pc for more than five years, and will remain low for several years to come. In Japan, by my calculations, the average inflation rate over the past 20 years was almost exactly zero. Once you are down there, it is hard to get up again.

In an age when official interest rates are close to zero, this particular characteristic of inflation targeting has become a problem

Why does inflation targeting not help? It has to do with averages. Take the perspective of a borrower who takes a 10-year fixed-rate loan. The real value of the debt repayments depends on future rates of inflation: the faster prices rise, the less consumption must be given up in order to pay off the loan. If inflation targeting works properly, there should be no surprises; when the loan is taken out, both borrower and lender have a pretty good idea of its real value. But now suppose that inflation undershoots the target in one year. The real value of the loan is therefore greater than expected. In a best-case scenario, inflation will gradually creep up over time. But even then, average annual inflation over the 10 years will be lower than the target rate.

Now consider an alternative policy. Instead of keeping prices rising at a steady rate, the central bank tries to hit a predetermined price level each year. If price rises undershoot the target in one year, the pace has to be quickened the following year to make up the gap.

For a borrower who takes out a 10-year loan, a price-level target offers more certainty, since missing the target in a single year need not any longer affect the average inflation rate for the decade as a whole.

Another option would be to target nominal output growth — which is best thought of as the sum of real economic growth and inflation. The trouble is that nominal output growth is so slow that, if you started this regime today, hitting the target would entail a larger stimulus programme than anyone would have the nerve to implement. If you are looking for a new policy, targeting the price level is a better choice.

The recent macroeconomic literature suggests that price-level targeting is superior to inflation targeting — if not in theory then in practice. Price-level targeting would be especially suited to the policy rules most central banks have adopted.

In the eurozone, most great ideas fail because of some legal or bureaucratic obstacle. Nominal GDP targeting may be considered incompatible with the price stability mandate of the European Central Bank. A price-level target, by contrast, fulfils that mandate by definition. The economic and legal arguments are, for once, both sound and aligned.

How would a price-level target work in the current setting? First you choose the historic price-level trend, from which we have departed. The policy goal would be to join up with the trend.

A price-level target constitutes a very strong form of forward guidance. You tell the markets that you will not raise interest rates immediately once inflation begins to rise. But you do not just make a vague promise. You commit to allowing inflation to rise for an extended period, as part of your main policy target.

I do not want to play down potential problems. For example, if inflation rates were to stay above the target for too long, a price-level targeting central bank would have to get really tough — much tougher than it would have to under an inflation target. Since central banks rarely want to precipitate recessions, the whole policy may lack credibility from the outset.

My answer to that criticism would be the same as the one I was given 20 years ago. If you are stuck in a low-inflation trap, that situation may never arise. We have bigger problems to worry about now.

Published in Dawn, Economic & Business, June 2nd, 2014