What Is Inflation Targeting? Promising a Number and Being Judged On It

The 1970s left a specific diagnosis. Governments facing elections had an incentive to push for faster growth and lower unemployment through loose monetary policy, which worked briefly and then produced inflation; once the public expected that behaviour, it built the expected inflation into wage and price setting in advance, so the loose policy bought no extra employment and only higher prices. Economists called this the time inconsistency problem, and the papers by Finn Kydland and Edward Prescott setting it out won a Nobel Prize. The solution proposed was to remove the discretion: give the decision to an independent central bank, give it a single clear objective, and make its performance measurable so that failure is visible. A public target does one further thing that is easy to underrate, which is to anchor expectations, because if firms and unions believe inflation will be two percent, they set prices and wages accordingly and the belief becomes self-fulfilling.

The frameworks vary and share a common set of elements:

• •A numerical target, most commonly two percent annual consumer price inflation, sometimes as a point and sometimes as a band, chosen to be low enough not to distort decisions and high enough to leave room for cuts and to avoid deflation
• •A defined horizon, typically about two years, which acknowledges that policy acts with a long lag and that chasing every monthly figure would cause more instability than it removes
• •Operational independence, so the bank sets the interest rate without government instruction, while the target itself is usually set by the elected government
• •Transparency obligations, including published forecasts, minutes, votes and, in the British case, an open letter from the governor to the chancellor whenever inflation misses by more than one percentage point
• •Flexibility, since almost no regime targets inflation alone: the standard formulation allows the bank to look through temporary supply shocks and to take account of output and employment while returning inflation to target over the horizon

Adoption spread from New Zealand to Canada, Britain, Sweden, Australia and then across Europe, Latin America and Asia, with the European Central Bank and the Federal Reserve arriving at explicit numerical objectives by different routes. The period from the early 1990s to 2007 saw low and stable inflation across the adopting countries, along with reduced volatility in output, and was named the Great Moderation. How much credit the framework deserves is genuinely disputed, since the same period included the entry of China into world trade, a large expansion of global supply chains and favourable demographics, all of which pushed prices down for reasons unrelated to any central bank. The more defensible claim is narrower: inflation expectations became firmly anchored, which is measurable in bond markets and surveys, and that anchoring meant temporary price shocks stopped feeding into wages the way they had in the 1970s.

The framework has been tested from both directions and struggled each time. After 2008 the problem was inflation persistently below target with interest rates already at zero, which meant the main tool was exhausted, and the response was asset purchases and forward guidance, neither of which was part of the original design and both of which had large distributional side effects. Japan had faced the same problem since the 1990s and failed to escape it for decades. Then in 2021 inflation rose sharply for the opposite set of reasons, and most central banks initially described it as transitory and delayed raising rates, which in retrospect nearly all of them acknowledge was too slow. The defence is that the shock combined a pandemic, supply chain disruption, a war affecting energy and food, and large fiscal transfers, and that no forecasting framework handled it well. The criticism is that a framework built around anchoring expectations is precisely the one that should have moved early, and that the credibility it spent decades accumulating was drawn down heavily.

Several alternatives are under discussion and none has displaced the target. Average inflation targeting, adopted by the Federal Reserve in 2020, aims at two percent on average over time, which permits a deliberate overshoot after a period below target and was tested almost immediately under the worst possible conditions. Price level targeting commits to returning the price level itself to a path, which implies making up for past misses in both directions and is theoretically attractive and hard to explain publicly. Nominal income targeting aims at the total money value of output, which automatically tolerates higher inflation when growth is weak. Raising the target to three or four percent has been argued as a way to give more room above zero interest rates. Meanwhile the mandates themselves have broadened in practice, with financial stability, employment and in some cases climate risk added to what central banks are expected to consider, which pushes against the original insight that a single clear objective is what makes performance judgeable.

Inflation targeting gives an independent central bank one measurable objective, usually two percent over about two years, with published forecasts and an obligation to explain misses, and it was designed to solve the problem that governments cannot credibly promise not to inflate before an election. Its main achievement was anchoring expectations. It failed in one direction after 2008, when rates hit zero and inflation stayed too low, and in the other in 2021, when most central banks called the surge transitory and moved late, which is why the framework is now under review.