In 2021-23, the world economy experienced supply shocks and inflation levels of magnitudes not seen since the 1970s. Central banks responded with policy decisions of magnitudes equally unseen in several decades. Between July 2022 and September 2023, the ECB hiked its policy rates by 450 basis points, the fastest increase since its inception. According to Lhuissier (2025)’s estimates, this lowered inflation by 2.5 to 3 percentage points in 2024.
This estimate, however, only accounts for part of the contribution of monetary policy to lower inflation. It measures the contribution of the monetary policy decisions taken from 2021 onwards, assuming that inflation expectations would have remained well-anchored even if the ECB had not reacted at all. Yet, well-anchored inflation expectations (Burban et al., 2024) were the result of previous decades of low and stable inflation, an environment that central banks delivered thanks to central bank independence, greater transparency and communication, and a commitment to price stability (Bignon and Gautier, 2023). The good track record of their past policies allowed them to maintain well-anchored expectations, an asset that proved very important in dampening the effect of the large supply shocks of 2021-23.
Learning gain as a measure of expectations anchoring
A common assumption on the way households and firms form expectations is to assume that they try to learn what inflation will be from what it has been in the past. Under this assumption, households and firms increase their inflation expectations whenever inflation turns out higher than expected. But they may increase them by a little or by a lot. A simple way to measure how well long-term inflation expectations are anchored is to estimate how much they increase with recently realised inflation.
This can be measured by the learning gain, which may be assumed to be constant over time (the assumption of constant-gain learning). Under a low learning gain, long-term inflation expectations depend little on the very recent realisations of inflation. Instead, they depend mostly on what inflation has been over a long period of time: expectations are well-anchored. Under a high learning gain, long-term inflation expectations depend heavily on the recent realisations of inflation, and the more distant past is quickly forgotten. Expectations are poorly anchored. A good track record of the central bank can deliver not only low inflation but also a low learning gain, guaranteeing that a short-lived burst of inflation has a limited effect on long-term inflation expectations.
Since the inception of the euro area, the learning gain is estimated to be small. Using the 5-year-ahead average inflation expectations from the Survey of Professional Forecasters (SPF) over 1999-2024, we find a quarterly gain of 0.01. This means that if inflation in a given quarter is one percentage point higher than expected, professional forecasters increase their inflation expectations by one basis point on average. Chart 2 shows how estimated long-term inflation expectations with a learning gain of 0.01 tracks the average 5-year-ahead inflation expectations of the SPF.
Chart 2: Long-run inflation expectations in the euro area