Non-Technical Summary
Do past inflation experiences leave a long-lasting imprint on how households form inflation expectations during periods of elevated inflation? A natural hypothesis—supported by a growing literature—is that the memory of past inflation can durably shape how people form beliefs about prices when inflation rises.
This paper exploits the century-long gap between two high inflation episodes in Austria. The first is the hyperinflation that followed World War I and the collapse of the Austro-Hungarian Empire. A hundred years later, Austria experienced a sharp inflation surge in the early 2020s, reaching levels not seen in decades. Our central question is whether individuals living in regions that were more exposed to Austria’s historical inflation episode report higher inflation expectations today.
To study this, we assemble and link two data sources. First, we use city-level prices for Austria’s ten largest cities over 1914–1922 to construct a local Retail Price Index (RPI) capturing the intensity of the historical inflation episode across locations, ranging from around 11 000% to 20 000% when cumulated over 8 years. Second, we use micro data from the European Commission DG-ECFIN quarterly consumer survey (2020–2024), which contains households’ 12-month-ahead inflation expectations along with socio-economic characteristics and geographic information.
Figure 1 illustrates the core pattern: inflation expectations are higher in cities that experienced more intense inflation in the early 1920s. Because the axis is in log historical (8-year) cumulative inflation, the unconditional fitted line implies a raw elasticity of about 0.35. This positive relationship is confirmed in richer specifications that control for common national conditions and respondent characteristics: our main result is a persistent association of about 0.37, conditional on common time effects and individual characteristics (a 1% higher regional cumulative inflation rate in the early 1920s is associated with about a 0.37% higher expected inflation today). The association is positive in both the low-inflation (May 2020–April 2021) and high-inflation (May 2021–August 2024) subperiods. In practical terms, moving from the least to the most historically exposed city is associated with inflation expectations about 9.5% higher (about 11.7% using the high-inflation-period estimate) — that is, on the order of one percentage point of expected annual inflation.
We also explore a potential channel through which such “inflation memory” may persist across generations: local media. We build regional indices of media exposure to inflation by combining regional readership patterns with counts of inflation-related newspaper articles. We find that areas with higher historical inflation tend to display greater exposure to inflation-related news over the period 2002-2024—both in the volume of inflation articles and in the share of coverage devoted to inflation—consistent with the idea that collective memory can be reinforced locally.
The implications for monetary policy are that the legacy of past inflation can generate heterogeneous expectations across regions within a country, even a century later. This can matter for the transmission of policy and for communication: clear, targeted communication may be especially valuable in regions where inflation history makes price increases more salient.
Keywords: Inflation, Inflation Expectations, Long-Term Persistence.
Codes JEL: D14, E31, E71, G41, N14