The 2021-23 high inflation episode and inequality: insights from the Consumer Expectations Survey

Prepared by Neus Dausà i Noguera, Georgi Kocharkov and Omiros Kouvavas

Published as part of the ECB Economic Bulletin, Issue 7/2025.

1 Introduction

The high inflation episode of 2021-23 gave rise to pronounced movements in inflation, labour income, mortgage rates and household net wealth in the euro area (Chart 1). Inflation outpaced nominal labour income growth, mortgage rates nearly doubled, and the growth of household net wealth slowed significantly.^[1] However, the implications of these aggregate developments for economic inequality remain unclear. This raises the broader question of how shifts in inflation and interest rates are reflected in distributional outcomes, an issue central to economic research and policy debates.

Chart 1

Developments in aggregate household income and wealth

a) Aggregate household income and inflation	b) Aggregate household net wealth and mortgage rates
(year-on-year percentage change)	(mortgage rate: percent per annum; wealth growth: year-on-year growth rates)

Income, consumption and wealth inequalities are key dimensions along which the distributional consequences of the 2021-23 inflation surge and higher interest rates can be understood. These disparities not only shape the economic realities of households, but also influence broader social cohesion and economic stability (Alesina and Rodrik, 1994; Alesina and Perotti, 1996, among others). The period since the pandemic has focused renewed attention on inequality, as the global economy has faced a series of major shocks and disruptions.^[2] This article uses the ECB’s Consumer Expectations Survey (CES) to examine perceptions and measures of inequality in the euro area. We show that while standard indicators of income, consumption and wealth inequality remained broadly stable between 2022 and 2025, perceptions of inequality rose sharply, with cost-of-living pressures seen as the main driver. To understand this disconnect, we analyse how inflation affected households across income groups by looking at differences in personal inflation rates and how behavioural differences in financial decisions, particularly regarding mortgages, shaped distributional outcomes during the period of rising interest rates.^[3]

The remainder of this article is structured as follows. Section 2 describes the evolution of perceptions and measures of inequality during the post-pandemic period of high inflation and interest rate changes, using survey-based perceptions of households as well as standard measures such as Gini coefficients of income, consumption and wealth and the at-risk-of-poverty rate, which focuses specifically on the lower end of the income distribution. Section 3 explores the direct effects of inflation on inequality by looking at how varying cost-of-living pressures have affected households. Section 4 examines how interest rate changes influence inequality through channels such as the timing of investments and the borrowing and saving decisions of households. Section 5 concludes.

2 Perceptions and measures of inequality

Recent evidence from the August 2025 CES wave provides insights into public perceptions of inequality during the period marked by high inflation and changes in interest rates. According to the survey, 73% of households in the euro area believe that inequality has increased either “a little” or “a lot” since the onset of the inflation surge in 2021 and the subsequent changes in interest rates. By contrast, only 5% of respondents believe that inequality has decreased.^[4] Chart 2, panel a) shows the net percentage of households perceiving an increase in inequality since 2021, broken down by income group and by country.^[5] Higher-income households are more likely to report rising inequality than lower-income households.

Chart 2

Perceived economic inequality

a) Perceived increase in economic inequality

(net percentage)

b) Reasons for perceived increase in economic inequality

(percentage of respondents)

Rising living costs were seen as the main drivers of inequality during the high inflation episode (Chart 2, panel b). When asked to identify the drivers of this perceived increase, 93% of respondents pointed to cost-of-living increases as the primary factor (with inflation considered the single most important driver, chosen by 84%).^[6] This indicates that households interpret inequality primarily through changes in purchasing power rather than through conventional distributional measures.

Other frequently cited reasons included unequal developments in wages (66%) or in financial wealth and asset prices (40%). Higher-income households (the top 20%) were more likely to attribute growth in inequality to factors such as unequal access to finance, investment opportunities or financial wealth and asset prices, whereas lower-income households (the bottom 50%) more often put it down to unequal employment developments or savings and borrowing costs.^[7] Concerning the latter, the bottom 50% of households were far more likely than the top 20% to view savings and borrowing costs as a key driver of inequality.

Measuring inequality requires a multifaceted approach, as different metrics capture distinct aspects of economic disparities. The Gini coefficient is a widely used measure that captures inequality across the entire distribution.^[8] It is particularly sensitive to changes around the middle.^[9] In comparison, the at-risk-of-poverty rate focuses more on the lower end of the distribution, highlighting the challenges faced by the most vulnerable populations.^[10]

Income, consumption and wealth inequality in the euro area remained broadly stable between 2022 and 2025 (Chart 3, panel a). The Gini coefficient of disposable household income (i.e. after redistribution through taxes and transfers) increased only very slightly, from 0.33 in 2022 to 0.34 in 2025, although this overall stability masked moderate cross-country differences.^[11] Consumption inequality also showed little change, with the Gini coefficient edging up from 0.31 to 0.32. Consumption inequality tends to be lower and less volatile than income inequality because households smooth spending over time and higher-income households have lower average propensities to consume.^[12] Wealth inequality displayed a mild increase, with the Gini coefficient of net wealth rising by about 0.02 points between 2022 and 2024, most of it occurring in 2022-23 during the period of higher inflation.^[13]

The risk of poverty remained broadly unchanged during the period from 2022 to 2025, suggesting relative resilience despite the inflation and interest rate shocks. The at-risk-of-poverty rate in the euro area increased only slightly, from 20.3% in 2022 to 20.7% in 2025 (Chart 3, panel b), suggesting that redistribution mechanisms and targeted interventions may have cushioned the shocks for low-income households.

Chart 3

Indicators of inequality

a) Gini coefficients of wealth, income and consumption

(Gini coefficient)

b) At-risk-of-poverty rate

(percentage of respondents)

3 Sources of inequality: inflation

Conventional indicators point to inequality remaining broadly stable, yet public perceptions tell a different story, particularly as regards the impact of the inflation surge across households. The specific composition of a household’s consumption basket, which is largely determined by income level, plays a critical role in shaping the inflation rate that a specific household experiences. For example, lower-income households tend to spend a higher proportion of their income on essentials like food and utilities, whereas higher-income households allocate more of their spending to discretionary goods and services such as recreation and leisure. During the 2021-23 inflation episode there were steep price increases across the board, but these were particularly pronounced in categories like food and utilities. This disproportionate rise in the cost of necessities had varying implications for households depending on their income and consumption patterns.^[14]

An analysis of the distributional effects of inflation highlights significant differences in how price changes affected households across income levels. We use individual-level data from the CES to construct personal inflation rates for households. This approach uses the self-reported consumption shares of households during the same period as weights to calculate individual-level price indices (see Box 1 for details on the methodology). Chart 4 illustrates how inflation rates varied across income quintiles, with a focus on the top 20% and the bottom 50%. When inflation peaked in October 2022, lower-income households experienced significantly higher inflation rates than their higher-income counterparts: the percentage point difference between the average inflation rates of these two groups was 0.55.^[15] However, as inflation began to moderate, this trend reversed: inflation in the services and recreation categories, which have a higher weight in the basket of higher-income households, remained elevated, pushing up the inflation rate for these households. In November 2023 the inflation rate for higher-income households was 0.84 percentage points higher than the rate for lower-income households.

Chart 4

Personal inflation rates

(left-hand scale: percentage points; right-hand scale: percentage point difference)

Measuring the cost of inflation in terms of the additional income needed to maintain the previous year’s consumption reveals that lower-income households faced a significantly higher burden during the inflation surge. For each household, we calculate the additional cost of maintaining the previous year’s consumption basket at current prices, expressed as a share of current disposable income. This measure aligns conceptually with the compensating variation used in welfare analysis.^[16] Chart 5 illustrates how this additional cost evolved across income quintiles over time. The findings reveal that since 2022 lower-income households have been bearing a disproportionately higher burden. In 2022 the cost of inflation for the lowest income quintile amounted to 12.28% of their current income, compared with 5.69% for the highest quintile. By 2025, as inflation returned to target, these differences narrowed, but were not reversed.^[17] Taken together, the evidence in Charts 4 and 5 suggests that differential inflation experiences widened the distribution of real consumption across households, even though standard income inequality measures remained broadly stable. While fiscal transfers, wage adjustments and redistribution mechanisms may have helped to stabilise disposable incomes, they did not fully offset the uneven burden of higher living costs.

Chart 5

Inflation cost by income

(percentage points)

Box 1
Measuring personal inflation using detailed consumption data from the Consumer Expectations Survey

Prepared by Neus Dausà i Noguera, Georgi Kocharkov, Omiros Kouvavas and Athanasios Tsiortas

The direct impact of inflation on households depends on their consumption basket and on changes in their consumption, income and wealth levels. Headline inflation provides a measure of the aggregate state of the economy but neglects the differences in prices experienced by households. Having an individual measure of inflation allows for a better understanding of the distributional effects of price changes. This box introduces a new measure of individual-specific inflation based on household-level consumption data and describes its novel features and methodology.

The new metric makes it possible to measure inflation across various socio-demographic groups (opening the door to more granular studies) and facilitates timely assessment by using up-to-date data. Moreover, by using concurrent expenditure shares it accounts for substitution effects at higher levels of aggregation (upper-level substitution), offering a more accurate representation of inflation dynamics.^[18]

A growing literature has developed alternative approaches to the measurement of household-level inflation. Kaplan and Schulhofer-Wohl (2017) show that inflation rates differ significantly across US households, mainly because of price differences for identical goods. Marenčák and Nghiem (2025) use CES data for the euro area but rely on a simpler aggregation that covers around 85% of HICP items. Kukk et al. (2025) study heterogeneity in Estonia by linking bank data with the Household Budget Survey. By contrast, our approach maps all HICP items to CES categories and follows the official HICP methodology, ensuring broader coverage and closer alignment with official statistics.

Our measure relies on two data sources: the CES and the HICP. The quarterly module of the CES collects data on individual nominal expenditure for 15 different consumption categories, covering all items in the HICP. This can then be used to create a consumption basket per respondent per year.^[19] The consumption basket is calculated using the average across the year for each category.^[20]

We match the 15 CES categories to items at the COICOP-5 level (220 series in total from official HICP data), which are then aggregated into broader groups. Thus, for each CES category, the resulting price index (PI) for yeary and month mis an unchained Laspeyres-type index where the weights (w) have been adjusted to reflect the relative importance within that specific category. Additionally, the category price indices of the COICOP-5 items have been unchained to represent the growth rate since the previous December.^[21] The resulting index is given by:

PIK,y,m=∑k∈KwK,k,y∑k∈KwK,k,y⋅PIK,k,y,mPIK,k,y-1,12,

where K indicates the CES category and k is a subcomponent of category K.

The final individual-level price index for a given year and month is then a Paasche-type price index defined as:

PIi,y,m=∑KwK,y,iCES^⋅PIK,y,m

These individual-level unlinked indices are then chain-linked again across years at any desired level of aggregation (e.g. by income quantiles or other demographic sub-groups).

However, another adjustment needs to be made before the indices can be aggregated into one price index per individual household. The weights used in these consumption baskets are nominal shares, capturing the effect of price changes. To address this, the weights need to be price-updated by converting the average price underlying expenditure shares into December prices, as follows:

wK,y,iCES^=wK,y,iCES12-1∑m=112PIK,y,m⋅PIK,y-1,12

Chart A shows the results of the aggregated personal inflation indices in comparison to the HICP. As these indices use consumption from the current year rather than the penultimate year like their underlying weights, in principle they demonstrate a higher substitution effect – taking account of higher-level substitution between the 15 consumption items elicited in the CES.

Chart A

Personal inflation and HICP

(percentages per annum)

Even though both indices shown in Chart A describe the price changes an average euro area household experienced over time with the same scope, the index based on personal consumption baskets appears to be slightly lower, particularly during the period of elevated inflation. However, it was higher than the HICP during the period when inflation eased. This observation appears to be consistent with the literature on substitution effects. Overall, results from the CES-based personal inflation index match the HICP dynamics well and allow for further decomposition analyses (some of which are covered in this article).

4 Sources of inequality: interest rates

Interest rate adjustments affect households unevenly, reflecting differences in income, wealth, financial constraints and financial literacy. The impact of interest rate levels and changes depends on a household’s capacity to adapt to new economic conditions and on its knowledge of appropriate financial responses. Both of these factors vary systematically with the level of wealth. This section examines the timing of households’ financial decisions under different interest rate regimes and the heterogeneous responses of households to changes in interest rates.

Exposure to interest rate risk is a key factor that shapes the impact of higher interest rates and varies markedly across income groups. At the onset of the monetary policy tightening cycle, lower-income households were disproportionately affected for two main reasons: (i) they relied more heavily on adjustable-rate mortgages, and (ii) they had less favourable options regarding the fixation length of fixed-rate loans. As shown in Chart 6, panel a), a larger share of lower-income mortgagors hold adjustable-rate mortgages, which are much more sensitive to rising interest rates. In addition, lower-income mortgagors often select shorter fixation periods on fixed-rate loans, leaving them more exposed to shifts in interest rates.^[22]

The approach of higher-income households to managing interest rate risk appears to be better-informed. These households seem to internalise whether the prevailing interest rate environment is relatively low or high and adjust the fixation length of their mortgage loans accordingly. Chart 6, panel a) shows that higher-income households tend to choose a median fixation length of 15 years during periods of low interest rates and a median fixation length of ten years during periods of high rates. By contrast, lower-income households adopt a mostly similar fixation strategy, regardless of the interest rate environment. Choices regarding loan fixation lengths and exposure to interest rate risk can exacerbate financial disparities, as the outcomes of these decisions are directly tied to households’ financial resilience during periods of changes in interest rates.^[23]

Chart 6

Fixation periods, interest rate risk and the effects on disposable income inequality

a) Fixation periods and interest rate risk

(left-hand and middle panels: percentage of mortgagors; right-hand panel: years)

b) Impact of increased payments on income

(left-hand panel: changes in Gini coefficient; right-hand panel: percentages)

The degree of household interest rate risk, coupled with the size of household mortgages, influences the effective impact of interest rates on disposable income and inequality. While higher-income households are more likely to hold larger liabilities (e.g. mortgages), their ability to adjust their financial strategies mitigates the adverse effects of rising interest rates. By contrast, lower-income households are more exposed to interest rate risk. Box 2 quantifies the differential impact of higher interest rates across income groups, expressed as a percentage of household income, revealing the disproportionate burden borne by lower-income households. Using the results from Box 2, in Chart 6, panel b) we show the effect on the disposable income (after debt repayments) Gini coefficient by adjusting for the cumulative increased payments per household and comparing this with the baseline without the interest rate increases. Overall, for mortgagors, post-repayment income inequality increases across most countries and in the euro area. However, this is balanced out by the effect of higher-income households having a higher share of mortgages (Chart 6, panel b, right-hand side). In the absence of differential interest rate risk across mortgagors, increases in interest rates would actually reduce post-repayment income inequality owing to the disproportionately larger share of liabilities held by high-income households, however this effect is reversed owing to unequally distributed interest rate risk.

Box 2
Who bears the costs of higher interest rates? A microsimulation using the Consumer Expectations Survey

Prepared by Luca Caprari and Omiros Kouvavas

To quantify the effect of interest rate hikes on households, we use household-level data – which include information on borrower characteristics, estimated risk premia by income quantile, type of loan and timing of mortgage adjustments – to create a full picture of interest rates changes and increased payment amounts across the period from 2022 to 2025.

Households that hold adjustable-rate mortgages have interest rates that adjust each period, whereas households with fixed-rate mortgages have predetermined mortgage payments for a fixed period of time. Taking this into account, and using the available data, we calculate household-level interest rate changes over time matched with actual time-specific loan amounts. This allows us to calculate the nominal amount of the increased payments owing to interest rate changes.

Chart A, panel a) shows the mean interest rate changes for mortgagors between 2022 and 2025 for the bottom 50% and top 20% of the income distribution. The increase in interest rates is greater for mortgagors in the bottom 50%, driven by their higher interest rate risk owing to their higher share of adjustable-rate mortgages and their higher share of fixed-rate loans expiring during the period in question. This can be combined with the updated loan amounts and translated into the size of the respective payment increase corresponding to that year for the specific household. Chart A, panel b) shows the average increase in payment costs for mortgagors normalised to income and by household income quintile.

Chart A

Changes in interest rates and payment costs for mortgagors

(percentage points)

Differences in financial constraints and financial literacy affect household responses during and after periods of higher interest rates. Lower-income households not only face greater exposure to interest rate risk; they also exhibit distinct behavioural patterns in their credit demand. Specifically, these households tend to increase their credit applications during periods of higher interest rates, often relying more heavily on adjustable-rate mortgages.^[24] This reliance further amplifies the vulnerability of lower-income households to rising interest rates, perpetuating financial fragility and widening inequality.

5 Conclusion

Despite major shocks, broader indicators of inequality remained relatively stable between 2021 and 2025. Income and wealth inequality showed little change and the poverty rate remained contained. This aggregate stability can be attributed, at least in part, to heterogenous compensating exposures across households and to risk-sharing mechanisms that cushioned vulnerable groups.

The overall stability in measured inequality may mask changes in the underlying distribution of income and wealth. As documented in this article, according to the CES, households differed significantly in how they experienced both the surge in inflation and the rise in interest rates. The 2021-23 inflation shock placed a heavier burden on lower-income households, although this effect receded as inflation subsided. Subsequent interest rate increases also had distributional effects: lower-income households, which more often rely on adjustable-rate mortgages, faced higher repayment burdens, whereas higher-income households were better able to adapt.

Perceptions of higher inequality following the inflation surge may persist, despite the stability in measured inequality. While standard measures of inequality may have remained stable overall over the past few years, this does not exclude the possibility that some redistribution of disposable income and purchasing power has indeed taken place over this period. Reconciling the relative stability of measured inequality with the sharp increase in perceived inequality thus remains a topic for future research. As perceptions of inequality are likely to influence households’ attitudes and behaviour, these should be closely monitored alongside standard measures of inequality.

References

Acemoglu, D. and Restrepo, P. (2020), “Robots and Jobs: Evidence from US Labor Markets”, Journal of Political Economy, Vol. 128, No 6, pp. 2188-2244.

Alesina, A. and Perotti, R. (1996), “Income distribution, political instability, and investment”, European Economic Review, Vol. 40, No 6, pp. 1203-1228.

Alesina, A. and Rodrik, D. (1994), “Distributive Politics and Economic Growth”, The Quarterly Journal of Economics, Vol. 109, No 2, pp. 465-490.

Baptista, P., Dossche, M., Hannon, A., Henricot, D., Kouvavas, O., Malacrino, D. and Zimmermann, L. (2025), “The transmission of monetary policy: from mortgage rates to consumption”, Economic Bulletin, Issue 4, ECB.

Bobasu, A., di Nino, V. and Osbat, C. (2023), “The impact of the recent inflation surge across households”, Economic Bulletin, Issue 3, ECB.

Boskin Commission (1996), Toward A More Accurate Measure Of The Cost Of Living – Final Report to the Senate Finance Committee from the Advisory Commission To Study The Consumer Price Index, United States Social Security Administration.

Charalambakis, E., Fagandini, B., Henkel, L. and Osbat, C. (2022), “The impact of the recent rise in inflation on low-income households”, Economic Bulletin, Issue 7, ECB.

Cowell, F. A. (2011), Measuring Inequality, Oxford University Press, Oxford.

ECB (2021), “Inflation measurement and its assessment in the ECB’s monetary policy strategy review”, Occasional Paper Series, No 265, ECB, Frankfurt am Main, September.

Eurostat (2024), Harmonised Index of Consumer Prices (HICP) Methodological Manual – 2024 edition.

Henricot, D., Pöschl, J. and Tsiortas, A. (2025), “The heterogenous transmission of monetary policy to household credit”, Economic Bulletin, Issue 7, ECB

Kaas, L., Kocharkov, G. and Preugschat, E. (2019), “Wealth Inequality and Homeownership in Europe”, Annals of Economics and Statistics, No. 136, pp. 27-54.

Kaplan, G. and Schulhofer-Wohl, S. (2017), “Inflation at the household level”, Journal of Monetary Economics, Vol. 91, pp. 19-38.

Kukk, M., Toczynski, J. and Basten, C. (2025), “Beyond the headline: How personal exposure to inflation shapes the financial choices of households”, Journal of Monetary Economics, Vol. 153, July 2025, 103800.

Marenčák, M. and Nghiem, G. (2025), “Personal inflation rates in the euro area”, CAMA Working Papers, No 17/2025, Centre for Applied Macroeconomic Analysis, Crawford School of Public Policy, The Australian National University.

Lerman, R. and Yitzhaki, S. (1989), “Improving the accuracy of estimates of Gini coefficients”, Journal of Econometrics, Vol. 42, No 1, pp. 43-47.

Orchard, J. (2025), “Non-homothetic demand shifts and inflation inequality”, Finance and Economics Discussion Series, No 2025-085, Board of Governors of the Federal Reserve System, Washington DC.

Pallotti, F., Paz-Pardo, G., Slacalek, J., Tristani, O. and Violante, G. (2024), “Who bears the costs of inflation? Euro area households and the 2021–2023 shock”, Journal of Monetary Economics, Vol. 148, November 2024, 103671.