WORLD BANK GROUP Poverty, inclusion, and employment: The path toward a resilient Europe Poverty, inclusion, and employment: The path toward a resilient Europe Poverty and Equity Global Practice Europe and Central Asia Region Monica Robayo-Abril, Nga Thi Viet Nguyen, and Lukas Delgado The World Bank 2 © 2023 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e- mail: pubrights@worldbank.orgpubrights@worldbank.org. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank, therefore, does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. 3 Table of Contents Acknowledgments ........................................................................................................................................ 6 I. The resilience of the EU labor market to the COVID-19 crisis and the Ukraine war ........................ 8 II. Welfare impacts of the cost-of-living crisis on European households in selected countries ......... 15 III. Policies toward a more resilient Europe ......................................................................................... 25 References .................................................................................................................................................. 27 Annex: Methodology .................................................................................................................................. 28 Methodology 1: Measuring Direct and Indirect Welfare Impacts of Food and Energy Inflation ............... 28 Methodology 2: Measuring Overall Impacts of Overall Inflation using Weighted Adjusted CPIs .............. 30 Figures Figure 1 Labor force participation, employment, and unemployment rates (%15-64), EU27, 2019Q3- 2022Q3.......................................................................................................................................................... 8 Figure 2 Employment Growth by sector (%), EU27, 2019Q3-2022Q3 .......................................................... 8 Figure 3. Job vacancies growth index (2019Q3=100), by subregions ........................................................... 9 Figure 4 Labor market tightness, 2019Q3-2022Q3, by subregions .............................................................. 9 Figure 5 Employment rate (%), by age and sex, EU 27 ............................................................................... 11 Figure 6 Youth unemployment Rates (% labor force ages 15-24), EU27, and subregions ......................... 11 Figure 7 Employment growth index (2019Q3=100) by education level, EU27 ........................................... 12 Figure 8 Employment growth index (2019Q3=100) by subregion, Workers with primary education ...... 12 Figure 9 Employment growth index (2019Q3=100), by occupation, 2019Q3-2022Q3, EU27.................... 13 Figure 10 Type of employment by gender index (2019Q3=100), EU 27, 2019Q3-2022Q3 ........................ 13 Figure 11 Percentage change in real labor cost index, EU countries, 2019Q3-2022Q3 ............................ 14 Figure 12 Percentage change in the working hours of full-time workers, 2019-2021 ............................... 15 Figure 13 Evolution of energy prices (index) by countries, 2020M1-2022M12 ......................................... 16 Figure 14 Evolution of food prices (index) by countries, 2020M1-2022M12 ............................................. 16 Figure 15 Difference in the inflation of bottom and top income quintiles by countries, 2020M1-2022M12 .................................................................................................................................................................... 16 Figure 16. Simulated changes in poverty due to food and energy inflation, Direct and Indirect Impacts, USD 6.85 poverty line (2017 PPP) ............................................................................................................... 18 Figure 17 Simulated changes in household disposable income, by income deciles (direct Effects) .......... 20 Figure 18 Simulated Changes in the Gini coefficient (direct effect) ........................................................... 20 Figure 19 Simulated changes in per capita disposable income, by rural and urban .................................. 21 Figure 20 Simulated changes in per capita disposable income, by population subgroups ........................ 21 Figure 21: Changes in consumption patterns in response to rising prices ................................................. 22 Figure 22: Chief areas of concern among Bulgarian households, April 2022 ............................................. 22 Figure 23: Harder to make ends meet compared to pre-crisis level .......................................................... 22 Figure 24: Croatian households reporting arrears in utility bills over time ................................................ 23 4 Figure 25: Croatian households reporting arrears in phone bills over time ............................................... 23 Figure 26: Romanian households' financial prospects over time ............................................................... 23 Figure 27: Share of households reporting worsening financial situation in the next 12 months, by country .................................................................................................................................................................... 23 Figure 28: Share of households expecting less spending on durable goods in the next 12 months, by country ........................................................................................................................................................ 24 Figure 29: Share of households expecting less spending on restaurants in the next 12 months, by country .................................................................................................................................................................... 24 5 Acknowledgments This work was prepared by a team from the Poverty and Equity Global Practice (GP). The lead authors were Monica Robayo-Abril and Nga Thi Viet Nguyen (both Senior Economists, Poverty and Equity GP). Lukas Delgado and Britta Rude (both Consultants, Poverty and Equity GP) made significant contributions to this report. The team was indebted to Tom Bundervoet and Moritz Meyer for their insightful comments. The note was prepared under the guidance of Gallina (Country Director, ECCEU) and Salman Zaidi (ECA Practice Manager, Poverty, and Equity GP). 6 Poverty, inclusion, and employment: the path toward a resilient Europe Abstract This policy note focuses on the welfare impact of the cost-of-living crisis across EU countries, regions, and sub-populations. The COVID-19 crisis and the spillovers from the war in Ukraine have had asymmetric effects both across and within countries depending on household characteristics and main sources of income. The recovery has been heterogeneous across regions in the EU and across population segments in the 4 EU countries in light of the cost-of-living crisis and expected growth deceleration. There are two key welfare channels affecting households: the employment and the expenditure channel. First, we describe the employment channel by characterizing the more recent labor market trends and the shape of the recovery across subregions and population subgroups for 2019Q3-2022Q3. Since the beginning of 2021, the labor market experienced a rebound in employment in line with the resumption of economic activity on the back of solid vaccination campaigns. Though promising, the headline figure does not accurately depict the differences in the labor market's rebound across countries or employment types. Since much of the inequality prevalent throughout the EU is due to inequality in the labor market, uneven recovery in the labor market has implications for widening income inequality. Second, we describe the expenditure channel: the welfare impacts of rising prices (overall, by regions and across groups) due to loss in purchasing power and impacts on living conditions. Our results show that welfare losses could be sizable, particularly at the bottom of the distribution, with indirect or second-order impacts playing an important role. Finally, given these challenges, we describe what we expect in the medium term and potential policy options to tackle them. 7 I. The resilience of the EU labor market to the COVID-19 crisis and the Ukraine war 1. The European Union (EU) has entered a recovery period during which employment and participation gradually rose, and unemployment dropped, reaching a complete rebound by the third quarter of 2022, compared to the same quarter in 2019. In the early stages of the pandemic, many businesses were forced to close due to lockdowns and social distancing measures. This led to a sharp increase in unemployment and a decrease in labor force participation that gradually improved as the economic and health impacts of the pandemic were much smaller. According to the latest quarterly labor data (third quarter of 2022), the employment and labor force participation rates were about 1.5 and 1.2 percentage points higher than in the same quarter of 2019, while the unemployment rate was slightly below (0.5 percentage points). In Q3 of 2022, the unemployment rate in the EU fell to a historically low rate of 6.1%. The core labor market indicators suggest that, in general, the labor market in the EU has been improving and remarkably resilient to COVID-19 and the Ukraine crisis (Figure 1). 2. While employment in most sectors has fully recovered, the lingering impacts of the crises still shadow some sectors, notably agriculture, accommodation, and food. Nearly all major industries experienced employment gains (Figure 2). Certain industries are well above their pre-pandemic levels; for example, information and communication and real estate activities had positive employment growth from the third quarter of 2019 to the third quarter of 2022, 20 and 14.7 percent, respectively. In the digital sector, the shift to remote work as a response to the pandemic has increased the labor demand for workers. Yet, even if supply shortages and business restrictions eased in 2021, some industries continued to suffer the effects of the COVID-19 pandemic. For example, the agricultural sector employed around 12.5 percent fewer individuals in the same period, followed by administrative service activities (7.5 percent) and accommodation and food activities (4 percent). Manufacturing employment growth was notably subdued due to critical upstream providers' insufficient supply of goods and services (EC, 2022). Agricultural employment was already decreasing even before the pandemic, partly due to the steady growth in the service sector, which created more job opportunities outside of rural areas. Figure 1 Labor force participation, Figure 2 Employment Growth by sector (%), EU27, employment, and unemployment rates (%15-64), EU27, 2019Q3-2022Q3 2019Q3-2022Q3 8 AGRICULTURE, FORESTRY AND FISHING -12.5 ADMINISTRATIVE AND SUPPORT… -7.5 75 9 ACCOMMODATION AND FOOD SERVICE… -4.0 8 MINING AND QUARRYING -2.4 Labor Force participation and Employment (%) 73 MANUFACTURING -1.5 7 WHOLESALE AND RETAIL TRADE;… -0.1 Unemployment Rate (%) 6 TRANSPORTATION AND STORAGE 0.7 71 CONSTRUCTION 1.0 5 WATER SUPPLY; SEWERAGE, WASTE… 3.2 4 HUMAN HEALTH AND SOCIAL WORK… 3.4 69 PUBLIC ADMINISTRATION AND… 3.7 3 ARTS, ENTERTAINMENT AND… 4.6 67 2 EDUCATION 4.6 1 ELECTRICITY, GAS, STEAM AND AIR… 6.0 FINANCIAL AND INSURANCE ACTIVITIES 6.1 65 0 PROFESSIONAL, SCIENTIFIC AND… 7.3 ACTIVITIES OF EXTRATERRITORIAL… 10.6 OTHER SERVICE ACTIVITIES 12.3 Labor Force Participation rate REAL ESTATE ACTIVITIES 14.7 Employment rate INFORMATION AND COMMUNICATION 20.0 Unemployment rate -15.0 -10.0 -5.0 0.0 5.0 10.0 15.0 20.0 25.0 Employment change EU27, 2019Q3-2022Q3 Note: The rates are calculated for Note: Total employment for individuals between individuals between 15 and 64 years. Source: 15 and 64 years. Source: Eurostat (lfsq_egan22d), 2019Q3- Eurostat (lfsq_argaed, lfsq_ergaed and 2022Q3. lfsq_urgaed), 2019Q3-2022Q3. These series are not seasonally adjusted 3. With the recent increase in job vacancies and the relatively stable unemployed population, the labor market is becoming tighter in EU27 and across most European subregions. During the pandemic, job vacancies decreased, at most, between 18 and 34 percent of their pre-pandemic levels, yet they recovered fastly for most European subregions1 (except for CEE). An Increasing labor market tightness (as measured by an index of vacancies over the number of unemployed) in the EU, driven in large part by WE and NE regions, indicates that the recent growth in vacancies has been faster than the growth in unemployed. This hints at some tightening of the labor market and that it is becoming more challenging for employers to fill job vacancies. This can increase the wages firms offer as workers' outside options improve, especially in WE and NE, where the most significant increases in labor market tightness exist. More recently, in Q3 of 2022, job vacancies started decreasing for NE, WE, and CEE, likely due to tightening economic conditions with increasing interest rates and prices, but nevertheless vacancy rates and labor market tightness remain above pre-pandemic levels. (Figure 3). Figure 3. Job vacancies growth index Figure 4 Labor market tightness, 2019Q3-2022Q3, by (2019Q3=100), by subregions subregions 1 European subregions are constructed as follows: Northern Europe (NE) consists of Denmark, Estonia, Finland, Latvia, Lithuania and Sweden; Southern Europe (SE) consists of Cyprus, Spain, Italy, Portugal, Malta and Greece; Western Europe (WE) consists of Austria, Belgium, France, Ireland, Luxembourg, Netherlands and Germany; Central Eastern Europe (CEE) consists of Bulgaria, Czech Republic, Hungary, Croatia, Poland, Romania, Slovakia and Slovenia. 9 0.5 180 0.45 0.4 160 Labor market tightness (v/u) 0.35 0.3 140 0.25 120 0.2 0.15 100 0.1 0.05 80 0 60 NE SE WE CEE EU 27 NE SE WE CEE EU 27 Note: Total job vacancies include services, Note: Labor market tightness is constructed as job industries, and construction sectors. There is no available vacancies for services, industries, and construction sectors over information for Italy, Germany, and Estonia. This unemployed people between 15 and 64 years. This excludes the excludes the primary sector and international primary sector and international organizations and activities of organizations and activities of households as employers. households as employers. The vacancies are not seasonally Vacancies are not seasonally adjusted. Source: Eurostat adjusted. There is no available information for Italy, Germany, and (jvs_q_nace2), 2019Q4-2022Q3. Estonia—source: Eurostat (jvs_q_nace2 and lfsq_ugaed), 2019Q4- 2022Q3. 4. The overall improvement of key labor market indicators masks the uneven path of recovery for specific subgroups; the recovery in youth employment is still lagging, and job quality remains a significant issue. The pandemic exacerbated the labor market challenges of the youth (15-24), who have experienced much higher employment losses than older adults (55-64). While highly resilient in the recovery period- with employment rates growing significantly faster - the rapid gains barely compensate for the larger post-pandemic losses, and their employment growth has been overwhelmingly concentrated in part-time jobs. Contrarily, older individuals were more or less resilient after the pandemic, but their employment rates grew slower afterward (Figure 5). The recent increase in employment for older workers is mainly due to the increase in full-time jobs. Part-time employment generally comes with many disadvantages vis-à-vis full-time jobs, such as lower overall hourly compensation and other benefits. Note that the participation rates evolve similarly to employment for these subgroups. 5. Despite the labor market's rebound, young workers continue to face high unemployment rates; gender gaps in labor force participation have barely closed, except for those in the CEE region. As the pandemic hit young workers exceptionally high -many of them were laid off from the most affected industries, like, hospitality, retail, and personal services- there is a strong recovery in the unemployment rate across European subregions. However, the youth unemployment rate remains substantially high, 10 hovering at around 15 percent in most parts of Europe. Moreover, South Eastern Europe stands out, with more than 25 percent of active people between the ages of 15 and 24 unemployed (Figure 6). Gender gaps in labor force participation persist in most subregions, hovering around 12 and 13 percentage points in the CEE and the SE regions, compared to small gaps in NE. Little improvements have been observed, with the exception of the CEE region, which experienced a 2.1 percentage point increase on average. Figure 5 Employment rate (%), by age and sex, EU 27 Figure 6 Youth unemployment Rates (% labor force ages 15-24), EU27, and subregions 70 40 68 38 40 Young Employment Rates (%) Older Employment Rates (%) Young Unemployment Rates (%) 66 36 35 64 34 30 25 62 32 20 60 30 15 58 28 10 56 26 5 54 24 0 52 22 50 20 NE SE WE CEE EU 27 Female, 55-64 Male, 55-64 Female, 15-24 Male, 15-24 Note: Employment rate for individuals Note: Unemployment rate for individuals between between 15 and 24 years (young) and 55 to 64 15 to 24 years. Information for Germany is not available. years (old). Source: Eurostat (lfsq_ergaed), Source: Eurostat (lfsq_ugad and lfsq_agaed), 2019Q3- 2019Q3-2022Q3. 2022Q3. 6. While low-educated workers2 faced the most significant employment contraction during the crises, they recovered faster but barely enough to reach pre-pandemic levels. In the depth of the pandemic, employment among workers with primary education dropped by nearly ten percent compared to the pre-crisis level. In comparison, the same declined by only four percent for the average worker. During the recovery period, employment among this group grew faster but barely enough to reach pre- pandemic levels. In fact, their employment did not reach pre-pandemic levels until Q3 of 2022 (Figure 7). Contrarily, high-educated workers experienced employment gains after the pandemic, and their employment growth during the recovery was significantly higher. 7. Nevertheless, the employment recovery pace for low-educated workers varies significantly across subregions, with those in Central Eastern Europe lagging far behind. In NE, employment for workers with primary education already reached 5 percent above the pre-crisis level in Q3 of 2022, after a sharp decline during COVID-19. Yet, their peers in Central Eastern Europe still had their employment 11 level below 20 percent of their pre-crisis rate, slightly similar to their deepest employment contraction in Q1 of 2021 (Figure 8). This is partly due to the slow recovery of the accommodation and food service sector, which employs primarily low-educated workers. Moreover, employment of young individuals in this region is also below its pre-pandemic levels. Thus, the slow employment recovery of low-educated workers in Central Eastern Europe implies continuous earning losses for this vulnerable group. Figure 7 Employment growth index (2019Q3=100) by Figure 8 Employment growth index education level, EU27 (2019Q3=100) by subregion, Workers with primary education 115 110 Employment growth index (2019Q3=100) 110 105 Employment growth index (2019Q3=100 100 105 95 90 100 85 95 80 75 90 70 Primary Secondary Tertiary Total NE SE WE CEE Note: Employment for individuals between 15 and 64 Note: Employment for individuals between 15 years. Primary refers to less than primary, primary, and lower and 64 years in less than primary, primary, and lower secondary education; secondary refers to upper secondary and secondary education. Source: Eurostat (lfsq_egaed, post-secondary non-tertiary education; and tertiary to tertiary lfsq_agaed, and lfsq_igaed), 2019Q3-2022Q3. Germany is education. Source: Eurostat (lfsq_ergaed), 2019Q3-2022Q3. excluded due to lack of data for some periods 8. Employment levels of blue-collar workers, regardless of their skill3 level, have not returned to pre-pandemic levels; on the other hand, white-collar workers have recovered fully. Low-skill blue-collar workers experienced the largest employment decline post-pandemic, and despite experiencing significant improvements during the recovery, their employment levels are still below pre-pandemic. High-skill blue- collar workers experienced moderate employment drops but minimal gains during the recovery. As a result, none of these groups have seen their employment levels return to pre-pandemic levels. Contrarily, employment among white-collar workers has surpassed pre-pandemic levels, as they have experienced solid employment growth in the recovery period. 3 Reliable and detailed data on skills is hard to get, as a detailed measure of skills should include basic education and training. Here, we proxy skills by using the (ISCO) data on occupations, as commonly done for the EU, using Eurostat data. The classification aggregates occupation categories from the ISCO 88 classification into 4 groups (White-collar high-skilled (WCHS), White-collar low-skilled (WCLS), Blue-collar high-skilled (BCHS), Blue-collar low-skilled (WCLS). 12 9. The type of jobs available has evolved differently during Covid-19. Full-time employment is above pre-pandemic levels for both males and females, while part-time employment is below. At the pandemic's start, part-time relative to full-time employment increased more for males than females, mainly due to different policies aimed at decreasing mass layoffs at the beginning of the covid-19 shock. For that reason, firms decide to hire or retain workers in part-time schemes rather than in full-time jobs. During the recovery, the trend reversed, with full-time employment increasing more than part-time employment, particularly for females. Figure 9 Employment growth index (2019Q3=100), Figure 10 Type of employment by gender index by occupation, 2019Q3-2022Q3, EU27 (2019Q3=100), EU 27, 2019Q3-2022Q3 106 106 Employment growth index (2019Q3=100) Employment growth index (2019Q3=100) 104 104 102 100 102 98 100 ) 96 98 94 92 96 90 94 2019Q3 2019Q4 2020Q1 2020Q2 2020Q3 2020Q4 2021Q1 2021Q2 2021Q3 2021Q4 2022Q1 2022Q2 2022Q3 Low-skill blue collar High-skill blue collar Full-time male Part-time male Low-skill white collar High-skill white collar Full-time female Part-time female Note: Total employment for individuals between 20 and 64 Note: Part-time and full-time employment for individuals years. Source: Eurostat (lfsq_egais), 2019Q3-2022Q3. Low-skill between 15 and 64 years. Source: Eurostat (lfsq_epgais), blue collars (ISCO codes 8 and 9) include plant and machine 2019Q3-2022Q3. operators and assemblers and elementary occupations; High- skill blue collars (ISCO codes 6 and 7) include skilled agricultural and fishery workers and craft and related trades workers; Low-skill white collars (ISCO codes 4 and 5) include clerks and service workers and shop and market sales workers; High-skill white collars (ISCO codes 1,2 and 3) include legislators, senior officials and managers, professionals and technicians and associate professionals. It excludes employment in the armed forces and non-responses. 10. Labor Force participation is above pre-pandemic levels, including the one for females, but the gender gaps remain large, especially in SE and CEE. The gender-specific burden of the pandemic led to asymmetric shocks in the labor market of females. However, the recovery has been stronger for females in the latest periods than for males in all the country groups. For example, in the SE region, labor force participation among females have increased by 4.2 percentage points during the recovery period, compared to 3.3 percentage points among the male counterparts. In the CEE region, the increases have been 3.5 and 1.1 percentage points, respectively, for females and males. However, even if they are growing at a higher rate, the gender gap in participation is still very large for certain subregions. For 13 example, in SE and CEE, the difference is -13.5 percentage points and -12.4 percentage points, respectively. 11. Despite tighter labor markets, which are expected to lead to higher salaries, real wages declined significantly due to rising inflation. While nominal hourly labor costs4 have been increasing in most countries in the EU, real labor costs have been declining as a result of the rapid growth of inflation, especially since Q3 of 2021 (Figure 11). Bulgaria is the only country with slightly positive real wage growth between Q3 of 2022 and Q3 of 2021. For the rest of the EU, real wages declined significantly, with the most striking drop observed in Estonia (-15.9 percent), Latvia (-15.5 percent), and Czech Republic (-12.5 percent). Moreover, the gap between real and nominal labor costs has been widening in recent quarters as inflation continues to surge. While most EU countries have increased (or announced to increase) their statutory minimum wages to soften the impacts of inflation on low-wage workers, these increases do not fully offset the high inflation. As of August 2022, a declining real minimum wage occurred in all EU countries except Belgium, Hungary, Greece, Romania, and France (EC 2022). Figure 11 Percentage change in real labor cost index, EU countries, 2019Q3-2022Q3 15.0 10.0 Change in real labor costs (%) 5.0 0.0 EL EE LV LT ES LU NL BE EU27 IT IE AT FR SI DK FI SK RO DE SE BG CZ MT PL PT CY HR HU -5.0 -10.0 -15.0 -20.0 Change 2019Q3-2020Q3 Change 2020Q3-2021Q3 Change 2021Q3-2022Q3 Note: The real labor costs are constructed by deflating the labor cost index with the HICP index. This index assumes a consumer basket that is representative of the entire EU. Labor costs refer to wages and salaries in the sectors of services, industries, and construction. Source: Eurostat (lc_lci_r2_q and prc_hicp_midx), 2019Q3-2022Q3. 12. The average working hours of full-time workers dropped in the first year of the pandemic across all EU countries; more recently, working hours have not normalized across all countries. The labor market after COVID-19 adjusted, partly through changes in the intensive margin, leading to lower working hours among employees. In the more recent expansionary phase from 2020 to 2021, in some countries, firms responded to increasing demand by hiring more workers or asking for extra hours for existing ones. This translates into 4 Includes wages and salaries only. 14 growth in overall working hours. However, working hours have not come back to pre-pandemic levels in many countries. Figure 12 Percentage change in the working hours of full-time workers, 2019-2021 4 3 Change in working hou rs (full-time) 2 1 0 LU LT EE LV EL ES IT RO EU27 IE AT BE FR SK NL DK SI BG MT FI SE DE PL CY PT CZ HU HR -1 -2 -3 Change 2019-2020 Change 2020-2021 Note: Average working hours for full-time employees between 15 and 64 years. Source: Eurostat (tps00071), 2019- 2021. No more recent data is available. II. Welfare impacts of the cost-of-living crisis on European households in selected countries 13. This section assesses the potential welfare impact of rising inflation – aggravated by the global supply chain disruptions stemming from Russia’s invasion to Ukraine– on household income and living conditions in four EU Countries, namely Bulgaria, Croatia, Poland, and Romania.5 Rising food and energy prices in 2021-2022 left the poorest with less disposable income for other essential needs (e.g.education and health), and expose their vulnerability to further shocks. Understandably, their level of resilience against these exogenous shocks will depend on governments’ support measures and expected consumer prices in the coming months (or years). 14. In December 2022, consumer prices increased by 10.4 percent in EU27 compared to December 2021, according to the 12-month percentage change in the Eurostat harmonized consumer price index (HICP)6. The overall HICP is mostly driven by food and energy inflation that has been accelerating in the same period, reaching an increase of 17.8 percent and 37.2 percent, respectively. Across all four EU countries selected for this analysis, their overall HICPs were higher than the EU-27 average, ranging from 12.7 percent in Croatia to 15.3 percent in Bulgaria. Notably, energy inflation in these countries, especially Bulgaria and Croatia, has been substantially lower than the EU27 average (Figure 13) due to the 5 Where the World Bank has engagements through Country Partnership Frameworks. 6 The official monthly inflation rate for goods and services. 15 government’s energy price caps. However, the trend is the opposite for food inflation despite some efforts from the governments to reduce food prices7 (Figure 14). Figure 13 Evolution of energy prices (index) by Figure 14 Evolution of food prices (index) by countries, 2020M1-2022M12 countries, 2020M1-2022M12 180 140 170 135 130 160 125 150 Energy prices (index) Food prices (index) 120 140 115 130 110 120 105 110 100 100 95 90 90 BG HR PL RO EU27 BG HR PL RO EU27 Note: The index is constructed for electricity, Note: The index is constructed for food and gas, and other fuels. Source: Eurostat (prc_hicp_midx), non-alcoholic beverages. Source: Eurostat 2020M1-2022M12. (prc_hicp_midx), 2020M1-2022M12. 15. The inflation gap between wealthier and poorer households has been widening recently. Not only can inflation impacts be asymmetric across the income distribution, but the inflation pressures have also become much higher among poorer households, increasing the uneven losses from price increases. In particular, for December-2022, the difference in inflation between the bottom and top quintiles increased to 3.6 percentage points in Bulgaria, 3.1 percentage points in Romania, 1.9 percentage points in Croatia, and 1.4 percentage points in Poland (Figure 15). This implies that the recent inflationary pressures have disproportionately affected the poorest. Figure 15 Difference in the inflation of bottom and top income quintiles by countries, 2020M1-2022M12 7 For example, in Croatia, the government also introduced VAT reduction for certain food items. 16 5 4 Gap bottom vs. top inflation 3 2 1 0 -1 -2 BG HR PL RO Source: Bruegel (https://www.bruegel.org/dataset/inflation-inequality-european-union-and-its- drivers), 2020M1-2022M12. 16. Moreover, poorer households spend larger shares of their income on food items, making rising inflation harder to cope with. Households at the lowest end of the income distribution in the 4 EU countries spend more than half their total budget on food and energy - a much larger share than their peers in higher income brackets. For example, in Bulgaria, the share of household expenditure devoted to food is two times larger in the bottom (poorest) consumption decile than in the highest (richer) decile. The bottom decile spends nearly half its budget on food, while the upper decile spends slightly more than 20 percent. This means any change in food inflation can potentially lead to larger welfare losses among the poor or greater vulnerability to rising food prices. If a change in food prices results in a higher share of total household income spent on food, this can result in the household being more liquidity- constrained. Moreover, households could change their consumption pattern as a response, substituting food staples for cheaper and lower quality options and/or reducing consumption of other non-food expenditures and human capital investments in education and health. 17. Higher energy gas prices are likely to be particularly challenging, especially when, in some countries, a large share of the population could not keep their homes warm even prior to the energy crisis. About 23.5 percent of Bulgarian households cannot keep their homes adequately warm, and the share reached 42.6 percent among income-poor households (those with incomes below 60 percent of the median equivalized income). Additionally, high energy costs and/or low household income often force people affected by energy poverty to fall behind on the payments of their utility bills. In 2020, approximately 19.2 and 7.3 percent of Bulgarian and Romanian households could not pay utility bills (heating, electricity, gas, water, etc.) on time due to financial difficulties. This share was even higher in Croatia, reaching 15.2 percent. 18. Income poverty is expected to face significant downside risks in 2022 as the four EU countries grapple with the aftermaths of Russia’s invasion to Ukraine and the consequential supply chain disruptions and inflation. Inflation can affect households' monetary welfare through multiple channels. First, it directly reduces the purchasing power of households, which can lead to a decline in the regular consumption of goods and services – these are the direct effects of rising food and energy prices on households’ income. Second, the surge in energy and food prices can increase prices of other goods in the economy that use energy and food as inputs, hence raising the overall inflation rates. With rising costs of 17 materials, firms might also react by hiring fewer workers or stopping giving pay raises to cope. Finally, savings and investments can be affected, as higher interest rates can decrease the real value of assets over time. This analysis will address (i) the direct impacts of rising food and energy prices on households' income and (ii) the indirect impacts of rising food and energy prices on prices of other goods in the economy and their overall effects on households’ income. Compared to the direct impacts, the indirect impacts on households might take more time to materialize as energy-using companies in different sectors (e.g., transportation, agriculture, construction, etc.) need some time to adjust their operations and pass on higher energy prices to consumers. See Annex for Methodological Details. 19. When combining the impact of rising food and energy prices without government income support measures, poverty rates (6.85 USD per day in 2017 PPP) are expected to rise in the four countries, and indirect impacts account for most of the welfare losses. Microsimulation results show that the direct effects are welfare reducing in all four countries but generally contained, ranging from 0.2 to 1 percentage point in Poland and Bulgaria. This is understandable as the observed CPI of food and energy already reflects governments’ price caps put in place to shield consumers. Most of the poverty increases are due to the pass-through from energy prices to food and other products and services prices (indirect impacts), with poverty rates increasing by 0.3, 0.5, 1.7 and 1.8 percentage points in Poland, Croatia, Bulgaria, and Romania, respectively (Figure 16). These estimates do not account for additional income support measures (e.g. social transfers, one-time payment to pensioners, etc.), which are often temporal with limited generosity in most EU countries, but could potentially mitigate these losses, if well designed. As a result, people perceived these measures as insufficient (Box 1). Figure 16. Simulated changes in poverty due to food and energy inflation, Direct and Indirect Impacts, USD 6.85 poverty line (2017 PPP) 4 Percentage point poverty change 3 2 1.8 1.7 1.0 0.9 1 0.5 0.3 0.3 0.2 0 BG HR PO RO Simulated change in poverty (indirect effect) Simulated change in poverty (direct effect) Note: Simulated changes in poverty represent the difference between baseline poverty rates, constructed from household income per capita in EU-SILC 2020 and 2019 for Poland, and simulated poverty, resulting from microsimulations. For the simulation of direct impacts, we use inflation changes from October 2022 relative to October 2019 and assume that the price elasticity of demand is the same for food and energy, but the price elasticity differs across the income distribution. For the simulation of the indirect effect, we use detailed price changes as a result of the energy shock in a CGE model. Source: World Bank micro simulations based on Eurostat Harmonized Indices of Consumer Prices (HICP) and the latest consumption data available from the HBS and income data from the EU-SILC. 18 Box 1: Governments' responses to rising prices and households' perception Since early 2022, Bulgaria, Croatia, Poland, and Romania governments have swiftly put in place support schemes to shield consumers and businesses from rising prices, particularly energy prices. In Croatia, the government introduced a price cap on gas, oil derivatives, and electricity, a reduction in the value-added tax (VAT) rate on various food items and energy, in addition to subsidies and social transfers to firms and households. The Bulgarian authorities froze power and heating prices from December to March 2022 to reduce vulnerability in the winter months and provided temporal heating assistance to poor households. In Poland, the government implemented an "anti-inflation shield," which temporarily reduced VAT on food, gas, fertilizers, petrol, diesel, and heating. Moreover, the government strengthened Special Needs Allowance for housing-related costs and allowance to help households struggling with energy bills. Finally, the Romanian government imposed a temporary cap on electricity and natural gas prices and introduced a series of grants and vouchers (e.g., Heating Energy Allowance, Natural Gas Allowance, Solid Fuel or Oil Allowance) to help vulnerable Romanians and businesses. However, facing rising living costs, most people believed these policy responses were insufficient. In Romania and Croatia, where we conducted surveys after the governments announced support measures, more than 60 percent of respondents reported insufficient support. In Croatia, people in lower income brackets and rural areas were more likely to express dissatisfaction. However, in Romania, middle-income and urban residents were more affected. % of Romanian adults believing the % of Croatian adults believing the government's energy responses were insufficient government's energy responses were insufficient 70 70 60 60 % respondents 50 % respondents 50 40 30 40 20 30 10 20 - 10 - National Bottom Mid 40 Top 20 Rural 40 Source: Romania rapid survey, June 2022 Source: Croatia rapid survey, April 2022 20. Our microsimulations show asymmetric welfare impacts across the income distribution, with losses concentrated among lower-income deciles. While household income is expected to decrease among all income deciles due to the loss in purchasing power, the decrease is significantly higher among lower-income households (Figure 17). The largest decrease is observed in Bulgarian households at the bottom of the income distribution, followed by Romanian households. These results are mostly due to their consumption patterns: they spend higher budget shares on food and energy, and they have less capacity to substitute for other goods. 21. While inflation affects the poor more than the rich, the impact on inequality is smaller, partly due to the governments’ price caps. The simulated Gini shows that the Gini coefficient increases in Bulgaria and Romania by 0.8 and 0.5 percentage points, respectively, due to the direct effects of inflation, 19 while in Croatia and Poland increases by 0.2 percentage points. The effect is slightly higher when simulating the indirect effects in Croatia and Romania, yet in Bulgaria and Romania the indirect effect is much smaller (Figure 18). The higher pass-through from energy prices to food prices, than to other products and services prices, in Croatia and Romania explains that poorer households, who spend relatively more on food, are more affected and therefore inequality increases. On the other hand, in Bulgaria and Romania the energy prices translates similarly to food and other products and services prices, and as wealthier households spend a larger share on other products, the impact on inequality is less clear. Figure 17 Simulated changes in household disposable income, Figure 18 Simulated Changes in the Gini coefficient (direct by income deciles (direct Effects) effect) 0.00 2.0 % growth in household income due to -1.00 1 2 3 4 5 6 7 8 9 10 1.8 Percentage point Gini change -2.00 1.6 -3.00 1.4 -4.00 1.2 inflation -5.00 1.0 0.8 -6.00 0.8 -7.00 0.5 0.6 -8.00 0.3 0.4 0.3 0.2 0.2 -9.00 0.2 0.0 0.0 -10.00 0.0 Income deciles BG HR PL RO BG HR PL RO Simulated change in Gini (indirect effect) Simulated change in Gini (direct effect) Note: The income measure is per capita disposable income based on 2020 EU-SILC for all countries, and 2019 EU-SILC for Poland. For the Note: The income measure is income adult equivalent in EU-SILC direct impacts, we assume that the price elasticity of demand is the 2020 and 2019 for Poland. For the direct impacts, we assume that same for food and energy, but the price elasticity differs across the the price elasticity of demand is the same for food and energy, but income distribution. For the simulations, we use inflation changes from the price elasticity differs across the income distribution. We use October 2022 relative to October 2019. inflation changes from October 2022 relative to October 2019. For Source: World Bank micro simulations based on Eurostat Harmonised the indirect effect, we use the price pass-throughs from the CGE Indices of Consumer Prices (HICP) and the latest consumption data model. available from the HBS and income data from the EU-SILC. Source: World Bank micro simulations based on Eurostat Harmonised Indices of Consumer Prices (HICP) and the latest consumption data available from the HBS and income data from the EU-SILC. 22. However, some population groups – i.e., single-elderly households – are particularly susceptible to rising energy and food prices; rural households are also affected more in some countries. Single- elderly households experienced higher than average annual income losses (3.6 percent in Croatia, 4.5 percent in Romania, 7.4 percent in Romania, and 8.7 percent in Bulgaria). The worrisome news is that the pre-crisis poverty level among single elderly households was between 17 and 34 percentage points higher than national poverty rate, despite their access to pension income. Their retirement income sources tend to be fixed, giving less space to adjust to price increases. In Romania and Bulgaria, these households also have a relatively high energy share in overall expenditures. In Bulgaria, losses among rural households 20 are also higher, consistent with the energy consumption patterns. Moreover, job opportunities for individuals in rural areas or older adults without the required skills for the labor market are more scarce. Thus, policymakers need to consider tailored policies to alleviate the burden of inflation among these groups. Figure 19 Simulated changes in per capita Figure 20 Simulated changes in per capita disposable income, by rural and urban disposable income, by population subgroups 0.0 0.0 % growth in per capita income due to inflation -1.0 % growth in per capita income due to -1.0 -2.0 -2.0 -3.0 -3.0 -4.0 -2.8 -2.8 -2.9 -5.0 inflation -4.0 -4.0 -4.0 -4.0 -6.0 -5.0 -7.0 -6.5 -6.3 -6.0 -8.0 -6.9 -7.3 -7.7 -7.4 -7.5 -9.0 -7.0 -6.5 -6.3 -8.7 -6.8 -10.0 -7.3 -7.0 -8.0 -7.8 -9.0 All Urban Rural BG HR PO RO BG HR PO RO Note: The income measure is per capita disposable income in EU-SILC 2020 and 2019 for Poland. For the simulation of direct impacts, We use inflation changes from October 2022 relative to October 2019 and assume that the price elasticity of demand is the same for food and energy. Still, the price elasticity differs across the income distribution. Source: World Bank micro simulations based on Eurostat Harmonised Indices of Consumer Prices (HICP) and the latest consumption data available from the HBS and income data from the EU-SILC. 23. A complementary analysis using a decile-specific CPI shows that the welfare impacts can be even higher8. As a result of inflationary pressures, poverty (6.85 USD per day) is expected to increase from 1.2 percentage points in Croatia to 4.1 percentage points in Romania. Poland and Bulgaria can experience changes of 1.3 and 3.3 percentage points, respectively. In terms of inequality, the effect is slightly higher than the one estimated in method 1 (direct effects), with the Gini coefficient increasing by up to 0.8 percentage points in Bulgaria and 0.2 percentage points in Croatia. 24. Going beyond monetary poverty, rising prices of essential goods and declining real wages affect affordability and alter households' consumption patterns toward buying fewer and cheaper products. At least 70 percent of European consumers reported trying to reduce their bills by changing their purchasing behavior. As of June 2022, 86 percent of Romanian households bought fewer goods in general, and another 85 percent of households would sacrifice other purchases to cover the rising costs of essential goods (Figure 21). For Croatian households, these figures are 79 percent and 78 percent, respectively. 8 Overall estimating the poverty and distributional impact of inflation is challenging given the multiple channels of impact and the limited availability of real time data. Therefore, we adopt multiple approaches to get a range of estimates. 21 With inflation front of people's minds, European consumers also changed how they shopped: about 75 percent of Romanian and Croatian consumers seek money-saving tactics such as buying cheaper brands or products on sale. Moreover, shifts in consumer purchasing patterns could signal changes in production and retailers' marketing strategy in the upcoming shopping season. Figure 21: Changes in consumption patterns in response to rising prices 53% 33% Romania Sacrifice other purchases to cover rising prices… 56% 29% 45% 29% Buy brands on sale/promotion 42% 36% 16% 63% Croatia Sacrifice other purchases to cover rising prices… 19% 59% 16% 58% Buy brands on sale/promotion 17% 52% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% Strongly agree Agree Source: Croatia rapid surveys in April 2022, Romania rapid survey in June 2022. 25. As the rising cost of living outpaces economic and labor recovery, more households – especially the poorest - face difficulties making ends meet compared to their pre-crisis level. In all four countries, summer 2022 is marked with an increasing proportion of households reporting that it is harder to meet ends than their pre-crisis level (Figure 23). In Romania, nearly twice as many households reported so between spring 2021 and summer 2022 at 37 percent and 63 percent, respectively. A comparable increasing trend can be found in Croatia (from 33 percent to 58 percent), Bulgaria (from 33 percent to 46 percent), and Poland (from 48 percent in Spring 2021 to 55 percent in Fall 2021). The unfortunate reality is that the poorest households struggled even more: in spring 2022, seven in ten Croatian households at the bottom 40 percent of the income distribution declared greater difficulties in making ends meet. This figure was the same for Romanian households in the summer of 2022. Figure 22: Chief areas of concern among Bulgarian Figure 23: Harder to make ends meet compared to households, April 2022 pre-crisis level Energy prices 98% 37 Romania 47 54 63 Food prices 97% 48 Poland 55 Economic situation in Bulgaria 95% Political situation in Bulgaria 90% 33 Croatia 38 58 Health 86% 32 Bulgaria 46 Covid restrictions 60% Job situation 50% 0 20 40 60 0% 20% 40% 60% 80% 100% Spring 2021 Fall 2021 Spring 2022 Summer 2022 Source: Bulgaria rapid survey, April 2022 Source: Rapid surveys across countries 22 26. Furthermore, challenges in households' real purchasing power will lead to an uptick in defaults on obligations in 2022. In parallel with rising costs of living, the proportion of households reporting arrears in paying utility bills and phone bills rose in recent months, about twice the rate for the bottom 40 percent as the general population, revealing the heterogeneity in households' struggles across the income distribution (Figure 24 and Figure 25). In addition, the share of households in arrears is likely to rise further if the temporary government support schemes currently in place do not extend to the coming year. Figure 24: Croatian households reporting arrears in Figure 25: Croatian households reporting arrears utility bills over time in phone bills over time 23 25 21 25 22 20 17 15 20 15 13 15 14 10 10 13 8 12 11 10 11 5 5 8 8 6 - Dec 2020 Mar 2021 Jun 2021 Sep 2021 Apr 2022 0 Dec 2020 Mar 2021 Jun 2021 Sep 2021 Apr 2022 Average Income bottom 40 Average Bottom 40 income Source: Croatia rapid surveys 2020-2022 Source: Croatia rapid surveys 2020-2022 27. Although at the aggregate level, economic forecasts suggest positive economic growth and slower inflation in 20239, households still express a bleak prospect for their financial situation and expect persistent struggle in the coming year. In Romania, the spring and summer of 2022 mark an uptick in the share of households expecting a worsening financial situation in the next 12 months. In mid-2022, 54 percent of Romanian households reported a significant jump from 40 percent in early 2021 (Figure 26). In fact, by mid-2022, about half of households across Bulgaria, Croatia, and Romania expressed a pessimistic view of their financial prospects (Figure 27). Moreover, a strong economic projection does not reflect in households' expected spending: at least half of the residents in Bulgaria, Croatia, and Romania expected less spending on durable goods and restaurants in the coming year (Figures 16 and 17). Alarmingly, pessimism is much higher among households in the bottom 40 percent of the income distribution signaling persistent effects of the crisis on poor households. Figure 26: Romanian households' financial prospects Figure 27: Share of households reporting over time worsening financial situation in the next 12 months, by country 9 World Bank Macro-Poverty Outlook 2023. 23 60 70 62 59 Shuare of households 50 55 Share of households 60 54 40 27 27 48 49 25 50 30 25 24 21 22 40 20 10 20 20 27 27 30 19 16 16 0 20 10 0 Bulgaria Croatia Romania Much worse Worse National Bottom 40 percent Source: Romania rapid surveys 2020-2022 Source: Bulgaria and Croatia rapid surveys in April 2022, Romania rapid survey in June 2022. Figure 28: Share of households expecting less Figure 29: Share of households expecting less spending on durable goods in the next 12 months, by country spending on restaurants in the next 12 months, by country 80 73 73 80 75 69 70 70 70 63 60 60 60 55 50 48 % households 50 % households 50 42 40 40 30 30 20 20 10 10 - 0 Bulgaria Croatia Romania Bulgaria Croatia Romania National Bottom 40 National Bottom 40 Source: Bulgaria and Croatia rapid surveys Source: Bulgaria and Croatia rapid in April 2022, Romania rapid survey in June 2022. surveys in April 2022, Romania rapid survey in June 2022. 28. Social safety nets are vital in protecting households from various risks; yet, in most countries, these programs are not sufficient to support the poorest and most vulnerable. In Romania, social spending is low, inefficient, and increasingly skewed toward pensions, making it less effective at reaching the neediest. Social transfers are also characterized by a prevalence of categorical programs, rendering them ineffective at reducing poverty. In Bulgaria, the fiscal system is characterized by limited progressivity. Direct transfer programs are progressive to varying degrees but are small in size. Between 21 and 77 percent of benefits accrue to the poorest quintile depending on the programs, though there is scope for improving leakage to the wealthiest households. Croatia's current social safety net is designed to mostly support categorical groups such as persons with disabilities and families with children. Meanwhile, poverty-targeted social assistance programs (i.e., Guaranteed Minimum Benefits and child 24 allowance) are small, suggesting that the safety net is not as progressive as it should be. Poland's social protection spending has a disproportionate focus on supporting families, with a significantly higher percentage of GDP and social protection expenditure being allocated to family benefits than the EU average (3.8 and 16.2 percent vs. 2.5 and 8.3 percent, respectively). However, there is a lack of adequate spending on social exclusion and housing, which is lower than the EU average. Despite well-targeted minimum-income programs for low-income households, only half of those living in legal poverty were reached in 2019. Additionally, there are gaps in social assistance coverage for the working poor who are not covered by targeted transfers.10 III. Policies toward a more resilient Europe 29. In the current context, countries should prioritize protecting the vulnerable through targeted support while keeping a tight fiscal stance to help fight inflation. In most EU countries, support measures aiming to shield consumers (individuals and firms) from rising food and energy inflation have been price caps with some income support measures. The general cap on energy prices and indirect utility subsidies can be regressive, or significantly less progressive than well-targeted social transfers, and pose a heavy burden on the fiscal budget. 30. Thus, governments must implement sound fiscal policies with several policy objectives, though trade-offs must be carefully assessed. These include improving the efficiency and equity of public spending, improving revenue measures, addressing the pension system's sustainability, eliminating quasi- fiscal deficits (especially in energy), and strengthening public debt management and public financial management. 31. Supporting the building blocks of a more efficient social protection system to deliver aid to the poorest households is critical, and even more so in an environment of rising food and energy prices. As discussed above, rising food and energy prices can adversely impact the poorest households, mostly from indirect effects. Therefore, social transfers need to be targeted at vulnerable groups. As governments may have competing demands and limited fiscal space, identifying and prioritizing vulnerable groups is important. The overall effectiveness of social programs can also be improved: higher spending on poverty- targeted programs, better targeting, and leakage reduction. 32. An often overlooked aspect of policy design needs to be addressed: how benefit levels change over time in response to increasing living costs; indexing government benefits and tax credits meant to provide relief is critical for optimizing their antipoverty effects. The value of several government benefits and tax credits meant to relieve struggling families is not keeping up because they were not designed to increase with inflation. As a result, the effectiveness of these benefit programs and tax credits is undermined, increasing the household risk of falling into poverty. 33. Minimum wage adequacy is also essential, as well as close monitoring of the ratio of the minimum wage to the cost of the basic food baskets is critical. As mentioned before, statutory minimum wages have decreased in real terms in most EU countries. To keep minimum wage adequacy, there is a need to establish clear criteria to set and update statutory minimum wages, assess their adequacy, establish rules to update them regularly, and ensure the timely and effective involvement of social partners. Moreover, in all Member States, the promotion of collective bargaining on wage setting (as well 10 Source: Romania, Croatia, Bulgaria and Poland CEQ Assessments. 25 as the coverage thereof) can also contribute to adequate minimum wage protection (European Commission, 2022). The compositional and distributional effects of rising minimum wages in real terms need to be carefully assessed. 26 References Badiani, Reena and Eva Militaru “The Distributional Impacts of Taxes and Social Spending in Romania – 2018 model”, unpublished manuscript. Bańkowski, K., Bouabdallah, O., Checherita-Westphal, C., Freier, M., Jacquinot, P., & Muggenthaler, P. (2023, February 13). Fiscal policy and high inflation. European Central Bank. Retrieved on March 9, 2023, from https://www.ecb.europa.eu/pub/economic- bulletin/articles/2023/html/ecb.ebart202302_01~2bd46eff8f.en.html#toc11 Claeys, G., L. Guetta-Jeanrenaud, C. McCaffrey, L. Welslau (2022) 'Inflation inequality in the European Union and its drivers,' Bruegel Datasets, first published October 26 European Commission, Directorate-General for Employment, Social Affairs and Inclusion, 2022, "Labour Market and Wage Developments in Europe, Annual Review" Freund, Caroline L., and Christine I. Wallich. "The welfare effects of raising household energy prices in Poland." The Energy Journal 17.1 (1996). Menyhert, B. (2022). The effect of rising energy and consumer prices on household finances, poverty, and social exclusion in the EU. Publications Office of the European Union, Luxembourg. Nguyen, T.V.N. & Rubil, I. (2021). The Distributional Impact of Taxes and Social Spending in Croatia. Sgaravatti, G., Tagliapietra, S., Trasi, C., & Zachmann, G. (2023, March 8). National fiscal policy responses to the energy crisis. Bruegel. Retrieved March 9, 2023, from https://www.bruegel.org/dataset/national- policies-shield-consumers-rising-energy-prices Vaughan, K. N., & Cabrera, M. V. (2022). The Distributional Impact of Taxes and Social Spending in Bulgaria with an Application to Green Fiscal Policies. 27 Annex: Methodology Overall, estimating the poverty and distributional impact of food and energy prices is challenging due to the multiple channels; our microsimulation focuses on the expenditure channel and the income channel. We developed several microsimulation models to estimate the quantitative welfare impacts of food and energy prices on households across countries in the EU region. We described below the results of two methodologies to measure poverty and the distributional impacts of rising food and energy prices in the 4 EU countries. When looking at the welfare impacts of inflation, we focus on the expenditure/consumption channel (losses in purchasing power) and the income channel. More generally, we look at the passthrough from food and energy prices to core and overall inflation, which reduces the real value of wages, government, and private transfers, as well as savings, thus affecting household welfare indirectly. The asset channel is not modeled. We simulate the welfare impact of food and energy inflation on households by accounting for direct and indirect/secondary impacts. First, we estimate the direct welfare losses of changes in energy and food prices by approximating purchasing power parity PPP losses (Freund and Wallich, 1997) over 2019- 2022. The relatively higher price of energy and food will constrain some of that consumer's budget and household income, which cannot be spent on other goods and services. This method considers that poorer and wealthier households respond differently to price increases (different food and price elasticities) and spend different shares of their household budget on these items. Second, we estimate the indirect welfare effects of changes in food and energy prices, which also translate into generalized inflation. Under this method, we first estimate a decile-specific counterfactual CPI using the information on the household budget shares (which vary by decile), and the CPIs for the specific groups11. Then, we calculate the change in this counterfactual CPI that occur with changes in the energy and food CPI. The difference between these two CPIs is interpreted as the inflation that occurs due to changes in food and energy components. Then, we simulate the impact of the counterfactual CPI inflation (that results from the increases in energy and food prices) on poverty and welfare. We do so by deflating the income aggregate with these counterfactual CPIs (see more details below). Methodology 1: Measuring Direct and Indirect Welfare Impacts of Food and Energy Inflation Direct Welfare Impacts To estimate the welfare impact of relative changes in food and energy prices, following Freund and Wallich (1995), we adjust the welfare aggregate to account for the loss in purchasing power (PPP loss) because households spend a larger share of the total expenditure on food and energy. Following the methodology used by Freund and Wallich (1995) and the Balancing Act (2013), the impact of the change in prices on the share of consumer surplus as a percentage of total household expenditures can be calculated as: (1) ΔCS/E = (S0 (P1 − P0) / P0) (ε + ε(P1 − P0) / P0 +1) 11 The CPI assume all households have the same reference basket. Hence, we compute a decile- level CPI, as the consumption baskets vary by decile. 28 where S0 is the initial budget share before the price change, and ε is the price elasticity of demand, and P1 and Po represent the initial and final relative prices. Notice that when ε is zero, this estimate of the direct welfare impact implicitly assumes that households do not substitute away from electricity, so it should be interpreted as either an estimate of the short- run impact (i.e., before households can adjust electricity consumption for other sources of energy) or as an upper bound of the long-run estimate. We analyze alternative scenarios (under different ε) to see the sensitivity of the results to this parameter. In particular, we made different assumptions on how elastic the demand for food is to price changes based on the available literature. The ability of households to substitute for other goods will be captured by different price elasticities, which will vary across the income distribution (as the poor usually have less ability to substitute). Table 1 Assumption of price elasticities by income deciles Income deciles Price elasticity of food and energy 1-3 -0.5 4-7 -1 8-10 -1.5 We implement this approach using the available household budget surveys and EUSILC surveys in the 4 EU countries. First, we estimate the food and energy shares using the country’s household budget surveys (HBS) by decile and impute them to each household in the EU-SILC by assuming a one-to-one relationship between consumption and income deciles. To do so, we assume the same share across households in the same decile. Then, we estimate the change in consumer surplus for each household in the EU-SILC using (1), given the estimated shares, (b) a range of price elasticities (decile dependent), and (3) the relative food and energy prices from Eurostat Harmonized Indices of Consumer Prices (HICP) over the period 2019- 2022. Finally, we use the PPP losses to estimate one counterfactual income distribution for the observed change in food prices and another for the observed change in energy prices. Some caveats apply. It is important to note that this approach would not allow one to account explicitly for the fact that households can be both producers and consumers of food. Also, it does not incorporate indirect and general equilibrium effects. For this, we use a CGE Model as detailed below. Indirect Welfare Effects The passthrough from food and energy prices to the core and overall inflation also reduces wages and non-labor income in real terms, thus negatively affecting household welfare. One way to estimate the indirect impacts of food and energy prices on households is by using input- output matrices if updated and available. This may be particularly important in the case of energy prices, which affect households not only directly but also through the household’s consumption of other products that use energy as inputs. Unfortunately, this is not the case in the four EU countries. For example, the latest input-output matrix for Romania is for 2020, while for Bulgaria, the IO matrix is quite outdated (2014). Therefore, for Bulgaria, we would have to make very strong assumptions: (i) using the input-output matrices of countries with similar economic structures - this is a very strong assumption as these matrices are very idiosyncratic, and not easy to extrapolate from one country to another; (ii) assuming the different 29 sectors' interdependencies have not changed and use the old input-output matrices. For this reason, in this analysis we did not use this method. We use the CGE model currently available for the 4 EU countries to estimate indirect effects of rising energy prices. Computable General Equilibrium (CGE) models can further aid in the assessment of the degree of pass-through from energy price increases to other sectors. We abstract from indirect effects of food inflation, as higher energy inevitably translate into higher production costs, while the effects from food tend to be of smaller magnitude. An energy price shock is introduced in the CGE model, and the model produces detailed counterfactual inflation estimates for 56 sectors as a result of the energy shock, based on the GTAP classification. Then, we map the GTAP codes to the COICOP codes in the household surveys12, and aggregate the price changes using the corresponding household budget shares (which are decile specific). This counterfactual CPI captures the change in overall inflation as a result of changes in energy prices. Then, we simulate the indirect welfare impacts of this inflation in the microsimulation model. Finally, to test how reasonable are our assumptions, we compare the observed CPI inflation with the counterfactual inflation produced by the CGE model. The magnitudes fall within a reasonable range. Methodology 2: Measuring Overall Impacts of Overall Inflation using Weighted Adjusted CPIs Inflation is usually calculated using overall price indices that are based on groups of goods and services purchased or consumed by the average person (the reference basket). Inflation is typically measured using aggregate price indices such as the Producer Price Index (PPI) or Consumer Price Index (CPI). While informative, these indices are based on aggregate bundles of goods and services sold or consumed by the “median” firm or household. However, the budget allocation for households varies widely among different income and demographic groups. Therefore, understanding the impact of inflation on household choices and well-being requires considering how changes in consumer prices affect different households differently. In this section, we construct a decile-specific counterfactual CPI and adjust it with the latest changes in food and energy prices (weighted by the respective shares); then, we look at the impacts of these counterfactual inflation on household welfare. We construct income-decile-specific price indices that capture heterogeneity in consumption bundles to investigate the inflation experience of different income groups in Romania and shed light on the role of varying inflation rates on poverty and inequality. We then used it to deflate the income aggregate and estimate poverty and inequality impacts. We construct our CPI using consumption shares for three types of products and differences in prices for these same three types of products. These are food and beverages, energy, and the rest (all excluding 12 Some caveats apply, as there is no one-to-one mapping from GTAP product categories in the CGE model and the COICOP consumption categories in the household surveys. Therefore, some assumptions are made to produce a unique mapping. We assume first that the price changes and budget shares in sectors are the same as for households. Next, we assume that the sectors of grains and crops, livestock and meat products, and processed food belong to food and non-alcoholic beverages in COICOP. And that the sectors of textiles and clothing, transport and communication, and other services belong to other products and services (which refer to the rest of COICOP categories excluding electricy, gas and other fuels consumption). 30 food and energy). We distinguish between these three products to better capture the larger consumption shares by the poor for food, beverages, and energy. The equation below details the baseline CPI: = ∑ ∑ ℎ ∗ _ =0 =0 , where c is country C (Bulgaria, Croatia, Poland, Romania), d are the deciles, and i are the three classes of goods (food and beverages, energy, and all other goods). ℎ refers to the consumption share of that income decile for the type of goods i in country C in 2018. _ is the ratio of the CPI for the type of goods i in country C for the years 2019 versus 2016. We use to transform income reported for 2019 in 2019 currencies to 2016 currencies in our baseline scenario. For the simulated scenario with higher inflation, we simply replace this value with the following: ̇ = ∑ ∑ ℎ ∗ _ + ℎ ∗ _ =0 =0 _ is inflation during the period 2019 to October2022 (2018 to October, 2022, in the case of Poland). Then, we estimate the impact of overall inflation on poverty. We then use ̇ to deflate household incomes from the EU-SILC. We consider two different poverty measures: international poverty (at the daily 6.85 2017-PPP-US-Dollar line) and the 2016 anchored at-risk-of-poverty line (from now on: arop2016). In the former case, we convert daily income per capita reported in the 2020 EU-SILC data to 2017-PPP-US-Dollar levels, using the counterfactual CPI. The poverty increase due to increased prices is the difference between the baseline poverty rate and the simulated poverty rate. Our exercise relies on several restrictive assumptions. First, we abstract from income increases based on parallel price increases (e.g., those producing food and beverages might benefit from a share of the witnessed price increases for these products). We also abstract from employers compensating employees for increased inflation by adjusting wages or governments introducing public transfers for those most affected by the increases. In addition, we assume that there are no behavioral adjustments to the witnessed increase in the food basket. The consumption share of different types of products remains constant. In our analysis, consumers do not replace more expensive products with cheaper ones (e.g., shopping mainly in discounters) or stop purchasing certain goods (e.g., replacing meat with eggs). Finally, we do not incorporate economic growth considerations into the microsimulations, as this exercise aims not to produce poverty projections but to estimate the impact of the recent food and energy inflation on poverty and inequality. 31