Energy poverty and deprivation in Ireland

Barrett, Michelle; Farrell, Niall; Roantree, Barra Research Report Energy poverty and deprivation in Ireland Research Series, No. 144 Provided in Cooperation with: The Economic and Social Research Institute (ESRI), Dublin Suggested Citation: Barrett, Michelle; Farrell, Niall; Roantree, Barra (2022) : Energy poverty and deprivation in Ireland, Research Series, No. 144, The Economic and Social Research Institute (ESRI), Dublin, https://doi.org/10.26504/rs144 This Version is available at: https://hdl.handle.net/10419/268076 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ ENERGY POVERTY AND DEPRIVATION IN IRELAND MICHELLE BARRETT, NIALL FARRELL AND BARRA ROANTREE RESEARCH SERIES NUMBER 144 JUNE 2022 E V I D E N C E F O R P O L I C Y ENERGY POVERTY AND DEPRIVATION IN IRELAND Michelle Barrett Niall Farrell Barra Roantree June 2022 RESEARCH SERIES NUMBER 144 Available to download from www.esri.ie  The Economic and Social Research Institute Whitaker Square, Sir John Rogerson’s Quay, Dublin 2 https://doi.org/10.26504/rs144 This Open Access work is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. ABOUT THE ESRI The mission of the Economic and Social Research Institute is to advance evidencebased policymaking that supports economic sustainability and social progress in Ireland. ESRI researchers apply the highest standards of academic excellence to challenges facing policymakers, focusing on 12 areas of critical importance to 21st Century Ireland. The Institute was founded in 1960 by a group of senior civil servants led by Dr T.K. Whitaker, who identified the need for independent and in-depth research analysis to provide a robust evidence base for policymaking in Ireland. Since then, the Institute has remained committed to independent research, and its work is free of any expressed ideology or political position. The Institute publishes all research reaching the appropriate academic standard, irrespective of its findings or who funds the research. The quality of its research output is guaranteed by a rigorous peer review process. ESRI researchers are experts in their fields and are committed to producing work that meets the highest academic standards and practices. The work of the Institute is disseminated widely in books, journal articles and reports. ESRI publications are available to download, free of charge, from its website. Additionally, ESRI staff communicate research findings at regular conferences and seminars. The ESRI is a company limited by guarantee, answerable to its members and governed by a Council, comprising 14 members who represent a cross-section of ESRI members from academia, civil services, state agencies, businesses and civil society. The Institute receives an annual grant-in-aid from the Department of Public Expenditure and Reform to support the scientific and public interest elements of the Institute’s activities; the grant accounted for an average of 30 per cent of the Institute’s income over the lifetime of the last Research Strategy. The remaining funding comes from research programmes supported by government departments and agencies, public bodies and competitive research programmes. Further information is available at www.esri.ie. THE AUTHORS Michelle Barrett is a Research Assistant at the Economic and Social Research Institute. Niall Farrell is a Senior Research Officer at the Economic and Social Research Institute and an Adjunct Associate Professor at Trinity College Dublin. Barra Roantree is a Research Officer at the Economic and Social Research Institute and an Adjunct Assistant Professor at Trinity College Dublin. ACKNOWLEDGEMENTS We are grateful to the Central Statistics Office (CSO) for facilitating access to the Survey of Income and Living Conditions (SILC) Research Microdata File used to construct the database for the SWITCH tax-benefit model, and to the Irish Social Science Data Archive for facilitating access to the Household Budget Survey (HBS). Results are based on eSWITCH version 4.5. This work was carried out with funding from the Community Foundation for Ireland, which is gratefully acknowledged. The analysis builds on previous work carried out under the ESRI’s Tax, Welfare and Pensions and Poverty Inequality and Living Standards Research Programmes. We are grateful to colleagues at the ESRI for their helpful suggestions, as well as to three anonymous referees for their comments on an earlier draft of this paper. All views and any errors and omissions remain the sole responsibility of the authors. This report has been accepted for publication by the Institute, which does not itself take institutional policy positions. The report has been peer-reviewed prior to publication. The authors are solely responsible for the content and the views expressed. FOREWORD Energy poverty is an equality issue. As fuel bills go up, it is people and families on lower incomes that suffer the most. In rapidly increasing numbers, households are facing the choice between putting food on the table, buying back-to-school clothes or heating their home. The increases in bills are already alarming. The potential for further increases risks creating a sense of desperation which requires assurances from policymakers that action will be taken before winter arrives. The challenge fuel poverty represents has been set out starkly in the findings of this report; up to 43 per cent of households could be at risk if energy price hikes continue and bills increase by a further 25 per cent. This is more than double the previous fuel poverty record in the early 1990s. The price increases that have already occurred and those yet to come are going to have an enormous impact. Cost increases per average household of €21.27 per week, or €38.63 when motor fuel is included, since January 2021 are already having an impact. This report underlines that households with lower incomes spend a much larger share of their income on fuel. That is where the need is greatest and where support must be targeted. Many of the 5,000 voluntary, community and charitable groups we work with will be looking at this report and no doubt will reflect on it as they make pre-Budget submissions to government. The options assessed by the ESRI to assist those who will be most affected by energy inflation need urgent government attention. Cutting indirect taxes does not deliver the response required – it also, and the report is clear, blunts the incentive to reduce the use of fossil fuels. The arguments for lump-sum payments to recipients of welfare payments, increases in the Fuel Allowance and increasing PRSI credits are compelling. This report would not have been possible without the support of anonymous donors to The Community Foundation for Ireland. We acknowledge their generosity in ensuring this timely and important research. Denise Charlton, Chief Executive, The Community Foundation for Ireland Table of contents | iii TABLE OF CONTENTS EXECUTIVE SUMMARY .......................................................................................................................... VII CHAPTER 1 INTRODUCTION .................................................................................................................. 1 CHAPTER 2 HISTORICAL TRENDS IN ENERGY POVERTY AND DEPRIVATION ......................................... 3 2.1 Measuring energy poverty and deprivation ............................................................ 3 2.2 Data and methodology............................................................................................. 4 2.3 Results ...................................................................................................................... 6 CHAPTER 3 IMPACT OF RISING ENERGY PRICES ON HOUSEHOLDS .................................................... 15 3.1 Modelling approach ............................................................................................... 16 3.2 Impact of recent energy price increases ................................................................ 17 3.3 Impact of potential future energy price increases ................................................ 19 CHAPTER 4 POLICY OPTIONS ............................................................................................................... 21 4.1 Indirect tax cuts ...................................................................................................... 21 4.2 Direct tax and welfare changes .............................................................................. 23 CHAPTER 5 CONCLUSION .................................................................................................................... 27 REFERENCES .......................................................................................................................................... 31 APPENDIX .............................................................................................................................................. 33 A. Imputation of energy expenditure into SWITCH SILC data .................................... 33 B. Additional figures and tables ................................................................................. 36 iv | Energy poverty and deprivation in Ireland LIST OF TABLES Table 3.1 Simulated impact of recent energy price increases on energy poverty ........................ 19 Table 3.2 Simulated impact of potential energy price increases on energy poverty .................... 20 Table B.1 Share of households in arrears on utility bill in population groups............................... 40 Table B.2 Estimates from probit regression model of fuel deprivation (for those not at risk of poverty) .......................................................................................................................... 41 Table B.3 Simulated impact of recent energy price increases, by household type ....................... 42 Table B.4 Simulated impact of potential future energy price increases, by household type ....... 43 Table B.5 Simulated impact of indirect tax measures (€ pw), by household type ........................ 44 Table B.6 Simulated impact of indirect tax measures (%), by household type ............................. 45 Table B.7 Simulated impact of welfare measures (€ pw), by household type .............................. 46 Table B.8 Simulated impact of welfare measures (%), by household type ................................... 47 Table B.9 Simulated impact of direct tax measures (€ pw), by household type ........................... 48 Table B.10 Simulated impact of welfare measures (%), by household type ................................... 49 Table B.11 Estimate of HICP inflation between Jan 2021 and Apr 2022, by group ......................... 50 LIST OF FIGURES Figure 2.1 Trajectory of energy poverty/deprivation in Ireland (1994–2020) ................................. 7 Figure 2.2 Rates of energy poverty and deprivation, by at-risk-of-poverty status .......................... 8 Figure 2.3 Composition of those experiencing energy poverty and deprivation, by whether atrisk-of-poverty ................................................................................................................. 9 Figure 2.4 Rates of energy poverty/deprivation by dwelling type ................................................. 11 Figure 2.5 Rates of energy poverty/deprivation by household tenure .......................................... 12 Figure 2.6 Energy poverty by fuel type ........................................................................................... 14 Figure 3.1 Change in selected CPI sub-indices (2015=100) ............................................................ 15 Figure 3.2 Simulated impact of recent energy price increases, by household type ....................... 18 Figure 3.3 Impact of simulated energy price increases, by household type .................................. 19 Figure 4.1 Simulated impact of indirect tax measures, by household type ................................... 22 Figure 4.2 Simulated impact of welfare measures, by household type ......................................... 24 Figure 4.3 Simulated impact of direct tax measures, by household type ...................................... 25 Figure B.1 Composition of energy poverty/deprivation by dwelling type ..................................... 36 Figure B.2 Composition of energy poverty/deprivation by household tenure .............................. 37 Figure B.3 Composition of energy deprivation by dwelling condition ........................................... 38 Figure B.4 Rates of energy deprivation by dwelling condition ....................................................... 38 Figure B.5 Components of self-reported energy deprivation (1994–2020) ................................... 39 Figure B.6 Change in selected CPI sub-indices 2003–2021 (2015=100) ......................................... 39 Abbreviations | v ABBREVIATIONS CSO Central Statistics Office ESRI Economic and Social Research Institute HBS Household Budget Survey LIIS Living in Ireland Survey SILC Survey of Income and Living Conditions SEAI Sustainable Energy Authority of Ireland 2 | Energy poverty and deprivation in Ireland This provides qualitative insight into the nature of energy poverty and deprivation in Ireland. The second contribution of this paper is to explore the effect that recent price increases have had on households and rates of expenditure-based energy poverty. Analyses such as this are often constrained by data availability; many data releases are intermittent and cannot capture the pace of energy price changes since summer 2021. We employ an imputation procedure to investigate the effect of recent price changes and future price increases on households accounting for changes in income and household composition (e.g. the rise in employment rates). The third and final contribution of this report is to consider the potential policy options to mitigate the impact of rising energy prices on households. We examine cuts to indirect taxes on energy, before turning to look at cuts to direct taxes on personal income and increases to social transfers. The structure of this report is as follows. Chapter 2 compares measures of selfreported energy deprivation and expenditure-based energy poverty over the past three decades. We employ two data sources – the Irish Household Budget Survey and the Survey of Income and Living Conditions – to compare rates of energy poverty to rates of energy deprivation. This comparison gives qualitative insight into the nature of energy poverty in Ireland. Chapter 3 examines the impact of recent changes in energy prices on households and measures of energy poverty, while Chapter 4 assesses options that policymakers might consider in trying to mitigate the impact of these rising energy prices on households. Chapter 5 concludes with consideration of the implications these developments have for policy. Historical trends in energy poverty and deprivation | 3 CHAPTER 2 Historical trends in energy poverty and deprivation This chapter will consider the incidence of energy poverty and deprivation among Irish households since 1994. The 2016 Strategy to Combat Energy Poverty3 defined energy poverty as the inability to heat or power a home to an adequate degree. While covering expenditures on heat and other energy services in the home, many studies in this field focus on the ability of a household to keep their home adequately warm (Boardman, 1991; Healy and Clinch, 2004; Hills, 2012; O’Meara, 2015). This chapter compares two contrasting measurement approaches: expenditure-based energy poverty and self-reported energy deprivation. Through this comparison, we provide a greater understanding of the qualitative nature of energy poverty in Ireland. We shed light on the relative contribution that heating and non-heating expenditures make towards energy poverty in Ireland. There are many ways in which households may respond to disproportionately high energy expenditures. Some households may go without energy services if they cannot afford the costs. Self-reported energy deprivation captures these effects by asking survey respondents whether they had been deprived of certain energy services during the period of analysis. Alternatively, constrained households may absorb the high cost, perhaps cutting back instead on other expenditures. Expenditure-based metrics are better able to capture this phenomenon. It is likely that both behaviours happen in varying degrees. By comparing trends associated with both metrics, insight into the factors contributing to their prevalence may be provided. This chapter proceeds as follows. The following section will provide an overview of energy poverty metrics and their calculation. The data and methods used in this analysis will be presented in Section 2.2. Section 2.3 will present the results. First, we will compare headline energy poverty rates as calculated by expenditure-based and self-reported measures. This will be followed by a discussion of how energy poverty/deprivation varies by general poverty status, tenure, dwelling type, age group and dwelling condition. 2.1 MEASURING ENERGY POVERTY AND DEPRIVATION Expenditure-based metrics measure energy poverty according to the proportion of disposable income spent on energy services in the household. While there are 3 Fuel poverty and energy poverty are often used interchangeably in the literature when discussing the affordability of adequate energy resources in the context of developing countries such as Ireland. This paper will use the term ‘energy poverty’. This should not be confused with the term ‘energy poverty’ when employed in an international development context, usually referring to inadequate access to energy. 4 | Energy poverty and deprivation in Ireland many metrics,4 the most common approach is to measure energy expenditures as a proportion of household income, with a household defined as being energy-poor if they spend more than 10 per cent of disposable income on energy services (Boardman, 1991). This metric is simple and transparent; however, it is not without criticism. Some households classified as energy-poor may be so because they have a large house or choose to keep it excessively warm as opposed to being forced into spending more than 10 per cent of their income on energy through necessity. This approach is most often applied using actual fuel expenditure data and so does not capture the extent to which households reduce their energy expenditures in response to price changes or income constraints.5 To overcome this, some applications use predicted rather than actual expenditure, where possible, with predicted values aligning with expected expenditure, absent a binding budget constraint. While predicting expenditures is one way to address this deficiency, self-reporting by households can also capture the extent to which a household goes without adequate heat in the home. These measures rely on self-reports made by householders on their capacity to afford the energy services they need (Watson and Maître, 2015). Often, the questions asked of householders cover whether they reduced their expenditure on energy services due to budget constraints, or whether their home is in such a condition that they cannot keep it adequately warm. This chapter will focus on understanding the extent to which Irish households are unable to afford adequate heat in their home by comparing the trajectory of expenditure-based energy poverty and self-reported energy deprivation. When considering self-reported deprivation, we will focus on those households who have reported an explicit inability to keep adequately warm. 2.2 DATA AND METHODOLOGY We use three primary data sources. The Household Budget Survey (HBS) is used to calculate expenditure-based energy poverty while both the Survey on Income and Living Conditions (SILC) and the Living in Ireland Survey (LIIS) are used to calculate self-reported measures of energy deprivation. 4 See Farrell (2021), BEIS (2021) and Tovar-Reaños and Lynch (2021) for a discussion. 5 Coyne et al. (2018) find evidence consistent with this in their study of an energy efficiency upgrade scheme in Ireland. This led to much smaller than expected energy savings, as households responded to the increased efficiency of their dwellings by increasing ‘thermal comfort’. Historical trends in energy poverty and deprivation | 5 Expenditure-based energy poverty The HBS contains a representative cross-sectional profile of household income and expenditure. All expenditures are recorded over a two-week period, including energy expenditures. Occupant socio-economic data are recorded, along with data on dwelling characteristics and appliance ownership. Responses are weighted to minimise any bias that may occur due to participant non-response. The HBS has been collected at regular intervals since 1987; the 1987, 1994, 1999, 2004/05, 2009/10 and 2015/16 waves are used in this paper. While many energy poverty metrics exist (most notably the Low Income High Cost of Hills (2012), and the recent Low Income Low Energy Efficiency approach of BEIS (2021), we focus on the 10 per cent income threshold approach for clarity of exposition. A household is deemed fuel-poor if they spend more than 10 per cent of their disposable income on energy services (electricity, heating oil, gas or solid fuels). Since we use actual energy expenditures, results should be construed in this context. For further discussion of expenditure-based approaches in an Irish context, and the nuanced differences in the calculation of energy poverty that result, see Farrell (2021). For the purposes of this analysis, we consider two energy expenditure scenarios to provide additional qualitative insight into the nature of energy poverty in Ireland. For our baseline calculation, we follow the literature (DCENR, 2011, 2016; Farrell, 2021; Scott et al., 2008; Tovar-Reaños and Lynch, 2022; Watson and Maître, 2015) and assume energy expenditures to include electricity, gas, oil and solid fuels. We adopt a secondary measure to focus on heating-related expenditures alone and exclude electricity. Space heating for about one fifth of households in Ireland is electric (c.18 per cent); the remainder are served by solid fuels (5 per cent), gas (39 per cent) or oil-fired central heating (36 per cent) (CSO, 2022). In addition, electricity is also used for services apart from heat. Self-reported energy deprivation Two data sources are used to calculate self-reported energy deprivation. From 1994–2001, the Living in Ireland Survey (LIIS) collected information on income and living conditions in Ireland.6 This survey included questions on ability to heat the home. Since 2003, these data have been collected as part of the EU Survey of Income and Living Conditions (SILC). The SILC is part of an EU project to provide harmonised data on income and living conditions, collected annually. As with the LIIS, the SILC contains questions on householder ability to heat the home. The Irish 6 We use data from waves 1–6 of the LIIS (1994–1999) to avoid potential concerns about the representativeness of later waves which were augmented with a booster sample: see Roantree et al. (2021) for further information. 6 | Energy poverty and deprivation in Ireland data are collected and managed by the Central Statistics Office (CSO) and used to monitor poverty and social exclusion in Ireland. Both the LIIS and the SILC collect information on the income and living conditions of households alongside a large range of sociodemographic information about the household members, ranging from personal characteristics to personal income, labour market position, education and health status (Watson and Maître, 2015). We use responses to questions in the LIIS and SILC relating to whether households had to go without heating during the last 12 months through lack of money, and whether they were unable to afford to keep the home adequately warm.7 Our measure of energy deprivation used throughout the paper is simply whether the head of household responds positively to either of these questions.8 2.3 RESULTS This section will compare the trajectory of energy poverty in Ireland according to self-reported energy deprivation and expenditure-based energy poverty metrics. The overall trajectory is first presented, followed by a discussion of how this varies by socio-economic cohort. The incidence of the measures by general poverty status, dwelling type, tenure, age group and housing condition is then discussed. When analysing these results, it should be noted that there is much less variation over time in expenditure-based metrics, in part because of the more intermittent nature of data collection. Figure 2.1 compares the trajectory of headline energy poverty and deprivation metrics. Both self-reported energy deprivation and expenditure-based energy poverty measures declined over the 1990s and early 2000s before rising in the aftermath of the Great Recession. This rise is more evident in the self-reported measure of energy deprivation, perhaps in part because the data are collected at more regular intervals than expenditure on energy. However, the greater sensitivity of self-reported energy deprivation to economic conditions also mirrors that of material deprivation more generally (Roantree et al., 2021). This is consistent with households reducing their expenditure on and going without adequate energy when incomes are squeezed. While energy deprivation fell from a peak of almost 20 per cent in 2013 to 9 per cent in 2018, it has started to rise again since, despite – what Figure B.6 in the appendix shows was – relative stability in the price of most fuels over this period. 7 The phrasing of these questions in the LIIS is slightly different to that in the SILC, so may give rise to a structural break in the series. We indicate this in what follows by not connecting the lines between LIIS and SILC estimates. 8 This differs slightly from the measure used by Watson and Maître (2015), which also includes whether households were in arrears on utility bills. Appendix Table B.1 presents estimates of the share of all and selected sub-groups of households in arrears over time. Historical trends in energy poverty and deprivation | 7 FIGURE 2.1 TRAJECTORY OF ENERGY POVERTY/DEPRIVATION IN IRELAND (1994–2020) Source: Authors’ calculations using the Household Budget Survey, Living in Ireland Survey, and Survey of Income and Living Conditions. The metrics of self-reported energy deprivation used in this paper place a greater emphasis on heat-related deprivation in the household. To compare expenditurebased energy poverty on a more similar footing to these metrics of energy deprivation, we also report the headline energy poverty rates excluding electricity. While both expenditure-based energy poverty and self-reported energy deprivation are more closely aligned once electricity expenditures are removed over the 1990s and early 2000s, this is less true in recent years. The rate of energy poverty and self-reported energy deprivation for the entire population can mask significant variation across the population. We now turn to look at the experience of different groups using both the measure of spending more than 10 per cent of income on energy, including electricity, and self-reported energy deprivation. First, we explore the extent to which energy poverty and deprivation overlaps with income poverty, as captured by households’ at-risk-ofpoverty status. This is defined as living in a household with less than 60 per cent of the median income level, adjusted (equivalised) for household size. Figure 2.2 shows the rates of energy poverty/deprivation for those at risk of poverty and for those that are not. We see a similar pattern for both self-reported and expenditure-based metrics; a much greater proportion of households that are at risk of poverty experience energy poverty/deprivation than households that are not at risk of poverty (i.e. above the poverty line). 0% 5% 10% 15% 20% 25% 1990 1995 2000 2005 2010 2015 2020 >10% income (incl. electricity) >10% income (excl. electricity) self-reported deprivation 8 | Energy poverty and deprivation in Ireland FIGURE 2.2 RATES OF ENERGY POVERTY AND DEPRIVATION, BY AT-RISK-OF-POVERTY STATUS Sources: Authors’ calculations using the Household Budget Survey, the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. Note: Energy poverty calculation includes electricity. Indeed, in 1999 almost 60 per cent of those at risk of poverty were also in energy poverty, though this has fallen in recent years: to 32 per cent in 2009/10 and 40 per cent in 2015/16. In contrast, less than 5 per cent of households who were not at risk of poverty experienced energy poverty in 2015/ 16. For self-reported measures, we estimate that 38 per cent of those at risk of poverty experienced energy deprivation in 1994, falling to 20 per cent by 2020. While for most of the period we observe that less than 10 per cent of those not at risk of poverty also reported enduring energy deprivation, this proportion rose significantly during the financial crisis and, after falling back to low levels during the recovery, has started to increase again in recent years.9 Although they face much higher rates of energy poverty and deprivation, those at risk of poverty make up a relatively small share – less than 20 per cent in 2019 – of the overall population. Thus those not at risk of poverty still make up a relatively large share of those experiencing energy poverty and deprivation. This is shown in Figure 2.3, which plots the composition of those in energy poverty and energy deprivation over time. While those experiencing energy poverty are increasingly also at risk of poverty (that is, below the income poverty line), most of those who report experiencing energy deprivation are not at risk of poverty. 9 Appendix Table B.2 shows that these households are more likely to be renters, lone parents, to live in households without anyone in paid work, and to live in poor-quality dwellings. 0% 10% 20% 30% 40% 50% 60% 70% 1990 1995 2000 2005 2010 2015 2020 Energy poverty rate: not at-risk-of-poverty Energy poverty rate: at-risk-of-poverty Energy deprivation rate: not at-risk-of-poverty Energy deprivation rate: at-risk-of poverty Historical trends in energy poverty and deprivation | 9 FIGURE 2.3 COMPOSITION OF THOSE EXPERIENCING ENERGY POVERTY AND DEPRIVATION, BY WHETHER AT-RISK-OF-POVERTY COMPOSITION OF ENERGY POVERTY BY AT-RISK-OF-POVERTY Sources: Authors’ calculations using the Household Budget Survey. Note: Energy poverty calculation includes electricity. COMPOSITION OF ENERGY DEPRIVATION BY AT-RISK-OF-POVERTY Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. 0% 5% 10% 15% 20% 25% 1994 1999 2004 2009 2015 Not at-risk-of-poverty but in energy poverty At-risk-of-poverty and in energy poverty 0% 5% 10% 15% 20% 25% 1994 1995 1996 1997 1998 1999 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Not at-risk-of-poverty but energy deprived At-risk-of-poverty and energy deprived 10 | Energy poverty and deprivation in Ireland Figure 2.4 shows rates of energy poverty and deprivation by dwelling type. These show some interesting contrasts, with apartment dwellers experiencing the highest rate of energy deprivation but the lowest rates of energy poverty throughout. Conversely, those living in detached houses report the highest rates of energy poverty but the lowest rates of energy deprivation. This difference likely reflects the nature of fuel poverty being experienced by different cohorts. It suggests that, while apartment dwellers are more exposed to being deprived of adequate warmth than those living in houses, those living in detached houses need to spend much more of their income to adequately heat their homes. There are many possible reasons for this, including that those who live in apartments have, on average, lower levels of resources, leading to a greater propensity to go without adequate heating. The divergence between expenditure-based and self-reported measures is also evident in housing tenure. Figure 2.5 shows that, according to expenditure-based metrics, a similar proportion of renters and homeowners experience expenditurebased energy poverty. A much clearer delineation appears with respect to selfreported energy deprivation, where we find that renters are more likely to experience energy deprivation. Figure 2.5 shows that a much higher proportion of renters, up to 35 per cent in 1994, report being deprived of adequate heat. While this fell to c.20 per cent in the late 2010s, this is much greater than the rate for homeowners of c.6 per cent. Indeed, the rate of homeowners who report being deprived of adequate heat was consistently below 14 per cent during the duration of analysis – and often below 5 per cent. There are many plausible reasons for a greater incidence of self-reported energy deprivation among renters than expenditure-measured energy poverty. This may reflect the other sociodemographic variables correlated with rental: younger and perhaps lower-income groups who are more likely to respond to an energy-related budget constraint by reducing their consumption. The analysis in this paper does not allow for identification of such effects, but nevertheless shows that expenditure-based and self-reported measures of energy poverty can lead one to identify very different groups vulnerable to the rising cost of energy, a topic that is analysed in greater detail in Chapter 3.10 10 It would also be useful to consider the prevalence of energy poverty and deprivation among dwellings with poor standards of insulation. However, this information is unavailable in the HBS and the SILC datasets. The SILC data, however, identifies dwellings reporting problems with leaks, mould or damp. Taking this as a proxy for housing condition, Figure B3 of the Appendix decomposes energy deprivation according to dwelling condition. Similarly, Figure B4 of the Appendix shows the rates of energy deprivation by dwelling condition. These results show that – as one might expect – levels of energy deprivation are far higher among those with dwellings in a poor condition, although given this group makes up a relatively small share of the population, most of those reporting energy deprivation live in dwellings of better condition. Historical trends in energy poverty and deprivation | 11 FIGURE 2.4 RATES OF ENERGY POVERTY/DEPRIVATION BY DWELLING TYPE RATES OF ENERGY POVERTY BY DWELLING TYPE Sources: Authors’ calculations using the Household Budget Survey. Note: Excludes ‘other’ housing type. Energy poverty calculation includes electricity. RATES OF ENERGY DEPRIVATION BY DWELLING TYPE Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. Note: Excludes small number in ‘other’ dwelling type. Apartment/bedsit series based on small number of observations in 1999 due to attrition from LIIS, so should be interpreted with caution. 0% 5% 10% 15% 20% 25% 30% 1994 1999 2004 2009 2015 Detached Semi-detached Apartment/Bedsit 0% 5% 10% 15% 20% 25% 30% 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Detached Semi Detached Apartment/Bedsit 18 | Energy poverty and deprivation in Ireland FIGURE 3.2 SIMULATED IMPACT OF RECENT ENERGY PRICE INCREASES, BY HOUSEHOLD TYPE Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Deciles constructed equivalising income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix; the energy CPI sub-indices from January 2021 and April 2022 are used to simulate the price rise. This is in part because these groups tend to have lower incomes, and in part because more of their overall spending goes towards these types of goods. While not captured in these estimates, it should also be borne in mind that many of these identified vulnerable groups are less likely to have savings or other sources of accumulated wealth to draw upon when responding to such financial shocks (Lydon and McIndoe-Calder, 2021). As we will see in Chapter 4, these gradients have important implications for how policymakers can most effectively target those most affected by energy inflation. Table 3.1 presents our estimate of the impact of these energy price increases on measures of energy poverty. The measures we estimate are expenditure-based, with households spending more than 10 per cent of income on energy classified as being in energy poverty. Chapter 2 showed that these expenditure-based measures are quite sensitive to what precisely is counted as energy expenditure and income. To isolate the effect of price changes in electricity and all other (predominantly heating) fuels, we estimate variants including and excluding electricity expenditure. Relative to 2015/16 data, we estimate that expenditurebased energy poverty has risen from 13.2 per cent to 29.4 per cent, including electricity. If we exclude electricity expenditure, those spending more than 10 per cent of their disposable income on all other energy sources rises from 5.1 per cent to 12.7 per cent. To an extent, this reflects the sensitivity of expenditure-based measures of energy poverty to changes in the numerator, with a cluster of 0% 1% 2% 3% 4% 5% 6% 7% 8% Total Income quintile At-risk-ofpoverty Location Tenure €0 €10 €20 €30 €40 €50 €60 % of disposable income € per week home heating & electricity (€) ... & motor (€) home heating & electricity (%) ... & motor (%) Impact of rising energy prices on households | 19 households located near the 10 per cent threshold (Tovar-Reaños and Lynch, 2022; O’Malley et al., 2019). TABLE 3.1 SIMULATED IMPACT OF RECENT ENERGY PRICE INCREASES ON ENERGY POVERTY Energy spending >10% of disposable income: Excluding electricity Including electricity 2015/16 5.1% 13.2% Nowcast 12.7% 29.4% Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: The energy CPI sub-indices from January 2021 and April 2022 are used to simulate the price rise. 3.3 IMPACT OF POTENTIAL FUTURE ENERGY PRICE INCREASES Figure 3.3 presents our estimates of the average impact of potential future energy price increases.15 Specifically, we simulate the impact of a further 25 per cent increase in energy prices, broadly calibrated to the anticipated rise in electricity and gas prices in May.16 We assume the same proportional increase in liquid fuel, solid fuel and motor fuel prices, though there is likely to be some variation in these. FIGURE 3.3 IMPACT OF SIMULATED ENERGY PRICE INCREASES, BY HOUSEHOLD TYPE Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Deciles constructed equivalising income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix. The energy CPI sub-indices from January 2021 and April 2022 are used to simulate the price rise, plus additional 25 per cent increase. 15 Table B.4 presents estimates for further sub-groups of households. 16 See https://www.rte.ie/news/business/2022/0330/1289331-electric-ireland-price-hike/ 0% 2% 4% 6% 8% 10% 12% 14% Total Income quintile At-risk-ofpoverty Location Tenure €0 €10 €20 €30 €40 €50 €60 €70 €80 €90 €100 % of disposable income € per week home heating & electricity (€) ... & motor (€) home heating & electricity (%) ... & motor (%) 20 | Energy poverty and deprivation in Ireland TABLE 3.2 SIMULATED IMPACT OF POTENTIAL ENERGY PRICE INCREASES ON ENERGY POVERTY Energy spending >10% of disposable income: Excluding electricity Including electricity 2015/16 estimate 5.1% 13.2% Forecast 20.5% 43.0% Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: The energy CPI sub-indices from January 2021 and April 2022 are used to simulate the price rise, plus additional 25 per cent increase. The bars show that such further levels of energy inflation would increase the impact on households to an average of €36.57 per week excluding motor fuel (4.0 per cent of disposable income) and €67.66 per week (7.4 per cent of disposable income) including motor fuel. Again, these impacts exhibit strong gradients with respect to household income, at-risk-of-poverty status, location and tenure. Table 3.2 presents our estimates of energy poverty levels incorporating the additional energy inflation presented in Figure 3.3. These are substantially higher than the estimates in Table 3.1, ranging from between 20.5 per cent excluding electricity to 43.0 per cent when electricity is included. Increases of this magnitude would leave energy poverty rates at their highest recorded level, surpassing rates seen in 1994. To summarise, the impact of recent – and potential future – increases in energy prices is substantial, especially for lower-income households whose main source of income is social transfers (CSO, 2022). We now turn to assess options that policymakers might consider in trying to mitigate these impacts. . Policy options | 21 CHAPTER 4 Policy options This chapter considers potential policy options to mitigate the impact of rising energy prices on households. It begins by examining cuts to indirect taxes on energy, before turning to look at cuts to direct taxes on personal income and increases to social transfers. 4.1 INDIRECT TAX CUTS Much of the public discussion around potential policy responses to rising energy costs has focused on cuts to indirect taxes on energy.17 There are currently three main indirect taxes on energy: excise duties (levied at a rate per litre or megawatt hour), the carbon tax (levied at a rate per tonne of CO2-equivalent) and VAT (levied at a reduced rate of 13.5 per cent on electricity and home heating fuels). While cutting any of these taxes would offer support to households affected by rising energy prices, doing so might also have unintended or undesirable effects. Rising prices can help avoid shortages or rationing in the event of supply constraints, with research showing that price caps can increase the risk of shortages and blackouts (Reiss and White, 2008). In the long run, price caps also weaken the incentive to invest in energy-saving technology and behaviour. In addition, cutting indirect taxes on energy exacerbates existing effective subsidies to burning fossil fuel (de Bruin et al., 2019), with, for example, reductions to VAT on electricity and gas further distorting consumption decisions towards such services and away from goods or services subject to the standard rate. Figure 4.1 shows the impact of various proposed cuts to indirect taxes on energy in terms of both cash gains (the bars, Euro per week) and as a percentage of disposable income (the connected lines).18 The blue series show that reducing VAT on electricity and gas from 13.5 to 9 per cent would lower households’ bills by an average of €1.16 per week (0.1 per cent of disposable income). Although gains would be larger in cash terms for higher-income households (an average of €1.53 per week for the highest income quintile compared to €0.70 for the lowest-income quintile), they would be smaller as a percentage of disposable income (0.1 per cent compared to 0.2 per cent). 17 See, for example, https://www.thejournal.ie/vat-energy-costs-5724186-Mar2022/ 18 We do not show the impact of measures on our estimates of energy poverty as these are predominantly determined by and sensitive to the numerator (energy expenditure), with changes to the denominator (after-tax and welfare income) of the magnitude we consider having little impact. Appendix Table B.5-6 contain estimates for other subgroups. 22 | Energy poverty and deprivation in Ireland FIGURE 4.1 SIMULATED IMPACT OF INDIRECT TAX MEASURES, BY HOUSEHOLD TYPE Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Deciles constructed equivalising income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix. Similarly, while cash gains are larger for homeowners and those at risk of poverty, gains as a percentage of disposable income are larger for renters and those who already reported an existing inability to heat their home. Both cash and proportional gains are larger for those living in rural than urban areas, reflecting rural households’ greater expenditure on energy in both absolute terms and as a percentage of income. The red series in Figure 4.1 shows the impact of removing the carbon tax on home heating oil. The average impact is similar to the cut to VAT on electricity and gas at €1.10 per week or 0.1 per cent of disposable income. However, although the results show the same pattern of gains in cash and proportional terms by household income, tenure and at-risk-of-poverty status, they differ substantially by location. This is because there is much greater reliance on home heating oil in rural areas that are not connected to the natural gas grid, and very little usage in urban areas. Finally, the orange series shows the impact of cutting excise duty on petrol and diesel by 50c per litre. We estimate that households would gain by an average of €1.71 per week (0.2 per cent of disposable income). However, unlike the other indirect tax changes considered, the gradient of gains as a percentage of income is flat across the lowest four income quintiles. This reflects the fact that a smaller 0.0% 0.1% 0.1% 0.2% 0.2% 0.3% 0.3% Total Income quintile At-risk-ofpoverty Location Tenure €0.00 €0.50 €1.00 €1.50 €2.00 €2.50 €3.00 % of disposable income € per week VAT cut (€) Carbon tax on heating oil (€) Cut to fuel duties (€) VAT cut (%) Carbon tax on heating oil (%) Cut to fuel duties (%) Policy options | 23 share of lower-income households’ total spending goes on motor fuel than of higher-income households’ total spending. In all cases, the indirect tax cuts would only compensate households for a fraction of the increase in energy prices that has been experienced since 2021. This is despite a non-trivial cost, which we estimate in excess of €110 million per year for the cut to VAT and the abolition of carbon tax on heating oil, and €170 million per year for the cut in motor fuel duties.19 Furthermore, although the gains from cutting VAT or the carbon tax on heating oil are largest as a share of disposable income for lower-income households proportionally most affected by energy price increases, most of the cost arises from cutting taxes for higher-income households. This is because higher-income households spend more in absolute terms on energy and motor fuel. As a result, about half of the overall cash gains from cutting indirect taxes on energy go to the two highest-income quintiles compared to less than a third to the two lowest-income quintiles. This illustrates the fundamental trade-off policymakers face between targeting measures towards those most affected and the cost of the measures, a topic we return to in Chapter 5. We now turn to assess potential direct tax and welfare changes, considering whether these can more effectively target support towards households most affected by energy inflation. 4.2 DIRECT TAX AND WELFARE CHANGES We examine direct tax and welfare changes that would each cost around €230 million per year. Figure 4.2 presents our estimate of the impact of social welfare changes: a double social welfare payment akin to the annual Christmas bonus, a doubling of the fuel allowance, and a €120 lump-sum electricity credit. While all three changes would increase average household incomes by the same amount (€2.30-€2.50 per week, or 0.25 per cent of disposable income), there are differences in the impact of the measures across groups. Unlike cuts to indirect taxes, the first two changes would benefit lower-income households most both in cash terms and as a share of disposable income. This is because social welfare payments are means-tested, meaning few higher-income households are in receipt of the payments and so cannot gain from their increase. The fuel allowance increase is slightly more progressive than the social welfare bonus – with an average gain of €5.59 vs €4.02 per week for the lowest income quintile – but is restricted to longer-term claimants of social welfare payments, such as Jobseeker’s Benefit, so excludes some low-income welfare-dependent households. 19 Our costings are based on the estimated gain to households from the tax cuts, but many non-household users (like businesses or public authorities) would also benefit; this is not captured in our survey data and would have an additional Exchequer cost. 24 | Energy poverty and deprivation in Ireland FIGURE 4.2 SIMULATED IMPACT OF WELFARE MEASURES, BY HOUSEHOLD TYPE Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Deciles constructed equivalising income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix. While both the social welfare bonus and double fuel allowance are also very well targeted towards those at risk of poverty, they are less well targeted towards renters who – as we saw in Chapter 2 – face higher rates of energy deprivation relative to homeowners. This reflects the fact that pensioners – who make up a large share of the beneficiaries of both measures – are much more likely to own their own home than low-income working-age adults. In contrast, the €120 lumpsum electricity credit results in proportionally larger gains for renters than homeowners, but is less progressive than the two other social welfare measures. This is because higher-income households gain by the same amount in cash terms as lower-income households. As a result, while still progressive, the electricity credit can be considered somewhat less well targeted towards those most affected by energy inflation (though better targeted than indirect tax cuts).20 While welfare changes can therefore be designed to accurately target the households most adversely affected by rising energy prices, the absence of an income support for low-income working individuals without children means that policymakers may also wish to look at changes to direct taxation.21 20 Appendix Table B.7 and 8 contain estimates for other sub-groups. 21 The introduction of such a support – a feature of social welfare systems in many advanced economies – has been suggested by the National Economic and Social Council (2020), Roantree (2020) and Keane et al. (2021) among others. 0.0% 0.2% 0.4% 0.6% 0.8% 1.0% 1.2% 1.4% 1.6% 1.8% 2.0% Total Income quintile At-risk-ofpoverty Location Tenure €0.00 €1.00 €2.00 €3.00 €4.00 €5.00 €6.00 €7.00 € per week Social Welfare Bonus (€) Double Fuel Allowance (€) €120 energy credit (€) Social Welfare Bonus (%) Double Fuel Allowance (%) €120 energy credit (%) Policy options | 25 FIGURE 4.3 SIMULATED IMPACT OF DIRECT TAX MEASURES, BY HOUSEHOLD TYPE Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Deciles constructed equivalising income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix. Figure 4.3 shows the estimated impacts of two direct tax measures aimed at lower earners: increasing the Pay Related Social Insurance (PRSI) credit from €12 to €33 and increasing the main income tax credits by €50 per year.22 We estimate that both these measures would increase households’ incomes by about €2.30 per week (0.3 per cent of disposable income). While neither tax cut is well targeted towards those at risk of poverty or the very lowest-income households (who typically do not have earnings high enough to pay income tax or PRSI), increasing the PRSI credit results in gains that are highest for those in the second-lowest income quintile. This contrasts with increasing the main income tax credits, which benefits households in the middle of the income distribution most as a share of income and results in largest cash increases for higher-income households.23 Taken together, these results show that the direct tax and welfare system provide a much more precise means of targeting support towards the households most adversely affected by rising energy costs than cuts to indirect taxes. The final chapter of this report summarises the findings of our research and its implications for policy. 22 The PRSI credit reduces the PRSI liabilities of lower earners. Increasing the credit by this amount could require making refunds to some taxpayers in a similar way to how Revenue made payments to firms during the course of the pandemic. 23 Appendix Table B.9 and 10 contain estimates for other sub-groups. 0.0% 0.1% 0.2% 0.3% 0.4% 0.5% 0.6% 0.7% Total Income quintile At-risk-ofpoverty Location Tenure €0.00 €0.50 €1.00 €1.50 €2.00 €2.50 €3.00 €3.50 €4.00 €4.50 % of disposable income € per week 175% increase in PRSI credit (€) €50 increase in the tax credit (€) 175% increase in PRSI credit (%) €50 increase in the tax credit (%) Conclusion | 27 CHAPTER 5 Conclusion This report has examined energy poverty and deprivation in Ireland, a topic that rapidly rising prices have recently returned to the forefront of public debate. While expenditure-based measures of energy poverty and self-reported measures of energy deprivation have both declined over time, the two approaches suggest that between 5 and 15 per cent of households were experiencing problems adequately heating their homes before the recent rise in prices. Variation in the level of these estimates is partly due to whether or not electricity is included in energy expenditure, but also reflects a conceptual difference in what is being measured. Self-reported measures of energy deprivation place a greater emphasis on heat-related deprivation in the household, while expenditure-based measures can include some households with large or excessively heated homes and exclude others whose incomes are too low to allow them to keep their homes adequately heated. We also highlight important socio-economic differences between groups identified as vulnerable to rising fuel prices by expenditure-based measures of energy poverty and self-reported measures of energy deprivation. First, while there is a substantial overlap between measures of energy poverty and income poverty (as captured by the at-risk-of-poverty line), there is less overlap between measures of self-reported energy deprivation and income poverty. Second, while expenditurebased measures of energy poverty are highest for those living in detached dwellings and lowest for those living in apartments, the reverse is true for selfreported measures of energy deprivation. Similarly, while expenditure-based measures of energy poverty show little difference between homeowners and renters, self-reported measures of energy deprivation are much higher for renters than homeowners. This suggests that, while homeowners and those living in detached dwellings spend a larger share of their income heating their homes, they also have a greater capacity to do so, likely reflecting their, on average, higher levels of income. By contrast, renters and those living in apartments are more likely to endure energy deprivation and go without heat because they cannot afford it. This illustrates the importance of developing more sophisticated measures of energy poverty, something previously highlighted by O’Malley et al. (2020) and Tovar-Reaños and Lynch (2020). From this perspective, the recent establishment of a research network on fuel poverty by the Department of Communications, 34 | Energy poverty and deprivation in Ireland The first stage estimates a probit model in order to calculate the probability of having non-zero expenditure: Pr(𝑎𝑎𝑁𝑁𝑐𝑐= 1)=𝜓𝜓(𝛼𝛼𝑐𝑐,0+�𝛽𝛽𝑐𝑐,0(ln 𝑁𝑁𝑁𝑁𝑁𝑁)𝑚𝑚+𝛾𝛾𝑐𝑐,0 𝒙𝒙+ 𝜖𝜖𝑐𝑐,0) 𝑚𝑚 The dependent variable is the probability of having positive expenditure on a given fuel, 𝑐𝑐. The exponent of NDE, 𝑚𝑚, is set to 2, while the vector of covariates contains the same demographic and household characteristics as before. The second stage uses an OLS regression to estimate budget shares of non-durable expenditure for a given fuel, 𝑤𝑤𝑐𝑐=𝐸𝐸𝑐𝑐 𝑁𝑁𝑁𝑁𝐸𝐸, conditional on having positive expenditure for that fuel. wc=𝛼𝛼𝑐𝑐+�𝛽𝛽𝑐𝑐(ln 𝑁𝑁𝑁𝑁𝑁𝑁)𝑚𝑚+𝛾𝛾𝑐𝑐 𝒙𝒙+ 𝜖𝜖𝑐𝑐 𝑖𝑖𝑖𝑖 𝑎𝑎𝑁𝑁𝑐𝑐= 1 𝑚𝑚 A second-order polynomial is again used for the expenditure term, while the control variables are the same as before.25 The estimated coefficients from both stages of this step are later used in Step 5. Step 3: Adjust the income distribution from SILC to match the HBS The distribution of disposable income from SILC is then adjusted to align it to the HBS data. We take the weekly disposable income for each household simulated by SWITCH under the baseline tax and welfare policy (described below) in 2015/2016 prices, then identify outliers in the distributions of disposable incomes in both the SILC and HBS datasets. This is done using the Chauvenet method for detecting outliers, an iterative procedure where an observation is marked as an outlier if it falls outside the criterion chosen. In this case, the criterion assumes a lognormal distribution: (𝑙𝑙𝑎𝑎 𝑦𝑦− 𝑙𝑙𝑎𝑎 𝑦𝑦 � � � � � ) /𝜎𝜎ln𝑦𝑦>𝑍𝑍(1−2𝑁𝑁)−1 where ln 𝑦𝑦 � � � � � is the mean and 𝜎𝜎 is the standard deviation of disposable income. Once outliers have been identified, the distribution of disposable income is standardised by scaling it using these moments of disposable income in the HBS, specifically: 𝑦𝑦�=�𝑦𝑦−𝑦𝑦� 𝜎𝜎𝑦𝑦�∗𝜎𝜎𝑦𝑦,𝐻𝐻𝐻𝐻𝐻𝐻+𝑦𝑦�𝐻𝐻𝐻𝐻𝐻𝐻 Step 4: Impute non-durable expenditure Once the variables in the SILC dataset have been constructed such that they have the same structure as their HBS counterparts, non-durable expenditure is imputed 25 Price is not included as a control in this equation as the exercise is a static microsimulation one, with the aim of imputing expenditures from the HBS to SILC rather than predicting what similar households would spend today given the higher level of prices. Such a dynamic microsimulation approach would require a more detailed demand system estimation like that adopted by Tovar-Reaños and Lynch (2019). Appendix | 35 into the SILC data. This simply involves taking the estimated coefficients from Step 1 above and the values of the variables to generate predicted values of nondurable expenditure for each household in our SILC dataset. The independent variables for income, as well as their interactions with the control variables, are the adjusted series (from Step 3) and not the original series in the SILC dataset. This means that our imputation procedure accounts for the lower level of employment and disposable income in 2020 arising from pandemic-related job losses, though it assumes that the relationship between expenditure and these (along with other demographic variables) is the same as in 2015/2016. The resulting imputation of non-durable expenditure is then adjusted in the same manner as income in Step 3 before being used as a variable in the next step. Step 5: Impute budget shares In the same manner as the previous step, budget shares are imputed following the two-stage procedure using the estimates from Step 2. The non-durable expenditure variables and their polynomials are again the adjusted series rather than the original prediction. 36 | Energy poverty and deprivation in Ireland B. ADDITIONAL FIGURES AND TABLES FIGURE B.1 COMPOSITION OF ENERGY POVERTY/DEPRIVATION BY DWELLING TYPE COMPOSITION OF ENERGY POVERTY BY DWELLING TYPE Sources: Authors’ calculations using the Household Budget Survey. Note: Energy poverty calculation includes electricity. COMPOSITION OF ENERGY DEPRIVATION BY DWELLING TYPE Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. 0% 5% 10% 15% 20% 25% 1994 1999 2004 2009 2015 Detached Semi-Detached Apartment/Bedsit Other 0% 5% 10% 15% 20% 25% 1994 1995 1996 1997 1998 1999 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Detached Semi-detached Apartment/Bedsit Other Appendix | 37 FIGURE B.2 COMPOSITION OF ENERGY POVERTY/DEPRIVATION BY HOUSEHOLD TENURE COMPOSITION OF POVERTY BY HOUEHOLD TENURE Sources: Authors’ calculations using the Household Budget Survey. Note: Energy poverty calculation includes electricity. COMPOSITION OF ENERGY DEPRIVATION BY HOUSEHOLD TENURE Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. 0% 5% 10% 15% 20% 25% 1994 1999 2004 2009 2015 Homeowner Renter 0% 5% 10% 15% 20% 25% Homeowner Renter 38 | Energy poverty and deprivation in Ireland FIGURE B.3 COMPOSITION OF ENERGY DEPRIVATION BY DWELLING CONDITION Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. FIGURE B.4 RATES OF ENERGY DEPRIVATION BY DWELLING CONDITION Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20% 1994 1995 1996 1997 1998 1999 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Bad Condition Good Condition 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 1994 1995 1996 1997 1998 1999 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Bad Condition Good Condition Appendix | 39 FIGURE B.5 COMPONENTS OF SELF-REPORTED ENERGY DEPRIVATION (1994–2020) Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. FIGURE B.6 CHANGE IN SELECTED CPI SUB-INDICES 2003–2021 (2015=100) Note: Authors’ calculations using CSO Table CPM16, indexed to average value in 2015. 0% 5% 10% 15% 20% 25% 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Cannot afford heat Lack of heat Either Both 0 20 40 60 80 100 120 140 160 2003M01 2005M01 2007M01 2009M01 2011M01 2013M01 2015M01 2017M01 2019M01 2021M01 Electricity Natural gas Liquid fuels (home heating oil) Solid fuels 40 | Energy poverty and deprivation in Ireland TABLE B.1 SHARE OF HOUSEHOLDS IN ARREARS ON UTILITY BILL IN POPULATION GROUPS Year Overall Not at risk of poverty At risk of poverty Renter Homeowner Detached House Semidetached House Apartment / Bedsit 1994 8.8% 6.7% 16.7% 20.1% 6.2% 4.9% 11.4% 12.6% 1995 6.1% 4.3% 12.4% 14.1% 4.2% 4.9% 7.1% 8.7% 1996 6.7% 5.0% 12.8% 16.5% 4.3% 4.6% 8.5% 6.1% 1997 5.3% 3.6% 10.7% 12.7% 3.3% 3.2% 7.4% 2.5% 1998 3.9% 2.7% 7.4% 10.7% 2.2% 2.0% 5.0% 9.2% 1999 3.9% 2.6% 7.6% 10.8% 2.2% 3.2% 4.7% 3.4% 2004 7.2% 5.26% 13.15% 19.4% 4.2% 3.3% 10.0% 14.7% 2005 6.1% 4.43% 11.74% 16.7% 3.1% 2.9% 8.2% 10.2% 2006 6.2% 4.57% 12.29% 17.4% 2.9% 2.7% 8.3% 11.6% 2007 6.0% 3.74% 14.16% 17.3% 2.6% 3.7% 6.9% 16.0% 2008 7.7% 6.26% 14.32% 18.4% 4.4% 4.0% 10.0% 15.7% 2009 9.6% 8.13% 16.99% 20.8% 5.6% 5.3% 12.1% 16.8% 2010 11.4% 9.47% 21.94% 22.2% 7.4% 6.7% 14.0% 17.5% 2011 12.7% 10.27% 23.01% 25.8% 7.0% 9.3% 14.4% 16.0% 2012 15.1% 12.71% 26.31% 25.4% 10.6% 10.0% 18.3% 16.4% 2013 16.4% 13.71% 29.56% 28.3% 11.5% 10.3% 20.1% 19.2% 2014 15.8% 13.24% 28.62% 25.8% 11.5% 12.1% 18.2% 18.8% 2015 13.4% 10.85% 25.42% 23.7% 9.2% 9.1% 16.0% 18.5% 2016 10.7% 8.07% 22.18% 19.9% 6.9% 7.0% 12.5% 16.3% 2017 8.8% 6.40% 19.50% 15.3% 6.2% 6.0% 10.7% 10.3% 2018 7.4% 5.37% 16.10% 14.8% 4.4% 4.4% 9.1% 10.4% 2019 7.9% 5.80% 18.12% 16.0% 4.4% 3.7% 10.9% 8.0% 2020 7.5% 5.65% 16.43% 17.1% 3.3% 4.7% 8.7% 11.5% Sources: Authors’ calculations using the Living in Ireland Survey and the Survey of Income and Living Conditions Research Microdata Files. Appendix | 41 TABLE B.2 ESTIMATES FROM PROBIT REGRESSION MODEL OF FUEL DEPRIVATION (FOR THOSE NOT AT RISK OF POVERTY) Average Marginal Effects Model 1 Model 2 Disability in Household 0.014* 0.011 Education: Lower secondary 0.024 0.022 Upper secondary -0.003 -0.002 (ref.: Post-secondary) Family Type: Single Adult 0.033* 0.032* Lone Parent Family 0.117*** 0.107*** (ref.: Couple) Couple with dependent children 0.015 0.017 Single Adult (65 years old or older) 0.019 0.019 Couple (at least on aged over 65) -0.007 -0.007 Non-related household 0.005 0.001 3 Adults + 0.001 0.0002 3 Adults + with dependent children 0.056*** 0.055*** Dublin City 0.007 0.006 Household with no-one in paid employment 0.061*** 0.058*** Household owns home -0.042*** -0.032*** House in poor condition 0.056*** Adj. R-squared 0.048 0.061 N 7,166 Sources: Authors’ calculations using the Survey of Income and Living Conditions Research Microdata Files. Notes: *** p<.001; ** p<.01; * p<.05; ± p<.10. 42 | Energy poverty and deprivation in Ireland TABLE B.3 SIMULATED IMPACT OF RECENT ENERGY PRICE INCREASES, BY HOUSEHOLD TYPE Household type Heating & electricity (€pw) Heating & electricity (%) … & motor fuel (€pw) … & motor fuel (% income) All €21.27 €38.63 2.3% 4.2% Income quintile Lowest €13.08 €20.36 3.8% 5.9% 2 €18.37 €30.27 3.1% 5.2% 3 €22.37 €41.46 2.6% 4.9% 4 €24.65 €47.42 2.3% 4.4% Highest €27.89 €53.68 1.6% 3.1% Can afford to heat home adequately Yes €16.71 €29.20 2.6% 4.6% No €21.75 €39.63 2.3% 4.2% Dwelling location Rural €24.20 €46.64 2.8% 5.4% Urban €19.21 €33.01 2.0% 3.5% Housing tenure Homeowner €24.63 €44.38 2.5% 4.5% Renter €13.31 €25.06 1.7% 3.2% Dwelling condition Good €21.60 €39.31 2.3% 4.2% Poor €18.94 €33.81 2.3% 4.2% Dwelling type Detached €27.19 €50.09 2.8% 5.2% Semi-detached €18.83 €33.69 2.1% 3.8% Flat/other €11.16 €20.25 1.4% 2.5% At-risk-of-poverty Yes €13.26 €21.80 4.1% 6.7% No €22.47 €41.16 2.2% 4.1% Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Quintiles of equivalised income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix. The simulated price rise shown is that experienced between January 2021 and April 2022. Appendix | 43 TABLE B.4 SIMULATED IMPACT OF POTENTIAL FUTURE ENERGY PRICE INCREASES, BY HOUSEHOLD TYPE Household type Heating & electricity (€pw) Heating & electricity (%) … & motor fuel (€pw) … & motor fuel (% income) All €36.57 €67.66 4.0% 7.4% Income quintile Lowest €22.78 €35.97 6.5% 10.3% 2 €31.47 €52.94 5.4% 9.1% 3 €38.69 €72.87 4.6% 8.6% 4 €42.52 €83.17 3.9% 7.7% Highest €47.43 €93.40 2.8% 5.4% Can afford to heat home adequately Yes €29.17 €51.73 4.6% 8.2% No €37.35 €69.33 4.0% 7.3% Dwelling location Rural €40.79 €80.58 4.7% 9.4% Urban €33.61 €58.59 3.5% 6.1% Housing tenure Homeowner €41.90 €77.13 4.3% 7.9% Renter €24.00 €45.33 3.1% 5.8% Dwelling condition Good €37.12 €68.82 4.0% 7.4% Poor €32.72 €59.43 4.0% 7.3% Dwelling type Detached €45.75 €86.40 4.7% 8.9% Semi-detached €32.92 €59.70 3.7% 6.7% Flat/other €20.25 €36.89 2.5% 4.6% At-risk-of-poverty Yes €23.19 €38.66 7.1% 11.8% No €38.58 €72.01 3.8% 7.2% Sources: Authors’ calculations using eSWITCH version 4.6 run on 2019 SILC data uprated to 2022 terms. Note: Quintiles of equivalised income using modified OECD equivalence scale. Spending imputed using approach detailed in Appendix. The simulated price rise shown is that experienced between January 2021 and April 2022, plus additional 25 per cent. 50 | Energy poverty and deprivation in Ireland TABLE B.11 ESTIMATE OF HICP INFLATION BETWEEN JAN 2021 AND APR 2022, BY GROUP Household type Food Non-energy industrial goods Energy Services Total All 1.1 1.2 3.9 3.0 9.2 Income quintile Lowest 1.5 1.1 4.7 2.8 10.1 2 1.4 1.3 4.4 2.7 9.8 3 1.2 1.3 4.1 2.9 9.6 4 1.0 1.3 3.5 3.2 9.0 Highest 0.8 1.3 2.9 3.6 8.6 Dwelling location Rural 1.2 1.4 5.0 2.7 10.3 Urban 1.1 1.2 3.5 3.2 9.0 Housing tenure Homeowner 1.3 1.3 4.6 2.8 10.0 Mortgage 1.1 1.4 3.8 3.1 9.3 Renter 1.2 1.0 3.1 3.4 8.8 Sources: Authors’ calculations using Household Budget Survey and monthly Eurostat HICP following approach of Box B in McQuinn et al. (2022) and Lydon (2021). Note: Quintiles constructed equivalising income using modified OECD equivalence scale. Whitaker Patrons and Corporate Members The ESRI plays a leading role in producing independent research which allows policymakers in Ireland to better understand the economic and social landscape which shapes Ireland. One goal of the ESRI is to ensure that the Institute has a strong revenue stream to fund valuable, independent social and economic research initiatives that will have a long-term impact on Irish society. In 2021, we launched a compelling new membership programme with two categories of membership: 1) Whitaker Patron and 2) Corporate Membership We would like to acknowledge the following organisations and thank them for their valuable support: WHITAKER PATRONS EirGrid ICON Davy ESB Intel Microsoft Kingspan Meta CORPORATE MEMBERSHIP Musgrave Group OHM Group (Jaguar) Tesco Enterprise Ireland Matheson PwC ABP Food Group KPMG Ibec DCC Department of Transport FBD Insurance Department of Further and Higher Education, Research, Innovation and Science IDA Ireland Irish Water Greencoat Capital Malin Fáilte Ireland SIPTU Bord Gáis Energy Department of Agriculture, Food and the Marine Irish Farmers' Association Irish Life National Treasury Management Agency Whitaker Square, Sir John Rogerson’s Quay, Dublin 2 Telephone +353 1 863 2000 Email [email protected] Web www.esri.ie Twitter @ESRIDublin