UNDERSTANDING SOCIO- ECONOMIC INEQUALITIES AFFECTING OLDER PEOPLE

Transcription

1 UNDERSTANDING SOCIO- ECONOMIC INEQUALITIES AFFECTING OLDER PEOPLE PAUL MCGILL

2 Acknowledgements This research forms part of a programme of independent research commissioned by the Office of the First Minister and Deputy First Minister (OFMDFM) to inform the policy development process and consequently the views expressed and conclusions drawn are those of the author and not necessarily those of OFMDFM. Paul McGill Centre for Ageing Research and Development in Ireland June

3 CONTENTS Acknowledgements 2 Executive summary 5 Introduction 7 Research questions 7 What are socio-economic inequalities? 7 1. Analysis of CARDI research reports 8 CARDI ageing research 9 Part 1: References to inequality 10 Part 2: Poverty and disadvantage 12 Part 3: Research projects with statistics on inequalities 15 Summary Distribution and deprivation of older people in NI 23 Part 1: Spatial distribution of older people in NI 26 Part 2: Income Deprivation affecting Older People 28 Part 3: A local profile of deprivation among older people 31 Summary Income inequalities in Ireland, North and South 36 Part 1: Earnings 37 Part 2: Pay inequalities within the older age group 41 Part 3: Income inequalities 45 Part 4: Accounting for greater income inequality 49 Summary Policy background 53 Part 1: Broad policy context 54 Part 2: Public policy in Ireland 56 Summary 58 3

4 5. Implications for policy 59 Social inclusion 61 Care 61 Poverty 62 Health and well-being 62 Income inequalities 63 Social transfers/benefits 64 Occupational and private pensions 65 Spatial deprivation and income inequalities 65 Implications for data collection 65 Conclusion References 67 Appendix 1: Data on inequalities within the older population, NI and ROI 72 Appendix 2: Income deprivation among older people (NI) 74 4

5 EXECUTIVE SUMMARY There is long-standing evidence of the impact of inequalities in areas such as life expectancy, health, employment, education, housing and social inclusion. This debate has been particularly intense in the recent years of economic austerity. However, less attention in this field has been given to the dimension of age. This report explores socio-economic inequalities, with a particular focus on older people, and seeks to answer four main questions: 1 Are there inequalities that affect older people as a group compared with younger people, or inequalities that exist within the older population? 2 If so, how are these inequalities changing over time? 3 Do these socio-economic inequalities have a detrimental impact on older people or on a substantial number of them? 4 How can any harmful socio-economic inequalities be reduced or eliminated and what are the implications for policy-making? It uses data from both Northern Ireland and the Republic of Ireland to examine the total income of older people and particularly earnings from employment, social transfers/benefits and occupational pensions. It also presents findings on socio-economic inequalities from CARDI-funded research. Finally it analyses spatial data in Northern Ireland to examine the geographical distribution of disadvantaged older people. Key FINDINGS* In RoI the poorest older people had a rise of 32 per week between 2004 and 2011 in total incomes while those with the highest incomes had a rise of 255 (CSO 2013). Total incomes of the poorest pensioner couples in NI did not change between and but the best off had a rise of 37 per week (DSD 2013). Employees aged 60+ earn 10,000 less per year than earners in their peak years in RoI and 2,400 less in NI (CSO Database and NISRA 2012) The richest older people in RoI earn 14 times more from employment than the poorest. In NI it is 36 times more for single pensioners and 44 times more for pensioner couples (CSO 2013; NISRA 2013). The gap in weekly earnings between top and bottom earners aged 60+ in NI rose from 294 to 430 between 2005 and 2012 (NISRA 2012). In the two years the incomes of the poorest older people in ROI declined by 24 per week (11.4%) (CSO, 2013). * the richest older people refers to the highest fifth by income while the poorest older people refers to the lowest fifth by income 5

6 A considerable body of evidence points to a knock on relationship between income or social inequalities with wider inequalities in areas such as health and social inclusion and the evidence in this report suggests that inequalities are increasing in some aspects among our ageing population. Policy implications To improve the lives of older people, policy, practice and resource allocation should be made on the best information available. The research summarised in this report points to important policy conclusions in a number of areas, including: the need to improve the incomes and living conditions of the poorest people in order to reduce inequalities; the benefits of improved transport in promoting social inclusion; pay and taxation policies that provide incentives for improved pensions; Initiatives to tackle structural inequality and promote healthy ageing; Improved data collection, especially on a cross-border basis, to enhance mutual learning. 6

7 Introduction Research questions This report explores socio-economic inequalities and seeks to answer four main questions. 1 Are there inequalities that affect older people as a group compared with younger people, or inequalities that exist within the older population? 2 If so, how are these inequalities changing over time? 3 Do these socio-economic inequalities have a detrimental impact on older people or on a substantial number of them? 4 How can any harmful socio-economic inequalities be reduced or eliminated and what are the implications for policy-making? The Equality and Human Rights Commission and age sector groups in the UK (EHRC 2009) state that: Socio-economic inequalities are defined as inequalities that relate to differences in income, social class, occupational background, educational achievement and neighbourhood deprivation. These are distinguished from socio-demographic differences, which relate to factors such as age, gender, ethnicity, marital status, number of children, household composition and living arrangements. (EHRC 2009: 44) What are socio-economic inequalities? The EHRC (2009) states that, to date, policies have not focused on inequalities that result from class differences and other socio-economic factors. This is despite evidence that people of all ages from lower socio-economic groups tend to have poorer outcomes: for example, to be in poor health, drop out of school, be unemployed, live in poor housing and/or go to prison. The growing body of literature on health inequalities referred to later in this report shows that social class can affect everything from birth weight through health and disability to mortality rates. Poor people are more likely to have bad health in childhood and this is likely to persist right through the life cycle and to cause earlier death than for people who are well-off. Socio-economic inequalities will be examined in this report in three contexts. Chapter 1 examines evidence in Ireland, North and South, contained in research reports funded by the Centre for Ageing Research and Development in Ireland (CARDI). Chapter 2 carries out an analysis of spatial deprivation, specifically the proportion of older people in small areas of Northern Ireland with low incomes. Chapter 3 examines the earnings and incomes of older people in Ireland, North and South, including the impact of the first few years of the recession. Chapter 4 then sets out some of the general policy background to inequalities and outlines policy developments in Ireland, North and South. Finally, Chapter 5 discusses the policy implications with reference to the preceding chapters. 7

8 1. Analysis of CARDI research reports 8

9 CARDI ageing research This chapter summarises research projects funded by CARDI with a particular focus on inequalities related to socio-economic status, deprivation or income. CARDI funded 25 cross-border and inter-disciplinary studies, which reported in the years , on themes such as social inclusion, health and illness, poverty, care systems, dementia, physical activity, rural older people and pharmacy. These studies had a research focus that has a bearing on policy and practice, with the aim of improving the lives of older people, especially those who are disadvantaged. Thirteen of these reports with findings related to inequality are summarised in Chapter 1 as follows: Part 1 five have brief references to inequalities; Part 2 four dealing broadly with aspects of poverty, disadvantage or inequality; Part 3 four present statistics on a range of topics, including variables related to social class or areas of disadvantage. Panel 1.1 summarises the theme, topic, lead applicant or author and date of the 13 research reports which contained findings related to socio-economic inequalities. Panel 1.1: CARDI-funded research projects relevant to socio-economic, income and deprivation inequalities among older people Theme Topic Reference Social inclusion Public transport Ahern and Hine 2010 Rural social exclusion Walsh, O Shea and Scharf 2012 Care Telecare Delaney et al 2011 Future demand for long-term care Wren et al 2012 Poverty Impact of the recession Hillyard et al 2010 Living standards Hillyard and Patsios 2011 Fuel poverty Goodman et al 2011 Older women workers and pensions Duvvury et al 2012 Health and well-being Prostate cancer Donnelly et al 2012 Depression and physical activity Morgan et al 2011 Multimorbidity Savva et al 2011 Food and nutrition Bantry White et al 2011 Memory clinics Barrett and Savage

10 Part 1: references to inequality This part of Chapter 1 sets out the brief references to inequalities in five studies funded by CARDI. Public transport Public transport emerged as a critically important service in several research reports funded by CARDI (Ahern and Hine 2010; HARC 2010; O Sullivan 2011). Good transport provides links to essential services such as post offices, shopping and health services and offers opportunities for social inclusion through visits to friends, relations, social and recreational activities or opportunities to work or volunteer. The high cost of running a car, especially for older people who may be faced with increased insurance costs, introduces an element of inequality since many older people may be unable to afford a car. Community Technical Aid (2003) identified poverty and low income in NI as one of the four factors causing access problems for disadvantaged and socially excluded older people. This is supported by figures which show that in ROI retired people spend an average of per week on transport (CSO 2007). In NI the average spend of people aged is (ONS 2009). There are other factors operating as well: for example, many older women have never learned to drive and older old people may suffer disabilities that prevent them from driving. In this context it is notable that in NI the average amount spent on transport by people aged 75+ falls to (ONS 2009). Ahern and Hine (2010) suggested that it would be cheaper for the state to give taxi vouchers to older people than to introduce new bus and rail services. Access to telecare Telecare can be defined as the remote or enhanced delivery of health and social services to people in their own home by means of telecommunications and computerised systems. It uses information and communication technology to trigger human responses, or shut down equipment to prevent hazards (Delaney et al 2011). In ROI access to telecare was initially based on ability to pay, with the result that people with greater resources were better able to afford the services on offer, except where voluntary and community organisations were able to step in. However, there are examples where policy interventions can make access more equal. In ROI, for example, the Department of Community, Rural and Gaeltacht Affairs offered grants for the installation of social alarms to people who could not otherwise afford them. This initiative was replaced by the Seniors Alert Scheme in May It supports a monitored personal alarm, monitored smoke detectors, monitored carbon monoxide detectors, additional pendant, external security lights and internal emergency lighting. Grants are administered by community and voluntary groups (Delaney et al 2011). Recruitment by memory clinics Dementia refers to a group of diseases characterised by a progressive and generally irreversible decline in mental functioning, predominantly affecting people over the age of 65. It is a major social and economic challenge for countries with ageing populations as growing numbers of people are living to an age when dementia is likely to occur (Barrett and Savage 2010). In a study of people presenting with symptoms of dementia to memory clinics in Belfast and Dublin, social class was not one of the variables recorded although education level was logged. Barrett and Savage (2012) found, using education level as a proxy of social class, that Mercer s clinic in Dublin may have recruited more advantaged patients and, conversely, that the Belfast clinic may have been serving less advantaged people. They found that the Dublin clinic may have had greater success at detecting cases of mild cognitive impairment and Alzheimer s disease among higher socio-economic groups and conclude that, in Dublin: 10

11 Referrals to the service of better-educated people are being made with greater regularity and, taken together with changes in other demographics, it could be tentatively suggested that secondary health service provision for dementia in the ROI increasingly favours higher socio-economic groups and, possibly, males. (Barrett and Savage 2012: 26) Social exclusion Walsh et al (2012), researching ageing in rural communities, found that the term social exclusion is useful for encompassing aspects of disadvantage other than shortage of income or wealth. However, the term presumes a dividing line between an included majority and an excluded minority. This can have several negative effects, including obscuring inequalities and differences among the included; framing inequality and poverty as an aberration rather than an integral feature of capitalist societies; and preserving existing structural inequalities between those who are marginalised and those who are not (Walsh et al 2012). Long-term care The issue of how to provide and pay for care in the home and in residential settings is becoming a major issue. Understanding what the demand for care will be is a major part of this consideration. Wren et al (2012) did not include socio-economic variables as part of the study on the demand for long-term care, North and South. However, a literature review by Wren et al (2012) stresses the importance of factors such as socio-economic status, resources and house ownership in determining the need for and access to care. Furthermore, people in low socioeconomic classes are more likely to suffer poor health, whereas wealthier people are more likely to be in good health and better able to afford the supports needed to live independently and thus postpone or avoid admission to residential care (Wren et al 2012, citing Grundy and Jitlal 2007). Wren et al, citing Breeze et al (1999), found lower risk of admission to long-term care among owner-occupiers as compared to renters. Furthermore, they found that men in rented accommodation had a 90% excess risk of institutionalisation and women a 40 45% excess risk. Partly, this may be because people who own their own homes are wealthier and more likely to be in good health than renters. It may also be partly explained by the deterrent effect of means testing, namely that housing assets are taken into account when deciding on payment for care. McCann et al (2011) agree that home ownership may act as a disincentive to enter residential care. They showed that the main difference in admission rates is between those who rent and those who own their homes but that the value of the house makes little difference. Families may choose to increase the amount of informal care-giving (and perhaps paid care) to prevent the sale of the home. 11

12 Part 2: poverty and disadvantage This section deals with the four reports funded by CARDI which had a strong focus on disadvantage and poverty, though not necessarily on socio-economic inequalities: Hillyard et al 2010; Hillyard and Patsios 2011; Goodman et al 2011 and Duvvury et al Poverty and the impact of the recession on older people Hillyard et al (2010), following detailed analysis of social datasets in Ireland, North and South, uncovered several factors causing differentials within the older population and found substantial differences between NI and ROI. First, NI has much higher rates of pensioner poverty than ROI. Second, the two jurisdictions are going in different directions, with a rapid decline in the proportion of pensioners in poverty in ROI and an increase in NI. Third, the actual numbers of both single pensioners and pensioner couples living in poverty have increased. Fourth, older people who live alone are at much greater risk than those who live with someone else. Fifth, a far higher number of women pensioners are likely to be in poverty than men and the older women get, the more likely this is to be the case. Figure 1.1: risk of poverty (60% median) among older people in NI and ROI (%) NI BHC NI AHC ROI Source: EU-SILC (CSO 2013) and Poverty in Northern Ireland 2010/11 (NISRA 2012b) Note: older people means 65+ except for women in NI, where it means

13 Since Hillyard et al (2010) collected their data (the latest year covered was 2009), the official poverty risk among older people has declined, as shown in Figure 1.1. With one exception, the risk of poverty among older people in ROI declined each year from 2003 to 2010, followed by a rise in In NI, the rate of poverty after housing costs declined from 2008 onwards and before housing costs a year later. In part the decline in poverty rates may be because average incomes in the population declined, especially in ROI, with the result that the benchmark figure of 60% of median income also declined; since the state pension was not reduced in ROI and was increased in NI, the result was that more older people ended up above the 60% figure. Hillyard et al (2010) argued that tax policies do not promote equality but appear to benefit those who are well-off. One example is the differential tax relief given for pension contributions, which is heavily skewed towards the better off. UK Treasury figures show that of the 37 billion of tax relief, some 60% goes to higher-rate taxpayers, with 25% nearly 10bn a year going to the top 1% of earners. In ROI Callan et al (2009) found that changes to tax relief on pensions would save public money and be fairer, and that tax relief at a standard rate could promote a more efficient and equitable public pension policy. Currently, over 8 out of every 10 of tax relief goes to taxpayers in the top fifth of the income distribution. Moloney and Whelan (2009) have noted that the pension subsidy is no more than 22% for persons in the first seven income deciles (incomes under 34,000 per annum) but is 34% for people on high incomes. Therefore the tax subsidy is highest for the top earners, in both percentage terms and overall amounts. Hillyard et al (2010) stated that taxation policies have boosted pension provision among the well-paid to a greater degree than among those on smaller salaries: The inequality arises not only because of inequality in salaries but also because high income earners are more likely to participate in pension schemes, more likely to make higher contributions, and the value of tax relief at the top rate of income tax is about double that for the standard rate taxpayer. There is a strong incentive for high earners to contribute to pension schemes, but a weaker incentive for those with low and middle incomes. (Hillyard et al 2010: 63) Living standards of older people Hillyard and Patsios (2011), in a study of living standards, noted that, on average, older people were better off than younger people in Ireland, both North and South, on most measurements of deprivation. However, in contrast to the reduction in at risk of poverty rates among older people, they found the recession is having a very real impact on all households, including both single pensioners and pensioner couples. There has been an increase in the proportion of single pensioners in NI unable to keep their house warm from 4.0% in 2007 to 5.5% in In ROI the proportion has nearly doubled from 2.4% in 2007 to 4.7% in In NI, the proportion of pensioner couples struggling to heat their home has more than trebled from 2.2% to 6.9%, while in ROI the increase has been more modest. The proportion of single pensioners in NI who cannot afford an annual holiday has gone down slightly, but there has been a large increase in ROI, from 13.3% in 2007 to 40.5% in The proportion of pensioner couples who were unable to afford an annual holiday rose from 14.8% to 19.3% in NI and from 14.9% to 26.3% in ROI. Hillyard and Patsios (2010) state that while many of the inequalities in older age are structured in terms of class, gender and ethnicity, there are other cross-cutting inequalities between those with occupational pensions and those without, between employees and the selfemployed, and between those with public sector pensions and those on state pensions only. 13

14 Fuel poverty Goodman et al (2011) argued that fuel poverty is caused by the interaction of high fuel prices, low income and poor energy efficiency in the home. Following a multi-methods research project, their main findings were that older people experience a dual burden. They are more likely to suffer fuel poverty and, because older people are more likely to have chronic diseases, they are also particularly vulnerable to poor health and social harm as a result. Low temperatures do not just have an immediate acute effect; a study in Dublin showed that a one-degree Celsius drop in temperature is associated with a 2.6% increase in deaths among older people over the following 40 days (Goodman et al 2004). The numbers of older people vulnerable to ill-effects from cold homes, North and South, will increase due to increases in the number of people aged 80 and over and those living with chronic illness or disability. Goodman et al (2011) note that in ROI between one-fifth and one-quarter of older people are in fuel poverty (that is, spending more than 10% of their disposable household income after tax on fuel and electricity) and this was true of over 35% of older people living alone. In NI 44% of households were in fuel poverty in 2009: 53% where the household reference person is aged and 76% where he/she is aged Older households account for a disproportionately large share of all households in fuel poverty; households where the head is an older person accounted for nearly half of all fuel-poor households in Northern Ireland in Moreover, older people tend to underestimate the problems facing them, so that their self-reported subjective measures of fuel poverty and levels of debt/arrears should be interpreted with caution (Goodman et al 2011). Several factors help to determine the extent of fuel poverty, including living in a private household; tenure; whether the home is a detached or semi-detached house, a terraced house or a flat; age and condition of the building; and the type of fuel used. One important factor is whether or not the resident lives alone; about a quarter of older people living alone in ROI and a third in NI face particular challenges in heating a home on a single income due to diseconomies of scale. The situation where a single older person may still occupy a family home with many vacant rooms is especially noteworthy (Goodman et al 2011). Older women and pensions Duvvury et al (2012) studied why older women workers have lower pension provision than men. The research illustrates the interaction of different forms of disadvantage over the life course. It found that women are at a disadvantage as a result of a male breadwinner model, especially because of absence from the paid workforce to raise children and take on other caring duties. Women in low-paid, temporary work lose out most because they often cannot afford to make pension contributions. Most older women depend heavily on the state pension: in NI nearly 56% of female interviewees in employment and 82% of those who were not relied solely on the state pension (Duvvury et al 2012, citing Evason and Spence 2002). This suggests the need to ensure that the basic state pension is adequate. Duvvury et al (2012) found that the current emphasis on occupational and personal pensions tends to reinforce the link between the pension system and earnings, length of service and employment status. Reinforcing this link is likely to exacerbate gender inequality in pension provision, given that women typically have lower earnings and interrupted employment records (Duvvury et al 2012: 9). Opportunities to earn income across the life course largely determine the amount of pension a worker can accumulate (Duvvury et al 2012, citing Giele and Elder 1998). Many factors influence pension-building behaviour and therefore wealth in older age, such as women s socio-economic class, parental influence at the outset of working life, level of education, internalised gendered caring norms, legislation and economic conditions (Duvvury et al 2012: 63). 1 By 2011, fuel poverty in households aged 75+ had declined to 66%; among those aged it was largely unchanged (NIHE 2012). 14

15 Part 3: research projects with statistics on inequalities This part deals with four reports which present statistics on a range of topics, including variables related to social class or areas of disadvantage. Prostate cancer Donnelly et al (2012) 2 investigated prostate cancer, the most common cancer in men in Ireland. They used data from 1996, 2001 and 2006 on rates of prostate cancer among men aged 70+ compared with younger men, with extent of deprivation of the area in which the patients lived as one of the variables. The areas were divided into five equal parts (quintiles) according to their degree of disadvantage based on the 2005 multiple deprivation measure in NI (NISRA 2005) and using five domains derived from the National Deprivation Index in ROI (Kelly & Teljeur 2004). An unusual pattern emerged in which NI and ROI were almost mirror images of one another (Figure 1.2). In NI the lowest rates of prostate cancer are found among people living in the least and most deprived areas and higher rates apply in the three middle quintiles. In ROI people in the least and most deprived areas have higher rates of the disease whereas lower rates exist in the three middle quintiles. Figure 1.2: men diagnosed with prostate cancer in 2006 by area deprivation (%) Donnelly et al 2012: 25 NI ROI Least deprived Most deprived 2. Donnelly was the main author but the project was led by Dr Anna Gavin at the Northern Ireland Cancer Registry. 15

16 However, the researchers indicate that there may not be a genuine socio-economic deprivation difference since the results could be an artefact of how deprived areas are defined in the distinct measures of deprivation in NI and ROI or could reflect differences in the two systems or the profile of men taking Prostate-Specific Antigen tests (Donnelly et al 2012: 33). Figure 1.3 shows that men in Ireland as a whole with prostate cancer were more likely to see a urologist in 2006 than was the case a decade previously. Moreover, in 2006, but not in 1996, there was an unbroken deprivation gradient: men from the least deprived areas were most likely to visit the specialist (91%) and the proportion declined for each level of deprivation, to 79% for men from the most deprived areas. Figure 1.3: NI and ROI men with prostate cancer who saw urologist in 1996 and 2006 by area deprivation (%) Donnelly et al Least deprived Most deprived 16

17 Important differences exist North and South. Between 1996 and 2006, the estimated proportion 3 of men seen by a urologist increased markedly in NI from 59% to 94%, but remained stable in ROI at approximately 81% (Donnelly et al 2012). Moreover a multivariate analysis revealed marked variations, as highlighted in Figure 1.4. Figure 1.4: estimated probability of seeing urologist by area deprivation (1996, 2001 and 2006 data combined) Donnelly et al 2012 NI ROI Least deprived Most deprived There is no difference in NI in the likelihood of men from the most and least disadvantaged areas seeing a urologist; men from the middle group are least likely to do so. In ROI, by contrast, there is a clear gradient, with men from the least deprived areas more likely to visit the specialist (86%) than men from the most deprived areas (76%) (Donnelly et al 2012). In 2006 men in ROI with prostate cancer were twice as likely (16.8% compared with 7.3%) to undergo radical prostatectomy as men in NI. Combining data for 1996, 2001 and 2006, men from the least deprived areas were 77% (or 1.77 times) more likely to have radical prostatectomy than men in the most deprived areas. There has been a big rise in the proportion of men receiving radical radiotherapy in the decade NI men living in the most deprived areas had a lower probability of receiving the treatment about 7% compared with more than 11% of men in the best off areas. Donnelly et al (2012) comment on these differences: Men in more socio-economically deprived areas in NI were less likely to receive radiotherapy. However, it should be noted that, after adjusting for other variables, men in NI s worst socio-economically-deprived group were receiving the same level of radiotherapy as men in ROI. It is not clear why there would be a deprivation gradient in NI as the NHS is a free system, equally accessible to all, but perhaps cultural attitudes determine uptake of treatment. (Donnelly et al 2012: 52). 3. Missing values were imputed, as explained in Donnelly et al 2012:

18 Depression and anhedonia Morgan et al (2011) undertook a data-mining study of depression and anhedonia (the inability to derive pleasure from activities). They analysed data about older people from SLÁN, the Survey of Lifestyle, Attitudes and Nutrition 2007 (DoH 2008) in ROI, which had 4,255 adults aged 50+, and the Northern Ireland Health and Social Wellbeing Survey (NIHSWS) , which had 1,904 adults aged 50+ (NISRA 2007). The variables included socio-economic status (SES), which can be combined into three groups as follows: I. SES 1 & 2 (professional, managerial and semi-professional): high II. SES 3 & 4 (lower non-manual and skilled manual): mid III. SES 5 & 6 (semi-skilled and unskilled manual): low. Depression in older adults is not just a serious illness but is also associated with increased risk of morbidity, suicide and self-neglect and it decreases physical, cognitive and social functioning. Low SES has been found to be correlated with a higher prevalence and incidence of depression (Morgan et al 2011, citing Everson et al. 2002). ROI respondents in high SES reported less psychological distress than those in low classes. This was also true in NI but the differences were not statistically significant. The results are shown in Figure 1.5, which also makes clear that rates of depression among people aged 50+ are higher in NI in each SES (Morgan et al 2011). Figure 1.5: people aged 50+ with depressed mood and/or anhedonia by social class (%) Morgan et al 2011 SES 1&2 SES 3&4 SES 5& ROI NI 18

19 A logistic regression undertaken by the researchers showed no SES association in NI, unlike ROI where the probability of people aged 50+ in low SES having depression and/or anhedonia was 43% higher than for high SES. Morgan et al (2011) concluded that depressive symptoms in ROI were negatively associated with male gender, older age groups, being a former smoker and moderate to high levels of physical activity, while depressive symptoms were positively associated with low social class and being a current smoker. In NI, depression was negatively associated with older age categories, being a former or current smoker and higher levels of physical activity. Being widowed was positively associated with elevated depressive symptoms. Multimorbidity Savva et al (2011) used SLÁN 2007 (DoH 2008) and NIHSWS 2005/6 (NISRA 2007) to study the causes and consequences of multimorbidity (having more than one chronic condition) in older people in Ireland. The literature review explains why SES is relevant to multimorbidity: Given the long-standing association between income inequality and incidence of chronic illness and the social determinants of health, it is important to advance the understanding of the impact of socio-economic status on health outcomes for older people with multiple chronic illnesses. The increasing gap in the health outcomes of older people across the socio-economic gradient will become an important public health issue, especially as the population ages. (Savva et al, 2011: 7) Savva et al (2011) estimated the relative risks of having one of eight chronic conditions 4 and also the risk of having two or more of these conditions. The risk of people in mid and low SES was compared with the risk of people in high SES, as shown in Figures 1.6a (one chronic disease) and 1.6b (multimorbidity). Figure 1.6a: Risk of mid and low SES having one chronic disease compared with high SES Savva et al (2011) Figure 1.6B: Risk of mid and low SES having 2+ chronic diseases compared with high SES Savva et al (2011) Low ROI NI Low ROI NI Mid Mid Note: in ROI risk of mid SES group having multimorbidity was the same as for the high SES. 4 These are heart attack, angina, stroke, diabetes, asthma, chronic obstructive pulmonary disease, musculo-skeletal pain (including rheumatism, arthritis and back pain) and cancer. 19

20 People in ROI in the low SES category are 13.5% more likely to have a chronic condition than people in high SES but in NI older people in low SES are nearly 70% more likely to have a chronic condition compared with those in high SES. Among older people with multimorbidity, the difference between the high and low SES increases; in ROI the gap grows to 64.5% and in NI to 117% (Savva et al 2011). There is therefore a socio-economic gradient in both parts of Ireland but it is considerably stronger in NI, where those in the low SES group are 2.2 times as likely to report multimorbidity as those in the high group. In ROI, the low SES group is only 1.6 times as likely to report multimorbidity as the high group. The data showed an unbroken social class gradient in both NI and ROI in the likelihood of reporting health as only poor or fair; in limitations of daily activity; and in having poor or fair quality of life: i.e. people in low SES were more likely to report worse outcomes on these three variables than people in higher SES. In all cases, the social class difference among NI respondents was greater than in ROI. Savva et al (2011) analysed the proportion of older people reporting fair or poor health by number of chronic diseases (Figures 1.7a and 1.7b). Again NI displays a clear social class gradient and respondents reported worse health than in ROI. People with two or more chronic conditions are much more likely to report that their health is only fair or poor compared to people with one or no chronic conditions. In ROI, among older people with multimorbidity in the highest SES, 53% reported fair or poor health but this was true of 74% of people in the lowest SES groups who had multimorbidity. Among older people with multimorbidity in NI, 75% of those with high SES said their health was poor or fair compared with 89% of those with low SES. Figure 1.7a: ROI older people reporting Savva et al 2011 Figure 1.7b: NI older people reporting Savva et al 2011 poor or fair health by morbidity and SES (%) poor or fair health by morbidity and SES (%) High Medium Low 74 High Medium Low No chronic disease 1 chronic disease 2+ chronic diseases No chronic disease 1 chronic disease 2+ chronic diseases 20

21 Savva et al (2011) found that many of the diseases investigated were more common in those from lower SES classes, both in NI and ROI. However, the social class gradient with respect to multiple chronic conditions is significantly stronger in NI. They indicate that the effect of SES on quality of life is largely explained by health status; after adjusting for disease and disability there is only a small independent effect of SES in reporting poor quality of life and it appears that this effect is restricted to those with multimorbidity, though this is difficult to test due to the small numbers. If this is the case it would appear that the socio-economic effect on health- related quality of life has two distinct components, first in the occurrence of chronic disease, possibly influenced by life-life course factors associated with low SES, and second by affecting the support available to people with disability and multimorbidity. (Savva et al 2011: 28) Food and nutrition Good nutrition is an important element of health and is essential to adequate functioning and quality of life for older adults. Malnutrition in older age is a significant public health problem which often goes undiagnosed. It is associated with outcomes such as prolonged hospitalisation and rehabilitation, infection, pressure ulcers, poor wound healing, reduced cognitive function, impaired muscle function and mortality (Bantry White et al 2011). In NI, an estimated four out of ten older people who are admitted to hospital are suffering from malnutrition on arrival. Patients over the age of 80 admitted to hospital have a five times higher prevalence of malnutrition than those under the age of 50. Six out of ten people are at risk of becoming malnourished (DHSSPS 2007). In ROI, it is estimated that 70,000 Irish people over 65 years of age may be either malnourished or at significant risk of malnutrition (UCD 2010). Bantry White et al (2011) analysed data from the Household Budget Survey (CSO), which covered 6,884 households, including 1,444 over 65, and the Expenditure and Food Survey (NISRA) with 533 households in NI, 144 over 65. This study included an income variable with respondents grouped into three categories: below 60% median income (at risk of poverty); between 60% and the median; and above the median. Bantry White et al (2011) found that both NI and ROI older households, especially aged 75 and older, spend less on food than any other age group. 5 The biggest spenders are aged in ROI and in NI. Much of this difference is related to household size since middle-aged groups are more likely to have children and, by comparison, the 75+ cohort includes many people living alone, mostly women. Another feature is that the weekly food spend in ROI is a lot higher than for NI in every age group; in both the and 75+ age groups, people in ROI spend 40% more than in NI. 5. Amounts were converted using the 2005 exchange rate of 1 equals

22 Spending covers all categories included in the survey, not all of them related to good nutrition: for example, spending on alcohol and tobacco may harm health and welfare. Households in ROI spend a lot more on this than in NI and young people far more than older people. Households in ROI aged spend an average of 82 per week on alcohol and tobacco and those aged spend 54, compared with 28 in the age range and 17 among people aged 75+. Similarly in NI, households in the two youngest age groups spend 29 and 27 respectively on alcohol and tobacco compared with 20 for those aged and 13 for those aged 75+. Gender differences exist as well. In single older households, men in ROI spend three times more on alcohol and tobacco than women and in NI they spend twice as much (Bantry White et al 2011). Bantry White et al (2011) found that: Total food spending of the high income group is double that of the low group in ROI but only one-third higher in NI. In ROI the greatest gap is in meals out, where the high income group spends four times more than the low group; in NI it is 85% more. In ROI there is a clear income gradient in spending on alcohol and tobacco; in NI the middle group, not high earners, spends most. In NI and ROI there is an income gradient in spending on both fruit and vegetables (Bantry White et al 2011). In NI there is little difference in expenditure on vegetables, butter, fat and oils as a proportion of all food spending. Low income households allocate 11.7% of their budgets to snacks and confectionery, compared with 7.7% for high earners; by contrast high income families spend a greater share of their food budgets on fruit than mid or low income households. Bantry White et al (2011) suggest that high-earning families in NI spend more on healthy food items and less on unhealthy ones than lower-income families. Summary In the context of the questions posed in the introduction, research funded by CARDI either cited or uncovered many examples of inequalities affecting older people. These include socio-economic differences in access to transport, memory clinics and long-term care; marked differences in income from pensions and other sources; greater fuel poverty and worse nutrition among poorer older people. Older people in higher social groups or least deprived areas were more likely to receive treatment for prostate cancer, less likely to have multimorbidity or a low quality of life and had the highest spending on food. Limited evidence exists on change over time, but poverty rates of older people have declined in recent years. The examples of inequalities cited relate to issues that can have a seriously detrimental impact on many older people s health, life expectancy, social inclusion and quality of life. The subsequent questions of how the inequalities can be reduced or eliminated and the implications for policy-making are considered in Chapter 5. 22

23 2. Distribution and deprivation of older people in NI 23

24 This chapter considers two important and related issues for policy-makers. The first is whether deprivation among older people can be tackled effectively through spatial strategies, as distinct from policies aimed at the individual. Secondly, if spatial strategies are appropriate, how best can they target the largest possible number of disadvantaged older people? This chapter focuses on NI due to the lack of official data in ROI. Part 1 of this chapter shows how the population of older people (women aged 60+ and men 65+) is distributed across NI. Part 2 gives a preliminary analysis of the relationship between Income Deprivation affecting Older People (IDAOP) and the more commonly used Multiple Deprivation Measure (MDM); and the health domain. It then looks at the distribution of older people according to IDAOP rankings and compares these with where older people live, as shown in Part 1. Part 3 profiles some of the most and least deprived Super Output Areas (SOAs) according to IDAOP to see what particular characteristics the different areas exhibit, which might give an indication of the best means to go about reducing inequalities. This analysis is based on MDM (NISRA 2010b), annual estimates of mid-year populations (NISRA 2011) and data on the Northern Ireland Neighbourhood Information Service (NINIS various dates). First MDM is described in Panel 2.1. Panel 2.1: NI Multiple Deprivation Measure 2010 The Multiple Deprivation Measure (MDM), published by NISRA, has been used for many years to measure spatial disadvantage in NI. It is made up of seven domains, which contribute to the overall MDM as follows: income deprivation 25% employment deprivation 25% health deprivation and disability 15% education, skills and training deprivation 15% proximity to services 10% living environment 5% crime and disorder 5% NISRA also published, along with the MDM, an index of Income Deprivation affecting Older People (IDAOP). The scores for income can be interpreted as the percentage of the relevant population that is deprived. The various indicators do not have uniform scales. Excluding Aldergrove 1 because it has no older people, scores for IDAOP range from 0.03 to MDM scores run from 1.65 for the least deprived area to for the most deprived; health scores range from 3.13 in the most deprived area to 3.10 in the healthiest area (NISRA 2010b). The most deprived SOA is ranked 1st and the least deprived is ranked 890th. See for full information 24

25 Several indicators are relevant to older people but only IDAOP refers specifically to them. This indicator merits further investigation as a means of identifying the spatial distribution of potentially vulnerable older people. A benefit of IDAOP is that it is reported for each of the 890 SOAs, which have an average total population of 2,000 (NISRA 2010b), of whom an average of 346 were women aged 60+ or men aged 65+ in 2010 (NISRA 2011). See panel 2.2 for a definition of IDAOP. Panel 2.2: Income Deprivation affecting Older People IDAOP is a stand-alone measure published with the MDM combining the percentage of a Super Output Area s population aged 60+ and their partners (if 60+), living in households in receipt of Income Support, State Pension Credit, income-based Jobseeker s Allowance, income-based Employment and Support Allowance, Housing Benefit, Working Tax Credit or Child Tax Credit. Tax credit claimants are included only where the equivalised income is below 60% of the NI median before housing costs. Scores represent the proportion of older people judged to be deprived e.g. a score of 0.92 means that 92% of over-60s in that area are income-deprived (NISRA 2010b). A limitation is that IDAOP is not reported at Output Area level (of which there are 5,022 with an average population of 360), which means it is not possible to identify small pockets of deprivation among older people. This is likely to affect rural areas in particular. A study of the composite MDM showed that no rural areas were in the top 10% most deprived SOAs in 2010 but that 15 rural areas featured in the most deprived 10% of Output Areas, indicating that smaller geographical units aided the identification of deprivation in rural areas (NISRA 2010c). 6 Men aged ,435; women aged ,885; total 308,

26 Part 1: spatial distribution of older people in NI In this part, spatial distribution of older people in NI at District Council and SOA level is studied more closely. In 2010 in NI as a whole, 17.1% of the population was made up of women aged 60+ and men aged 65+ (NISRA 2011). Older people were not evenly distributed across the 26 districts in NI, however, as is clear from Table 2.1. The proportion 7 of older people ranges from 14% in Derry to 21.7% in North Down. In Coleraine, Ards and Castlereagh older people constitute one-fifth or more of the inhabitants. By contrast, less than 15% of the population in Limavady, Dungannon, Newry & Mourne and Magherafelt are older people. North Down also has the highest proportion of people aged 85+, who make up 2.6% of its total population, followed by Castlereagh (2.2%). At the other end of the spectrum, only 1.1% of the population of Derry and Newry and Mourne are aged 85+. Table 2.1: NI Local Government Districts by number and percentage of older and very old people NISRA 2011 District Total all ages All 60/65+ Aged 85+ % 60/65+ % 85+ North Down 79,900 17,300 2, Coleraine 56,800 11,500 1, Ards 78,200 15,800 1, Castlereagh 67,000 13,500 1, Larne 31,700 6, Moyle 17,000 3, Ballymena 63,500 12,400 1, Newtownabbey 83,600 15,700 1, Carrickfergus 40,200 7, Belfast 268,700 47,200 5, Fermanagh 63,100 11,000 1, Ballymoney 30,600 5, Down 70,800 12,000 1, Armagh 59,400 10, Strabane 40,100 6, Lisburn 117,800 19,100 1, Banbridge 48,000 7, Craigavon 93,600 15,000 1, Antrim 54,100 8, Cookstown 36,700 5, Omagh 52,900 8, Limavady 33,600 4, Dungannon 57,700 8, Newry & Mourne 99,900 14,500 1, Magherafelt 44,700 6, Derry 109,800 15,400 1, Northern Ireland 1,799, ,400 29, Note: figures have been rounded to the nearest The proportion of older people is used because districts vary by size e.g. Belfast has almost 16 times more inhabitants than Moyle. Table 2.1 also gives the number of older people in each District. 26

27 Small area level If smaller areas such as SOAs are used to examine populations, the extremes are far greater. In three of the North Down SOAs, more than one-third of the population are older people and the same is true of Donaghadee South 2 in Ards. At the other extreme (and excluding the Antrim SOA which consists of an army base), there are four areas with less than 4% older people. If we rank all the SOAs according to their proportion of older people, we find that 15 of them fall into North Down and 13 into Belfast (which is by far the largest of the 26 districts). Close behind are Newtownabbey with 12, Coleraine with 11 and Ards with 10. The pattern is for concentrations of older people in districts in and around Belfast and, to a lesser extent, on the north coast. By contrast, there is a wedge of eight districts, running north-west from the triangle of Armagh-Craigavon-Banbridge through Dungannon, Cookstown and Magherafelt and west to Omagh and Strabane, which have no SOAs in which there are high proportions of older people. Some of the wide variation in the distribution of older people can be seen in Table 2.2, which lists the 25 SOAs with the highest and lowest proportions of older people. Table 2.2: SOAs by highest and lowest percentage of older people Source: NISRA (2011) 25 Highest 25 Lowest SOA District % OP SOA District % OP Churchill 1 North Down 37.8 Culmore 5 Derry 7.3 Groomsport North Down 35.3 Collin Glen 1 Lisburn 7.2 Donaghadee South 2 Ards 34.4 Kilwaughter 1 Larne 7.1 Princetown North Down 33.4 Culmore 3 Derry 7.1 Upper Malone 1 Belfast 32.7 Glencolin 1 Belfast 6.9 Carnmoney 2 Newtownabbey 32.5 Collin Glen 3 Lisburn 6.8 Gardenmore Larne 32.4 Shantallow West 1 Derry 6.8 Broadway 2 North Down 32.2 West 1 Strabane 6.3 Bryansburn 2 North Down 32.0 Cairnshill 1 Castlereagh 6.2 Springhill 2 North Down 31.8 Crevagh 2 Derry 6.2 Crawfordsburn North Down 31.6 Stranmillis 2 Belfast 5.9 Fortwilliam 1 Belfast 31.6 Shantallow West 4 Derry 5.8 Boneybefore Carrickfergus 31.0 Botanic 2 Belfast 5.3 Abbey 1 Newtownabbey 30.9 Crevagh 3 Derry 5.1 Bryansburn 1 North Down 30.9 Shantallow West 3 Derry 5.1 Mount Sandel Coleraine 30.8 Ballycrochan 2 North Down 5.0 Bradshaw's Brae 2 Ards 30.7 Botanic 4 Belfast 4.9 Crumlin 1 Belfast 30.4 Mallusk 2 Newtownabbey 4.9 Bangor Castle North Down 30.4 Loughview 2 North Down 4.8 Cherryvalley 1 Belfast 30.2 Derryaghy 1 Lisburn 4.6 Knockbracken 2 Castlereagh 29.9 Botanic 1 Belfast 4.4 Stormont 1 Belfast 29.6 Collin Glen 2 Lisburn 3.8 Cultra North Down 29.5 Botanic 3 Belfast 3.7 Burnthill 2 Newtownabbey 29.2 Derryaghy 2 Lisburn 3.2 Lagan Valley 2 Lisburn 29.0 Gresteel 2 Limavady 3.0 Note: older people (OP) means women aged 60+ and men aged

28 In the most extreme case, the proportion of older people in Churchill 1 in North Down is 12 times greater than in Derriaghy 2 and Gresteel 2. Overall, there are seven times more people aged 65+ in the 25 SOAs with the highest proportions of older people (31.6%) than in the 25 SOAs with the lowest older populations (4.2%). Gender In 2010 in NI as a whole 36.5% of older people were men and 63.5% were women, though the female group includes everyone aged 60+ whereas the male group is 65+ (NISRA 2011). There are marked variations in the gender balance at SOA level with the proportion of older women compared with older men ranging from 56.1% in Gransha, Banbridge, to 75.2% in Shantallow West 2 in Derry. The number of women aged 60+ as a proportion of all women averages 21.4% in NI as a whole but ranges greatly from 4.3% in Gresteel in Limavady to 47.4% in Churchill 1 in North Down. In the case of men aged 65+ as a proportion of all men, the NI average is 12.7%, ranging from 1.7% in Gresteel to 27.3% in Groomsport in North Down (NISRA 2011). Part 2: Income Deprivation affecting Older People This section studies the relationship between IDAOP, MDM and the health domain and the spatial distribution of older people according to IDAOP rankings. Multiple deprivation measure The MDM is the best known of the indicators of spatial disadvantage in NI and is used by government departments to identify deprivation. It is closely correlated by rank with the IDAOP (.9003 for all 890 SOAs) but the correlation for the 100 most deprived areas according to IDAOP is lower at.4365 (source: the author). There are several SOAs which have very different rankings on the two measures. For example, Culmore 4 in Derry has very high income deprivation among older people, with a rank of 23, but a much lower MDM rank of 149 (NISRA 2010b). Gresteel in Limavady has an IDAOP rank of 94 but it is 564 places further down in the MDM rankings at 658. Conversely, some areas which are high on the IDAOP measure are much lower on the MDM score. An example is Woodvale 1 SOA in Belfast, which is placed as 267th in IDAOP but on a much higher ranking of 57 on MDM. Likewise Ballymaglave in Down is nearly 300 places lower on IDAOP than on MDM (NISRA 2010b). Table 2.3 gives details of these and other examples of disparity between the indices. Table 2.3: selected comparisons of IDAOP and MDM scores and ranks Source: NISRA 2010b SOA District IDAOP score IDAOP rank MDM score MDM rank Coalisland South Dungannon Culmore 4 Derry Forkhill 2 Newry & Mourne Gresteel 2 Limavady Woodvale 1 Belfast Antiville Larne Scrabo 1 Ards Ballymaglave 2 Down

29 In summary the significance of such disparities is that if government departments and official bodies use the MDM scores and ranks to identify and target deprivation on a geographical basis, they may be missing several areas in which there is a high degree of income deprivation among older people. Examples of these are anti-poverty and neighbourhood renewal policies (see Chapter 5). Spatial strategies to combat inequalities in the general population or among older people in particular need to take account of the specific forms of deprivation they experience in order to be effective. Health deprivation Health deprivation and disability forms one of the seven domains of the MDM and accounts for 15% of the total weight. The health ranks correlate very strongly with IDAOP ranks across the 890 SOAs (.8830) but the correlation is weaker for the 100 most deprived SOAs on the IDAOP index (0.4248) (source: the author). Table 2.4 sets out some examples of where SOAs with high levels of IDAOP deprivation had much lower levels of deprivation (lower ranks) in the health domain. Feeny, for example, has a health ranking 375 places lower than its IDAOP rank. Twothirds of older people in Gresteel 2 have low incomes but they are living in an area which scores quite well in health deprivation ( 0.64) and overall deprivation (9.95). Table 2.4: selected comparisons of IDAOP and Health Deprivation & Disability Source: NISRA 2010b SOA District IDAOP score IDAOP rank MDM score Crossmaglen Newry & Mourne Ardboe Cookstown Culmore 3 Derry Forkhill 2 Newry & Mourne Feeny Limavady Gresteel 2 Limavady Older populations and IDAOP In summary, the evidence suggests that policies and initiatives that use the MDM or health scores to identify the most deprived areas are likely to overlook several SOAs with high levels of IDAOP. A further difficulty arises in targeting large numbers of income-deprived older people, namely that the areas with the highest levels of IDAOP tend to have the smallest numbers of older people living in them. Conversely, the areas with the largest concentrations of older people tend to have few low-income older people. If we divide the 890 SOAs into ten groups according to the size of their older populations (women 60+ and men 65+) (NISRA 2011), the first decile (with the highest populations) includes only two of the most deprived SOAs according to IDAOP and the next decile includes only four. 8 By contrast the ninth-lowest population decile contains 24 of the most low-income SOAs and the lowest-population decile contains 19 of them. If we plot all 10 deciles according to the number of older people, from largest to smallest older population, the average IDAOP score increases as population declines (see the trend line in Figure 2.1). In other words, having a lower older population in an area tends to be associated with a higher degree of income deprivation among older people. This means that spatial strategies directed towards the most deprived IDAOP SOAs will reach relatively few deprived people. The difference is substantial. The highest decile has an average of nearly 53,000 older people whereas the lowest one has only 13, See Appendix 2 for fuller figures and a more detailed explanation of this methodology. 29

30 Figure 2.1: average IDAOP score by size of older population by SOA in deciles Source: author, based on NISRA 2010b and 2011 Av IDAOP score Linear (Av IDAOP score) y = 0.013x R 2 = Note: 1st decile represents the SOAs with highest populations and 10th decile represents the SOAs with the lowest populations. Figure 2.2 illustrates the districts which have the highest numbers of SOAs with high income deprivation among older people. Whereas North Down includes 15 of the most populous SOAs by older population, it contains none of the SOAs with the highest rates of IDAOP. Forty-four of the most income-deprived SOAs are in Belfast and 18 in Derry. The next-largest numbers are in Craigavon, Lisburn and Newry and Mourne. This contrast between population and IDAOP strengthens the point that spatial policies and initiatives aimed at deprived older people may reach relatively small numbers of them initiatives aimed at areas with high proportions of older people, such as North Down, would reach few who are income-deprived whereas initiatives aimed at the most deprived according to IDAOP would target relatively small numbers of older people. Figure 2.2: NI Districts by SOAs in 100 most deprived by IDAOP scores NISRA 2010b Belfast 44 Coleraine 2 Craigavon 8 Derry 18 Dungannon Limavady 3 3 Lisburn 8 Newry & Mourne 7 Strabane 3 Others (1 each) 4 SOA s in worst

31 Part 3: a local profile of deprivation among older people Part 3 profiles some of the most and least deprived SOAs according to IDAOP to show what particular characteristics these areas exhibit. Extreme case studies The NI Neighbourhood Information Service (NINIS) 9 provides data at various geographical levels, including SOAs. It includes census data but is regularly updated using administrative data. This makes it possible to compare and contrast different small areas, in this case (Panel 2.3) the most and least deprived area as measured by the IDAOP index. Panel 2.3: comparison of the most and least deprived SOAs in NI Census 2011 Most deprived SOA Drumnamoe 1 in Craigavon is the most deprived SOA in NI as measured by IDAOP, with 95% of its older people experiencing income deprivation. It is also in the most deprived tenth of SOAs in the income, employment, health, education and child deprivation measures and in the overall MDM. It is in the least deprived fifth of the rankings in the proximity to services and crime and disorder domains. In the 2011 Census Drumnamoe 1 had 1,270 inhabitants, of whom 14.3% were aged 65 or older, close to the NI average of 14.6%. Population declined by 16% between 2001 (1,518) and The area is characterised by low educational attainment, with only 8% of the inhabitants aged 16+ holding degrees compared with a NI average of 24%; 64% had no or low-level qualifications, well above the NI average of 41%. The proportion of people with limiting long-term illness in 2011 was 34%, as opposed to 21% in NI and the number reporting at least good health (64.3%) was 15 percentage points below the NI average (79.5%). 50% of households lived in rented homes, compared with 30% in NI as a whole. Access to a car or van averaged 0.7 per household in 2011 and 47.4% of households had no car or van. In Drumnamoe 1 in 2011, 50% of people aged 16+ were economically active, lower than the NI average of 66%; unemployment at 9% was higher than the NI average of 5%. Administrative data for 2011 indicate that there were 155 income support claimants, 83 for incapacity benefit and 24 for employment and support allowance. If we express these as a % of men aged under 65 and women under 60 in 2010 (NISRA 2011), they work out at 13.6%, 7.3% and 2.1% respectively. Least deprived SOA Only 3% of older people are income-deprived in Jordanstown 2 in Newtownabbey, making it the least deprived SOA on this index (excluding Aldergrove 1, which has no older people). It is also in the least deprived tenth of SOAs on employment, income, health, education, living environment, crime and child deprivation indices and in the MDM. It is close to average in ranking on proximity to services. In 2011 Jordanstown 2 had a population of 2,047, of whom only 6.1% were aged 65 or older, well below the NI average of 14.6%. The population increased by 27% between 2001 (1,611) and The area has high educational attainment, with 43% of people aged 16+ holding degrees, substantially higher than the NI average of 24%; 20% had no or low educational attainments, well below the NI average of 41%. The proportion of people with limiting long-term illness in 2011 was 9.0%, far below the NI average of 21%, and the proportion reporting good health or better was 91.6%, well above the NI average of 79.5%. Only 4% of households rented their homes, far below the NI norm of 30%. Access to a car or van averaged 1.9 per household in 2011 and 2.3% of households had no car or van. In % of residents in Jordanstown 2 aged were economically active, 12 percentage points above the NI average; unemployment stood at 2.6%, about half the rate in NI. Data for 2011 show that there were less than five income support claimants, 25 for incapacity benefit and 11 for employment and support allowance. Expressed as a proportion of men under the age of 65 and women under 60 in 2010 (NISRA 2011), these equal 0.2%, 1.3% and 0.6% respectively. Note: see Table 2.5 for further details. 9. See for more information. 31

32 These case studies reveal that income deprivation among older people is not an isolated indicator. It is associated with many other forms of disadvantage. Drumnamoe 1 in Craigavon has higher rates of income deprivation among children and working-age people than Jordanstown 2, poorer health and much lower educational attainment. Closely related to these indicators is that Jordanstown 2 residents are far more likely to be in work and that, conversely, a much higher proportion of Drumnamoe 1 residents are in receipt of benefits. Jordanstown 2 inhabitants are far more likely to have a car and to own their own homes than are those in the Craigavon SOA. Data on most and least deprived areas The two case studies above highlight important differences between the most and least deprived areas as measured by IDAOP. However, IDAOP is only one measure of deprivation and big differences in rank can result from small changes in the number of income-deprived older people. This section examines whether the characteristics that distinguish the top and bottom SOAs are also evident in other areas. For this purpose Table 2.5 presents data on the three most and least deprived SOAs on the IDAOP measure. Since these are all urban areas, Crossmaglen is also included because it is the rural SOA with the lowest IDAOP scores (highest rank), to see if it shares the characteristics of the deprived urban areas. Table 2.5: rank of selected SOAs on deprivation measures NISRA (2010b) Ranks in 2010 Drum Ard Creg Cross Bally Up Mal Jordan IDAOP MDM Income Employment Health & disability Education, skills & training Proximity to services Living environment Crime & disorder Income deprivation among children Note: the names of the seven areas in full are Drumnamoe 1 in Belfast, Ardoyne 3 in Belfast, Creggan Central 1 in Derry, Crossmaglen in Newry and Mourne, Ballymaconnell 2 in North Down, Upper Malone 1 in Belfast and Jordanstown 2 in Newtownabbey. 32

33 A low rank on IDAOP is closely associated with low rank on child and overall income, employment, health, education and skills, living environment and MDM. Rural Crossmaglen scores somewhat better on these indicators than the deprived urban SOAs (Drumnamoe 1, Ardoyne 3 and Creggan Central 3), i.e. older people are somewhat more deprived than the general population (this may possibly indicate that there are low levels of occupational and private pensions in that area). Conversely the three areas with low income deprivation among older people (Ballymaconnell 2, Upper Malone 1 and Jordanstown 2) also do well on these specific indicators, typically falling into the top 10% or better. In the three least deprived areas, as measured by IDAOP, the proportion of people who reported good health in the 2011 Census ranged from 85% to 92%, which is better than Creggan Central 1 (76%) and Crossmaglen (77%) and much better than Drumnamoe 1 (64%) and Ardoyne 3 (67%) (Figure 2.3). The proportion of people with limiting long-term illness is markedly less in the three SOAs with low levels of poverty among older people, ranging from 9% to 18%, than in the four deprived areas, where the rate varies from 23% to 34%. Figure 2.3: health status of people in selected SOAs in 2011 Census 2011 Table KS301NI Limiting long-term illness Stated health good/very good Drum Ardoy Creg Cross Bally Up Mal Jordan Note: see Table 2.5 for the full names of the SOAs. At the time of the 2011 Census there was a large difference in the proportion of people with degrees or higher qualifications according to low income among older people. Figure 2.4 shows that in the urban deprived areas the rate was only 7 8%, while that for the rural deprived area was twice as high (16%). 33

34 Figure 2.4: education status of people aged 16+ in selected SOAs in 2011 (%) Source: Census 2011 Table KS501NI Age 16+ with degrees or higher Age 16+ with low/no qualifications Drum Ardoy Creg Cross Bally Up Mal Jordan Note: see Table 2.5 for the full names of the SOAs. Figure 2.5 (p34) shows that there are huge differences in the proportion of people who rent their homes between the urban deprived (50 72%) and urban non-deprived SOAs (4 12%) on the IDAOP index, with Crossmaglen coming in between (29%). Households in the least deprived IDAOP SOAs also have almost universal access to a car or van (92 98%), unlike the urban deprived areas (29 53%). Here too the rural area is different, with quite a high car access rate (77%). The 2011 Census also shows that the number of cars or vans per household was: Drumnamoe 1, 0.7; Ardoyne 3, 0.4; Creggan Central 1, 0.5; Crossmaglen, 1.3; Ballymaconnell 2, 1.8; Upper Malone 1, 1.6; and Jordanstown 2,

35 Figure 2.5: home and car ownership by household in selected SOAs in 2011 (%) Source: Census 2011 Tables KS402NI and KS405NI House rented Access to car or van Drum Ardoy Creg Cross Bally Up Mal Jordan Note: see Table 2.5 for the full names of the SOAs. Summary The number of older people is unevenly distributed across NI. At District Council level, the number of women aged 60+ and men aged 65+ ranges from 14% in Derry to 21.7% in North Down (NISRA 2011). At SOA level, the extremes run from 3% to 37.8%, excluding Aldergrove 1 because no older people live there. SOAs with high proportions of older people tend to have relatively few of them on low incomes and vice versa. There are also cases where SOAs have high ranks on IDAOP but much lower ranks on overall multiple deprivation and health and disability and cases where SOAs have low ranks on IDAOP but much higher ranks on the other two measures. There are very stark contrasts on many indicators between the least and most deprived areas as judged by IDAOP. An analysis over time, e.g. between the 2001 and 2011 Censuses, has not been provided but could be developed. The inequalities identified at SOA level affect very many older people on issues such as income, home ownership, access to a car, health status, employment and benefit dependency. 35

36 3. Income inequalities in Ireland, North and South 36

37 This chapter presents data on income inequalities in Ireland, North and South, as income is a useful indicator for social inclusion and a good metric of inequality. Part 1 deals with earnings from employment by 10-year age bands, while Part 2 is concerned with inequalities in pay within the older age group (60+). Part 3 reviews statistics on incomes generally and Part 4 identifies some of the factors behind inequalities by examining the sources of income of older people (aged 65+ in ROI and in NI women 60+ and men 65+). Part 1: Earnings Only a minority of older people are in paid employment but earnings from employment account for a great deal of the difference in overall incomes of the richest and poorest older people. In ROI, for example, the fifth of people aged 65+ with the lowest incomes earned 16 per week on average in 2011 whereas the fifth with the highest incomes earned 223 (CSO 2013). In NI earnings from employment averaged 10 per week for the lowest fifth of older couples in terms of income in , compared with 442 for the highest fifth; single pensioners on the lowest incomes earned an average of 1 in while those on the highest incomes averaged 36 (NISRA 2013). Hence, Part 1 of Chapter 3 examines differences in earnings. It might be assumed that older workers are better paid than younger ones because they have been in post longer and are therefore likely to be at the top of their pay scales. This is not the case in Ireland, North and South, where employees begin with low wages in their 20s and reach their peak earning years in their 40s; it remains high in their 50s (and in NI people aged are the highest earners in some of the years) but falls off quite sharply after that (NISRA 2012). Republic of Ireland The CSO database provides data on total annual earnings by age on its main database (Figure 3.1). 10 The difference in average annual earnings between workers in their 40s ( 46,049) and those aged 60+ ( 36,078) totalled almost 10,000 in Another feature is that average earnings of older workers declined in 2008 and then picked up again; overall, however, they were 623 per year lower in 2009 than two years previously. The group enjoyed higher earnings in 2009 than in 2007, an increase of more than 3,000 per year; the drop of about 300 between 2008 and 2009 indicates the early impact of the recession on earnings. 10 The CSO s Earnings Hours and Employment Costs Survey in ROI does not give earnings data by age. 37

38 Figure 3.1: ROI total average annual earnings by age group ( ) CSO Database* 42,870 46,376 46, ,701 34,706 36, * figures are shown only for the peak years (40 49) and 60+ Northern Ireland More detailed and more recent statistics are provided in NI by the Annual Survey of Hours and Earnings (ASHE), including a helpful analysis by age group (ONS 2012). 11 Figure 3.2: NI median weekly gross pay of full-time workers by selected age group ( ) ASHE 2012, ONS Note: 2012 data are provisional. 11 In some cases the coefficient of variation of ASHE figures is between 10% and 20%, which affects the quality of the estimates (ONS 2012); this is the case with many of the figures given for people aged

39 Unlike ROI, the oldest full-time workers (defined as 60+) have enjoyed the largest increases in median gross weekly pay in recent years, increasing by 46% between 2004 and 2012, from 326 to 477. People in their 50s had an uplift of 24% over the eight-year period. Notwithstanding this, older workers still had lower pay in 2012 than those aged 50 59, as shown in Figure 3.2. In 2012 employees aged 60+ were 9% behind those in their 50s, a difference of 47 per week or approximately 2,440 per year. Figure 3.3a shows that women workers are lower paid than men across the age range, with the largest differential of 82 per week occurring in the age range in Median weekly gross pay of full-time women workers is 95% that of men in the age range, 93% among workers in their 40s, 85% in their 50s and 88% among workers aged 65+. This supports the finding in several CARDI-funded research projects that older women are at a severe disadvantage compared with men, leading to lower female pension income in older age (Hillyard et al 2010; Hillyard and Patsios 2011; Duvvury et al 2012). Figure 3.3a also shows that both men and women aged 60+ are paid considerably less than men and women in their 40s and 50s respectively: e.g. men aged 60+ in work receive an average of 80 less than men in their 40s and women aged 60+ receive 99 less than women in their 40s (figures in the graph have been rounded). Figure 3.3a: NI median weekly gross pay of full-time workers by selected age group and sex 2012 ( ) ASHE 2012, ONS Male Female An analysis of hourly gross pay for all employees in 2012 (Figure 3.3b) shows the same broad pattern as weekly pay. Male employees aged 60+ earn 17% less per hour than those in their 40s and 13% less than those in their 50s; women aged 60+ earn 24% less per hour than women in their 40s and 19% less than women in their 50s. In addition, the increase in the gender gap with age is more pronounced in hourly than in weekly earnings. Women in their 30s earn 30 pence per hour less than men; in their 40s and 50s the difference is much higher at 2.23 and 2.10 respectively. Finally, women in their 60s are paid 2.59 less than men of the same age, 8.50 compared with Looking at female hourly earnings as a proportion of male hourly earnings, the figures fall from 97% aged to 83% aged 40 49, 83% aged and 77% aged

40 Figure 3.3b: NI median hourly gross earnings of all employees by selected age group and sex 2012 ( ) ASHE 2012, ONS Male Female In sum, this shows that older workers, North and South, are paid considerably less than workers in their 40s and 50s. In NI, but not in ROI, the gap is narrowing because older workers have gained larger increases in recent years than their younger colleagues. Women are paid less than men at all ages and the gender gap is wider at ages 50+ than at ages

41 Part 2: Pay inequalities within the older age group This part of Chapter 3 examines inequalities in earnings within the older worker group. Differences in earnings explain much of the variation in overall incomes, which in turn greatly influences access to goods and services and overall quality of life. It is simplistic to think of older workers as a homogenous group. The term covers a wide range of employees, from low-paid manual workers such as security staff or cleaners to highly paid professionals. In the case of earnings, the analysis is confined to NI because data are not available for ROI on the distribution of wages within the older population. Hourly earnings In analysing earnings, two measures are given in this chapter: the median hourly earnings of all employees and the median weekly earnings of full-time employees. These are analysed by comparing people on different ranks on the income spectrum. The literature on pay (see, for example, Brewer et al 2008) uses ratios such as the 90/10 comparison: i.e. the pay of somebody at the 90th percentile compared with an employee at the 10th percentile. ONS (2012) does not give a figure for the 90th percentile for employees aged 60+ as it is unreliable, so the analysis below uses alternative ratios: 75/25 and 80/ Figure 3.4a: NI gross hourly earnings of workers aged 60+ on the 25th and 75th percentile ( ) ASHE 2012, ONS 75th 25th The coefficient of variation of the 75th and 80th percentiles is 10 20%. 41

42 Taking the 75/25 ratio for hourly pay first (Figure 3.4a), the income of the quarter of people aged 60+ with the lowest earnings rose by 1.46 between 2004 and 2012 (and 1.11 between 2006 and 2012), from 5.86 to The top quarter s hourly pay rose by 3.12 over the same period (and 2.65 from 2006 to 2012), from to In 2012 the highest group was earning just over twice as much per hour as the lowest group. In the three years , marked by economic austerity, the average older employee on the 75th percentile had a rise of 1.33 per hour, while one on the 25th percentile had a decrease of 14 pence. An alternative measure is the earnings of people on the 80th and 20th percentiles, though Figure 3.4b also shows how low pay has essentially tracked, and may have been influenced by, changes in the national minimum wage (NMW). Here we see that inequality is growing more rapidly. The bottom fifth of employees aged 60+ had an hourly pay rise of 1.18 over the seven-year period compared with an increase for the top group of 3.65, so that the gap between them has risen from 8.44 to Older people on low pay had an average rise in the three years of the recession of 10 pence per hour, compared with a 35 pence increase in the NMW. People in the top fifth may have been less affected by the recession in , with an increase in hourly earnings of The impact of these changes on people aged 60+ was that the ratio of earnings of people on the 80th percentile to those on the 20th percentile increased from 2.28 in 2009 to 2.57 in 2012 (author s calculations). Figure 3.4b: NI gross hourly earnings of workers aged 60+ on the 20th and 80th percentiles ( ) ASHE 2012, ONS 80th 20th NMW Note: no figure is given for the 80th percentile in

43 Weekly earnings This section compares weekly earnings of full-time older workers in NI on the 75th and 25th percentiles (Figure 3.5). There is a widening gap, from 244 in 2004 (and 267 in 2006) to 296 in 2008 and then to 369 in The higher-paid group did not just gain three times more than the lower group in cash terms ( 197 compared with 71) over the eight years, but it also had a bigger percentage increase (38.8% as opposed to 26.9%). In 2004, full-time employees aged 60+ on the 75th percentile earned 1.9 times more than those on the 25th percentile (also 1.9 in 2006). In 2012 they earned 2.1 times more. In the three years of the recession , the top quarter enjoyed a weekly rise of 92, compared with 30 for the bottom quarter. Figure 3.5: NI gross weekly earnings of full-time employees aged 60+ on the 25th and 75th percentiles ( ) ASHE 2012, ONS 75th 25th Low pay The final paragraphs of Part 2 consider the position of older people at the bottom end of the spectrum, i.e. those with the lowest earnings, compared with average and higher-paid employees (Figure 3.6). Older workers on the 75th percentile had an overall pay rise of 197 (38.8%) between 2004 and 2012 ( 155 in ). People on average earnings received a larger percentage increase of 46.3%, which amounted to an additional 151. The lowest-paid 10th percentile had the smallest increase of any group in both cash and percentage terms, rising by 60 per week or 28.5%. The ratio of weekly pay received by older people on the 75th percentile compared with those on the 10th percentile has risen from 2.4 to 2.6 between 2004 and 2012 and the 50/10 ratio has gone up from 1.5 to 1.7 (there was no change in these ratios between 2006 and 2012). In other words, low-paid workers have fallen further behind those on both average and high earnings between 2004 and The 80/20 ratio is not examined here because ONS rates the 80th percentile figure of weekly earnings for people aged 60+ as unreliable for five of the years covered. 43

44 Figure 3.6: NI gross weekly earnings of full-time employees aged 60+ on median, 10th and 75th percentiles ( ) ASHE 2012, ONS 75th 50th 10th A feature of Figure 3.6 is the pay increases the top groups have enjoyed in the last three years, despite economic austerity. The top quarter had pay rises of 15%, adding 92 to their weekly pay; median pay increased between 2009 and 2012 by 24%, or 91 per week. The lowest 10% of older workers had a pay rise over the three years of 10%, amounting to 25 per week. Hourly rates for the lowest-paid 10% of older workers rose by 1.37 between 2004 and 2012, which works out at 27.5% (in the increase was 1.14 or 21.9%). This is lower than the increase in the NMW, which went up by 1.58 per hour (35.1%) between 2004 and In the three most recent years, the low-paid older workers had an increase in hourly pay of only 35 pence. In summary, NI data show that earnings inequalities are increasing over time, including the period of the recession ( ), whether measured in hourly or weekly pay. The earnings gap between the lowest-paid fifth of workers aged 60+ is increasing compared with both average earners and the highest-paid workers of the same age. 44

45 Part 3: Income inequalities Earnings from employment form a large part of the income of many older people but most are not in paid employment at all. Consequently, earnings account for only a small part of the total incomes of all older people: 16% of gross equivalised income of people aged 65+ in ROI in 2011 (CSO 2013), 23% of NI pensioner couples mean gross income after housing costs and 4% in the case of single pensioners in (DSD 2013). This part of Chapter 3 therefore focuses on the total incomes of older people. Northern Ireland incomes Median net incomes of the poorest pensioners in NI have barely changed in recent years whereas the richest have enjoyed larger incomes (DSD 2013). In the case of incomes before housing costs (BHC), the median net incomes of the poorest fifth of pensioner couples increased by 2 per week between and and the income of the next-poorest fifth increased by 6. The incomes of the richest fifth of older couples increased by 20 per week. The pattern was broadly similar for single pensioners BHC, with the income of the poorest fifth declining by 1 per week between and and the richest fifth receiving an increase of 7. These BHC statistics are not shown but Figures 3.7a and 3.7b show that the picture is similar for incomes after housing costs (AHC). Figure 3.7a: NI pensioner couples net weekly income after housing costs by quintile ( ) DSD, Pensioners Income Series, Q1 Q2 Q3 Q4 Q5 Among pensioner couples, median net incomes AHC of the poorest quintile remained unchanged. The richest fifth increased by 5.7%, boosting their incomes by 37 per week. The second-poorest quintile of pensioner couples had an increase of 16 per week, the middle group 23 and the second-richest quintile an increase of

46 Figure 3.7b: NI single pensioners net weekly income after housing costs by quintile ( ) DSD, Pensioners Income Series, Q1 Q2 Q3 Q4 Q5 The incomes of single pensioners AHC are only just over half those of pensioner couples, emphasising the particular vulnerability of older people who live alone, most of whom are women. In the case of inequality within the single pensioner population (Figure 3.7b), the poorest single pensioners had an increase in income of 2 per week, from 99 to 101. The next three quintiles rose by 9 10 whereas the weekly income of the richest fifth of single pensioners AHC expanded by 20 per week. These data indicate that inequalities among pensioners in NI have increased, regardless of which measure is used. Using the AHC figures, the net income ratio of the highest to lowest quintiles of pensioner couples increased from 3.7 to 3.9 and for single pensioners it rose from 3.0 to 3.2 (author s calculations using data in DSD 2013). The ratio for pensioner couples is similar to that for the UK as a whole (3.8), as is the ratio for the single pensioner group (3.1 in the UK) (Thane 2012: 25). The statistics for NI are presented in Figure 3.7c for the top and bottom quintiles. Figure 3.7c: NI pensioner units net weekly income in lowest and highest quintiles AHC ( ) DSD, Pensioners Income Series, lowest fifth highest fifth COUPLES COUPLES SINGLES SINGLES

47 Low-income pensioners are distributed quite unevenly across the 26 local government districts in NI (Figure 3.8). Five districts, headed by Armagh and Down, are better off than or equal to the UK average of 16% of pensioners below 60% median income and another four are at or below the NI average of 19%. In Banbridge the rate of pensioner poverty at 38% is twice as high as the NI average (19%). The three next-highest rates are in Dungannon, Carrickfergus and Larne and Moyle (which are combined due to sample size requirements), all with poverty rates of 30 32%. Figure 3.8: NI pensioners below 60% median income by local government district 2009/10 AHC (%) ARMAGH 7 DOWN 9 DERRY 13 ARDS 15 NORTH DOWN ALL UK ANTRIM 17 CASTLEREAGH 18 COLERAINE COOKSTOWN & MAGHERAFELT ALL NI NEWTOWNABBEY 20 NEWRY & MOURNE 21 CRAIGAVON LISBURN BALLYMONEY BELFAST FERMANAGH OMAGH & STRABANE BALLYMENA 27 LIMAVADY 29 CARRICKFERGUS LARNE & MOYLE DUNGANNON 32 BANBRIDGE 38 Source: DSD, Households Below Average Income Report , (2011b) Note: six LGDs have been combined into three pairs due to sample size requirements: Cookstown and Magherafelt; Omagh and Strabane; and Larne and Moyle. 47

48 Republic of Ireland incomes The bottom fifth of people aged 65+ in ROI in terms of weekly equivalised income had incomes equal to 53.2% of the average income in 2004 but this fell to 47.0% in 2010 and to 45.0% in Compared with the best-off older people (top quintile), the poorest fifth fell from 26.2% in 2004 to 22.0% in 2011 (CSO 2013). The quintile data for 2004 and 2011 are set out in Figure 3.9, 15 which shows that the lowest group had a modest overall increase in weekly income of 32, compared with 85 for the next group, 97 for the middle group, 123 for the second richest fifth and 255 for the richest older people. The ratio of incomes in the top and bottom quintiles has widened from 3.8 to 4.5. Figure 3.9: ROI equivalised weekly incomes of people aged 65+ by quintile ( ) CSO Q1 Q2 Q3 Q4 Q5 Changes between 2004 and 2011 disguise an increase and a subsequent decline. For example, average incomes of people aged 65+ increased by 48.4% between 2004 and 2009 but then declined by 5% between 2009 and This indicates that older people have not been protected from the effects of the recession in ROI, where gross domestic product declined in each of the years (IMF 2013). More importantly, the incomes of the poorest fifth of older people increased by less than average between 2004 and 2009 (36.2%) and then declined more sharply between 2009 and 2011 ( 11.4%); in cash terms the decline in the latest two years amounts to 24, from 209 to 185 per week. The incomes of the richest fifth of older people rose faster than for any other quintile between 2004 and 2009 (55.2%), followed by a two-year decline of 7.6%. 15. More detailed figures are given in Appendix 1. 48

49 Part 4: Accounting for greater income inequality This part of Chapter 3 cites evidence of the sources of income for rich and poor pensioners and how these have changed in recent years. This will help us to understand the main factors underlying income inequality. Figure 3.10 shows that between 2004 and 2011 the main factors explaining the growing differential between the bottom and top quintiles in terms of income were social transfers (see below) and occupational pensions, which increased by an average of 104 (85%) per week for the highest quintile and by 2 (55%) for the lowest quintile. On the other hand, the lowest quintile had an extra 8 from earnings over the seven years whereas earnings for the top quintile declined by 21. ROI sources of income Figure 3.10: ROI sources of weekly income of people aged 65+ in top and bottom quintiles, 2004 and 2011 ( ) CSO 2013 All Other Income Occup Pension Social Transfers Earnings ,706 Q Q Q Q The broad pattern was an increase in income between 2004 and 2009 for all quintiles and sources of income, followed by a more mixed pattern between 2009 and The lowest quintile had big relative declines in income from occupational pensions, investment and property, but these were from a small base and were offset by higher earnings. Older people in the highest quintile experienced a decline of 54 per week in earnings and 23 in property income from 2009 to 2011 as well as a drop in other income. In ROI most of the decline in the incomes of the poorest older people between 2009 and 2011 came from a 25 per week reduction in social transfers. 16 Figure 3.11 illustrates that the richest fifth of older people receive much more in social transfers than the poorest fifth. In 2011, for example, the top income quintile received an average of 314 per week in benefits whereas the poorest fifth received only 159, a difference of 156. Moreover the gap between them has grown enormously from 23 in CSO includes under social transfers unemployment-related payments, old-age related payments, family/children-related allowances and other transfers such as sickness or disability benefits. 49

50 Figure 3.11: ROI average weekly income by quintile from social transfers 2004 and 2011 CSO Q1 Q2 Q3 Q4 Q5 NI sources of income The most recent available data on components (or sources) of gross incomes in NI, which are presented separately for pensioner couples and single pensioners, relate to the three-year average figures for and (DSD 2013). This means that it is now possible to estimate an initial recession effect in the NI data. As in ROI, the richest fifth of pensioners (both couples and singles) in NI receive more in benefits than the poorest fifth (see Figure 3.12) but an important difference is that the gap is getting smaller rather than bigger. In the quintile of pensioner couples with the highest incomes received 202 per week in benefits and the bottom quintile 160, a margin of 42. In the amount received by the top fifth declined to 197 and the sum received by the lowest quintile increased by 9. This reduced the gap in benefit receipts for the top and bottom fifths of pensioner couples to 28 per week. 17. In NI pensioner means women aged 60+ and men aged 65+ though the retirement age for women is increasing to 65 by See Appendix 1 for fuller details of components of gross income. 19. This section is based on gross incomes whereas the earlier part of this chapter used net income data. 50

51 Figure 3.12: NI pensioner couples main sources of gross weekly income by quintile ( ) NISRA Pensioners Income Series 2013 All Other Income Earnings Occupation & Personal Pension Benefits , , ,706 Q1 Q2 Q3 Q4 Q5 Figure 3.12 also shows that the main factor distinguishing between rich and poor pensioner couples is earnings from employment. The poorest fifth received only 13 per week on average from this source in , declining to 10 in All of the higher quintiles had larger and increasing earnings, especially the richest quintile, whose average earnings rose from 368 to 442 per week. Apart from benefits and earnings, the main sources of difference between the richest and poorest pensioner couples in were occupational and personal pensions ( 17 for the bottom quintile and 414 for the top) and investments ( 1 per week for the poorest and 122 for the richest). In the case of gross incomes, the ratio between the highest and lowest pensioner couples was 5.6 in , rising to 6.0 in In cash terms the gap has risen from 889 to 986 per week between the richest and poorest pensioner couples. Incomes of single pensioners are just half those for couples. The gap between the richest and poorest single pensioners in benefit income reduced slightly because the top group had a drop of 2, from 206 to 204 while the bottom quintile had an increase from 111 to 115. Money received from occupational pensions, rather than earnings, is the main factor explaining the difference between rich and poor single pensioners, as shown in Figure

52 Figure 3.13: NI single pensioners main sources of weekly income and ( ) NISRA, Households Below Average Income (2013) All Other Income Earnings Occupation & Personal Pension Benefits ,701 34,706 Q Q Q Q Occupational and personal pensions made a negligible contribution to the weekly budgets of the poorest single pensioners, 5 per week in both and By contrast, the richest pensioners received 141 from this source in and 160 in The smaller range of incomes among single pensioners compared with pensioner couples meant that the ratio between top and bottom incomes was less extreme. Moreover, unlike the pattern with net incomes, the gap between the highest and lowest single pensioner groups has become smaller in recent years due to a decline in gross income of the highest quintile from 461 to 419 per week ( 42). This means the gross income ratio declined from 3.8 to 3.4 between and The examination of sources of income shows that in ROI the growing gap between the richest and poorest quintiles in the years is largely explained by big increases in social transfers and private and occupational pensions by the highest group. In NI the gap in receipts from benefits has reduced in the latest time period ( to ) but the richest quintile of pensioner couples has gained much more from earnings. Summary This chapter, which focused on income inequalities in Ireland, North and South, outlined the big pay differences between older and younger workers, amounting to 10,000 per year in ROI and 4,000 in NI. Within the older worker group there are big differences in earnings and the gap between the highest and lowest earners is getting greater. Even during the austerity years, , top earners received larger rises while the low-paid were virtually at a standstill. In the case of total incomes, the poorest pensioner units in NI have had a drop in net incomes (but not in gross incomes) in recent years while the richest have had an increase. In ROI both the richest and the poorest older people had lower incomes in 2011 than in 2009 but in the seven years covered by SILC data ( ) the lowest fifth of older people had a growth in equivalised income of 32 per week while the richest fifth had an increase of 255. Most of the difference is accounted for by greater receipts by the richest older people from social transfers, occupational pensions and earnings from employment. 52

53 4. Policy background 53