Chapter 13: Inequalities and Poverty in Retirement, (2012 edition)

• Although retired households’ income is lower than that of non-retired households, the relative position of retired households has improved over the last three decades (Figure 13.2).

• In 2010/11, income from private pensions accounted for a similar proportion of the average gross income of retired households as the state pension (Figure 13.6): 41 per cent and 38 per cent respectively.

• In 2010/11, pensioners with private pensions were more likely to be in the higher income quintile groups (Figure 13.8).

• 37 per cent of white pensioners were in the top two income quintile groups, while 35 per cent of Pakistani and Bangladeshi pensioners were found in the bottom income quintile (Figure 13.12).

• In 2010/11, an estimated 14 per cent of pensioners in the UK were living on less than 60 per cent of equivalised contemporary median income after housing costs (see Definitions). This is compared to 28 per cent in 1999/2000 (Figure 13.15).

• In 2010/11 around 9 per cent of pensioners were classified as being materially deprived. Material deprivation refers to whether people have access to common necessities.

• There were 1.1 million households of single people aged 60 or over in ‘fuel poverty’ in England in 2010 according to data published by the Department of Energy and Climate Change (Figure 13.19).

This chapter looks at inequalities and poverty in retirement. Inequality is a relative concept, in which the position of one individual or household in a distribution, normally an income distribution, is compared with that of other individuals or households. There is no single definition of poverty. The concept can refer to an absolute standard, such as a minimum level of income which society agrees to be acceptable for providing a basic standard of living (a ‘poverty line’). Alternatively, poverty can be measured in relative terms.

The analysis begins by looking at inequalities. First, it compares average incomes and income inequalities of retired and non-retired households over the last three decades. Then it looks at the effect that different sources of income have on a pensioner household’s position within the income distribution. The discussion of income inequality concludes with an exploration of other characteristics associated with a household’s position in the income distribution, such as the type of household, age, ethnicity, employment status and housing tenure.

The second part of the chapter looks at pensioner poverty. It also presents one of the officially used measures of poverty, which is based on a proportion of contemporary median income (see Definitions). An alternative indicator to this, which is the analysis of the number of households falling below absolute poverty income thresholds, is also discussed. This is defined within the Households Below Average Income (HBAI) data series as those households which fall below a specific percentage (50, 60 and 70 per cent) of inflation adjusted median income in 1998/99. It also discusses alternative measures, including attempts to establish minimum income standards, measures of material deprivation and measures based on expenditure.

The equivalisation of household income

The needs of a household grow with each additional member, but the need for housing space, electricity and the like, will not be three times as high for a household with three members as for a single person household. With the help of equivalence scales, each household type in the population is assigned a value in proportion to its needs. Equivalence scales conventionally take an adult couple without children as the reference point, with an equivalence value of one.

The calculation process then increases the income of single person households (since their incomes are divided by a value of less than one) and reduces the incomes of households with three or more persons, which have an equivalence value of greater than one. The household income data presented in this chapter has been adjusted using the modified Organisation for Economic Co-operation and Development (OECD) scale.

Household income before and after housing costs

In the modified OECD equivalence scale two separate versions are used, one for income Before Housing Costs (BHC) and another for income After Housing Costs (AHC). BHC measures of income are often used when considering the distribution of income across the entire population, whereas AHC measures of income subtract housing costs from income. AHC is mainly used when looking at the living standards of individuals at the lower end of the income scale or when measuring pensioner poverty.

The main issues when deciding which measure to use are whether people face genuine choices over their housing and whether differences in housing costs reflect differences in housing quality. For individuals who exercise a considerable degree of choice over housing costs and quality, a BHC income measure may be appropriate as housing can be seen as a consumption good. Where individuals have little choice about housing and where rents may be set with little reference to housing quality, for instance in the social housing sector, it can be argued that an AHC measure of income is more suitable.

AHC measures of income are also considered appropriate for pensioners, as around 70 per cent of pensioners own their homes outright. This group of pensioners can attain a higher standard of living than those with the same income level BHC but with mortgage or rental payments which reduce their AHC incomes. However, an AHC measure of income will understate the relative standard of living of those benefiting from better quality of housing by paying more for their accommodation.

Household income data sources

Information on household income in this chapter comes from two sources:

• A report on the Effects of Taxes and Benefits on Household Income (ETB) which is produced by the Office of National Statistics (ONS).

• The Households Below Average Income (HBAI) data series which is produced by the Department for Work and Pensions (DWP).

The income distribution used in both the ETB and HBAI analyses are based on a ranking of households by equivalised disposable household income (see The equivalisation of household income). All data published by the ETB series are BHC because of the difficulty in producing AHC adjusted values from the Living Costs and Food (LCF) survey.

Disposable income in both sources includes income from employment and self employment, state benefits, pensions and investment income, less payments of income tax, National Insurance contributions and Council Tax.

One of the most important differences is that for the ETB the unit of analysis is the household, while for HBAI the primary unit of analysis is the person. The income distribution appearing in the HBAI series is not a distribution of households, but rather a distribution of people according to the income of the household in which they live.

In the current ETB series, an individual is defined as retired if:

• They describe themselves as ‘retired’.

• If they are over the minimum National Insurance pension age and are either ‘unoccupied’ or ‘sick or injured but not intending to seek work’.

A household is defined as retired in ETB when more than 50 per cent of the household income is received by retired members of the household.

In the HBAI series, pensioners are defined as those over State Pension Age (SPA) at the time of the interview and, with the exception of the data used in Figure 13.10, results for pensioner couples exclude partners who are of working age. The SPA has recently undergone changes and has resulted in an increase of time needed for women to work before they can begin collecting their state pension. The SPA for women is set to rise to 65 by November 2018.

It was decided to use data sets from the ETB in this chapter instead of the DWP’s Pensioner Income (PI) series, as used in Pension Trends Chapter 11, because of the extensive time series it provides. The start year from which ETB data can be sourced is 1961/62. However, only after 1977 can data be accurately compared between years.

Income in retirement

The incomes of retired and non-retired households differ in terms of their level, distribution and composition. The following section focuses on the income of retired households, presenting results for non-retired households for comparison.

Figure 13.1 shows that over the last thirty years, the average equivalised disposable income has been lower for retired households than for non-retired households. Incomes of both retired and non-retired households have risen in real terms.

The average disposable incomes of non-retired households are subject to fluctuations associated with economic cycles. In the recession of the early 1990s, they peaked just before the recession and then fell, recovering in the second half of the decade. There is a similar pattern of decline in the incomes of non-retired households associated with the 2008/09 recession.

In contrast, incomes from state pensions and from private pensions, which are the main components of retirement income (see Income inequalities), show little fluctuation from year to year. Consequently, the incomes of retired households have grown more steadily than those of non-retired households over the last three decades.

Although retired households’ income is lower than that of non-retired households, the position of retired households has improved since 1977. Figure 13.2 shows that between 1977 and 1991 the average equivalised income of retired households was, on average, 61 per cent of that of non-retired households. For the period between 1992 and 2010/11 this proportion was around 65 per cent.

Income inequalities

The Gini coefficient is a widely-used measure of income inequality. Gini coefficients take a value from 0 to 1, with zero indicating total income equality (that is each household has an equal share of total household income) and higher values indicating greater income inequality. These proportions can also be expressed as percentages (0 to 100 per cent).

Figure 13.3 shows that over the last three decades income distribution has been less unequal (the Gini coefficient is lower) for retired households than for non-retired households. However, it should be noted that this result does not say anything about the poverty of retired households compared with that of non-retired households. Poverty is discussed later in this chapter.

The distribution of income has become more unequal over the last three decades for both retired and non-retired households, particularly between 1977 and 1990. Over this period, the Gini coefficient increased from 21 per cent to around 30 per cent for retired households and from 25 per cent to 35 per cent for non-retired households.

The similarity of these changes may be considered surprising, as the main sources of income of retired and non-retired households are quite different. Between 1991 and 2010/11 there has been a fall in the Gini coefficient for pensioners, representing an increasing equality in disposable pensioner income. In 1991 it was at a maximum of around 30 per cent for pensioners and this has fallen by 4 per cent to 26 per cent in 2010/11. There has been some variation, but little change overall for non-retired households over the same period.

Figure 13.4 examines the percentage shares of total equivalised disposable income (BHC) for retired households by income quintile group. Quintiles divide the data, ranked in ascending order, into five equal parts, from households with the lowest incomes in the bottom quintile group to households with the highest incomes in the top quintile group. Between 1977 and 2010/11 the top two income quintile groups of retired households increased their share of total retired household income while the bottom two quintile groups saw the size of their shares fall. The bottom income quintile group saw the largest fall, from around 13 per cent in 1977 to 10 per cent in 2010/11.

It is clear from figure 13.4 that the overall trend can be split into two distinct phases. The first is between 1977 and 1990, when the top quintile group of retired households increased its share of total household income from 34 per cent to 41 per cent at the expense of the lowest three income quintile groups. In the second phase, between 1991 and 2010/11, the share of total income held by the top quintile group fell to 36 per cent, while that for the second, third and fourth quintile groups increased.

While Figure 13.4 divides retired households into income quintile groups, Figure 13.5 details the proportion of retired households within each income quintile group.

Overall, the proportion of retired households in the bottom income quintile has fallen since 1977, while the proportion in the four higher quintile groups has increased. In 1977, retired households made up the largest proportion of the bottom income quintile group (54 per cent). By 2010/11, this had fallen to 26 per cent. Over the same period, the proportion of retired households in the four higher income quintiles increased. This was most pronounced in the third (middle) quintile, where the proportion of retired households increased from 8 per cent in 1977 to 23 per cent in 2010/11.

Income sources and inequalities

The main sources of income for retired households have changed substantially over the last three decades. Figure 13.6 shows that in 1977, income from private pensions accounted for just 18 per cent of the average gross income of retired households, compared with 53 per cent from the state retirement pension (the basic state pension and additional state pensions combined).

The state retirement pension was the largest source of income for retired households until 2000/01. Income from private pensions increased almost continuously over the three decades to 2006/07. It fell slightly the following year, but then recovered. In 2010/11, income from private pensions accounted for a similar proportion of the average gross income of retired households as the state retirement pension: 41 per cent and 38 per cent respectively.

Investment income was also an important source of income for retired households between 1977 and 2010/11. The level of investment income depends on the level of financial assets held by retired households and the returns on these investments in the form of dividends and interest payments. Investment income rose from 14 per cent of average gross income of retired households in 1977 to peak at 21 per cent in 1991. However, it represented only 6 per cent of the total average income for retired households in 2010/11 and this can largely be attributed to the recession.

Income from benefits doubled in real terms between 1977 and 1994/95. The largest increases were in housing, invalidity, disability and carers’ benefits. Income from benefits fell slightly in the remainder of the 1990s, before increasing in the early part of the 21st century, probably due, at least in part, to the introduction of Pension Credit in October 2003 (see Pension Trends Chapter 5). However, the average contribution to income provided by benefits has changed little since 1977. Between 1977 and 2010/11 it varied between 11 and 15 per cent and in 2010/11 it represented 11 per cent of the average income of retired households.

The sources of income of retired households are related to their position in the income distribution. Figure 13.7 shows percentage shares of income from private pensions for retired households by retired household income quintiles from 1977 to 2010/11. Income from private pensions has become more evenly distributed across the retired population since the mid-1990s, although more than half still goes to those in the top income quintile group. In 1991, around 66 per cent of income from private pensions went to retired households in the top income quintile, falling to nearly 55 per cent in 2010/11.

Figure 13.8 takes data from the Households below Average Income (HBAI) series to illustrate the effect of private pensions on the income levels of single pensioners and pensioner couples.

In 2010/11, pensioners with private pensions were more likely to be in the higher income quintile groups than those without private pensions. Pensioner couples where both partners had private pensions were the most likely of all the categories in Figure 13.8 to be in the top two income quintile groups and the least likely to be in the bottom two groups. Over three-fifths of pensioners with no private pensions were in the bottom two income quintile groups. Single people with no private pension were the least likely to be in the top two income quintile groups and couples with no private pension were the most likely to be in the bottom two.

The higher the value of pensioners’ savings, the more likely they are to be at the upper end of the income distribution. Figure 13.9 shows that over three-fifths of pensioners with no savings were in the bottom two income quintiles in 2010/11. At the other extreme, around half of pensioners with savings of £20,000 or more were in the top two income quintiles. Pensioners with this level of savings were also the least likely to be found in the bottom two quintiles.

Other characteristics associated with income inequality

Family and household characteristics are also related to the individual’s position within the income distribution. Figure 13.10, which is for the whole population including non-pensioners, shows that the group with the largest proportion of individuals in the bottom two income quintiles in 2010/11 was working age single parents (67 per cent), followed by single female pensioners (56 per cent) and single male pensioners (49 per cent). Single parents and single pensioners also had the smallest proportions in the top two quintiles, with 13 per cent of single parents, 21 per cent of single female pensioners and 28 per cent of single male pensioners in the top two quintile groups.

The results in Figure 13.10 are presented for income Before Housing Costs (BHC). On an income After Housing Costs (AHC) basis the proportion of single pensioners in the lower income quintiles is smaller. In 2010/11, 47 per cent of single female pensioners and 37 per cent of single male pensioners were in the bottom two income quintiles, while 31 per cent of female and 39 per cent of male pensioners were in the top two income quintiles.

Figure 13.11 shows (on an income AHC basis) that in 2010/11, the younger the pensioner, the more likely they were to be in the upper income quintile groups for the whole population. Pensioners aged 60 to 64 were the most likely to be in the top income quintile (26 per cent in this age group). Pensioners aged 85 or more were the least likely to be in the top income quintile group (9 per cent).

Figure 13.12 shows that Pakistani and Bangladeshi pensioners were the most likely of all ethnic groups to be found in the bottom two income quintile groups (71 per cent). White pensioners were the most likely to be in the top two income quintile groups (37 per cent), closely followed by those from Chinese or other ethnic groups (30 per cent) and Indian pensioners (28 per cent). White pensioners were the least likely to be found in the bottom income quintile group (12 per cent), while Pakistani and Bangladeshi pensioners were the most likely (35 per cent).

One reason for this may be that people from ethnic minority groups are more likely than those from white backgrounds to have characteristics associated with low pension incomes. These include lower rates of pension scheme membership, lower rates of employment and earnings, lower rates of eligibility for the additional state pension, and lower rates of employment at older ages. Pakistani and Bangladeshi workers are also more likely to work part-time than white workers³.

For pensioners, being in employment is associated with a better position within the income distribution (Figure 13.13). In 2010/11, pensioner households with one or more members working were more likely to be found in the top income quintile group (36 per cent) than those in which no one was employed (12 per cent). These households were also less likely to be found in the bottom two income quintiles (17 per cent) than those where no one was employed (44 per cent).

There is also a relationship between housing tenure and position in the income distribution. Figure 13.14 shows that in 2010/11, pensioners who were social housing tenants were least likely to be found in the top two income quintile groups (16 per cent). Pensioners who rented privately were most likely to be in the bottom income quintile group (28 per cent). Pensioners who owned their own homes were most likely to be in the top two income quintiles groups (42 per cent).

Notes for Inequalities

1. Figure 13.6 can be interpreted as presenting the breakdown for the ‘average household’ (based on the mean). However, the distribution of income from pensions and other sources varies considerably between households (see Pension Trends Chapters 11 and 12)

2. In figure 13.6, the fall in income from private pensions and annuities in 2007/08 should be treated with caution as it may be the result of sampling variation

3. Information regarding the working characteristics of ethnic minorities is sourced from Steventon, A and Sanchez, C (2008) The under pensioned: disabled people and people from ethnic minorities Research Report No. 5 Pension Policy Institute, Equality and Human Rights Commission

Whereas the previous section presented measures of income inequality for pensioners by examining their position within the income distribution, this section explores indicators of poverty. The UK has no official ‘poverty line’ (a minimum level of income which society agrees to be acceptable for providing a basic standard of living).

One of the official poverty measures used by the Government is a relative one: a proportion of equivalised contemporary median household income After Housing Costs (AHC) (see Definitions). For other measures please see the recent Households Below Average Income report (HBAI) published by the Department for Work and Pensions (DWP).

The equivalised contemporary measure shows a downward trend in the proportion of pensioners living in poverty since the mid-1990s. Figure 13.15 shows the proportion of pensioners falling below 50, 60 and 70 per cent of equivalised contemporary median income AHC, which is a relative poverty income measure (see the ‘HBAI’ report published by the DWP for a description of the alternative absolute poverty measure). In 1994/95, when contemporary median income for a couple with no children was £277 per week in 2010/11-equivalent money values, 40 per cent of pensioners fell below the 70 per cent threshold.

By 2010/11, when equivalised contemporary median income AHC was £359 per week, this proportion had fallen to 24 per cent. In 1994/95, 28 per cent of pensioners fell below the 60 per cent threshold; by 2010/11 this had fallen to 14 per cent. The proportion of pensioner households falling below 50 per cent of contemporary median income has varied between 8 and 13 per cent since 1994/95.

For the HBAI absolute poverty measure in 2010/11, where current incomes are compared against inflation adjusted median income values in 1998/99, 14 per cent of pensioners fell below the 70 per cent threshold, 9 per cent of pensioners fell below the 60 per cent median income threshold and 5 per cent fell below the 50 per cent threshold. The proportion of pensioners falling below the 50 per cent absolute poverty income threshold has varied between 8 and 5 per cent since 2000/01.

Returning to the AHC meaure, in 2006/07, there was an increase in the proportion of pensioner households falling below all three thresholds (50, 60 and 70 per cent). However, this increase has been reversed since 2007/08.

The most commonly used proportion of contemporary median household income AHC is 60 per cent. In 2010/11, an estimated 1.7 million pensioners in the UK were living on less than 60 per cent of equivalised contemporary median income AHC, compared with 2.8 million in 1999/2000.

Figure 13.16 shows that a larger proportion of female pensioners than male pensioners live in households with incomes below 60 per cent of contemporary median income. However, the gap has narrowed since 1994/95, when 31 per cent of female pensioners and 24 per cent of male pensioners were living in households with less than 60 per cent of contemporary median income. In 2010/11, these proportions had fallen to 15 per cent and 13 per cent respectively.

Figure 13.17 shows what proportion of the total pensioner population living in households with less than 60 per cent of contemporary median household income (AHC) are men, and what proportion are women. Despite a slight reduction since 1994/95, over two-thirds of pensioners living in relative poverty are still women. This is partly because there are more female than male pensioners in the pensioner population as a whole, and partly because of the low income levels of many women pensioners.

The low income levels of women pensioners reflect the gender pay gap during people’s working lives as well as interrupted working histories due to caring responsibilities, which impair the ability of women to save for retirement though the pension system. In addition, many older women pensioners are widows, and the level of occupational pension received by a widow is generally only half the level of the pension received by the couple when the man was alive.

The measure of poverty discussed in the previous section was a relative measure. This section discusses two different measures which could be used for measuring pensioner poverty in the UK. The first is the Standard Minimum Guarantee associated with the ‘guarantee credit’ element of Pension Credit, which is designed to provide a minimum acceptable level of income for pensioners (see Pension Trends Chapter 5).

The standard guarantee credit level is set by Parliament, and in 2011/12 was £137.35 per week for a single person and £209.70 per week for a couple (two people living together). This is set to rise to £142.70 and £217.90 respectively in 2012/13. In addition to the standard amount, carers and disabled pensioners may receive additional supplements.

A similar measure has been developed by the Centre for Research in Social Policy (CRSP) at Loughborough University and the Family Budget Unit (FBU) at the University of York. This is known as the Minimum Income Standard (MIS) for Britain. It is based on public consultation and expert opinion. The MIS shows the cost of basic goods and services for different household types and is calculated by specifying baskets of goods and services required by different households in order to meet their needs and to participate in society.

The MIS claims to represent a level of income below which it is socially unacceptable for anyone to live. Almost all households officially defined as being in relative poverty, with incomes below 60 per cent of contemporary median income, are also below the MIS. In April 2012, the MIS for pensioner couples was £205.93 (AHC).

The measures of poverty discussed so far in this chapter are based on income. An alternative to this is to consider indicators of standards of living or material deprivation.

Material deprivation is calculated from the responses people give to interview questions about how accessible they find everyday necessities. Their responses are weighted by the prevalence of the item within the population so that the more commonplace an item is, the more deprived someone is who cannot afford it. A record of material deprivation for pensioners aged over 65 has been collected over the past two years within the Family Resources Survey (FRS) and this is published in the Households Below Average Income (HBAI) report.

The essential items queried in this survey include the ability of a person to socialise with friends or family, their ability to have one filling meal a day, their access to a car or taxi and whether or not they could afford an unexpected two hundred pound expense.

According to the 2010/11 HBAI report around 9 per cent of pensioners aged 65 or over were classified as being materially deprived (a 0.1 per cent drop since 2009/10). Generally, the bottom two income quintile groups gave more negative responses to the survey (AHC) than the top two and key findings of the 2010/11 results include:

• 51 per cent of pensioners within the bottom quintile group were not able to take a holiday away from home.

• 86 per cent of the bottom quintile group had regular access to a car or taxi. This is in contrast to 98 per cent in the fourth quintile.

• in the bottom two quintile groups 30 per cent and 29 per cent of pensioners respectively were not able to go out socially once a month. This compares with 24 per cent and 11 per cent in the top two quintiles.

The published results have also shown the importance of receiving payments from private pensions. In 2010, according to the FRS the proportion of the 26 per cent of pensioners who did not have a private pension and who were classified as materially deprived was around 60 per cent. This is in contrast to those who were receiving a private pension, where around 40 per cent were described as being materially deprived. However, care should be taken with these figures as the provision of a private pension is one factor out of the many that are required to determine the risk of being materially deprived.

Standard poverty measures based on income, whether designed to measure relative or absolute poverty, ignore factors which may be associated with increased income needs (such as those of people living with health problems or disabilities). The final section of this chapter examines measures that consider households’ expenditure needs.

Expenditure levels vary greatly between types of retired and non-retired household, according to results from the Office for National Statistics (ONS) Living Costs and Food Survey (Figure 13.18). Couples had higher average expenditure levels than single households in 2010. Non-retired households with two adults had the highest levels of expenditure, while single retired households had the lowest levels of expenditure. One- and two-adult retired households had around two-thirds of the expenditure levels of their non-retired counterparts without children.

The average proportion of total expenditure spent on different items also varies according to household type. In 2010, a larger proportion of the expenditure of one- and two-adult retired households was on net housing costs, fuel and power, and household goods and services than it was for one- and two- adult non-retired households (without children). For single retired households, expenditure on housing, fuel and power, and household goods and services made up over one-quarter of total expenditure, compared with nearly one-fifth for two-adult non-retired households with children (see also Pension Trends Chapter 11).

A related concept is that of ‘fuel poverty’. Households are considered to be in fuel poverty if they have to spend more than 10 per cent of their income on fuel to maintain a satisfactory level of heating (21 degrees Celsius for the main living area and 18 degrees Celsius for other occupied rooms). A report published by the Department of Energy and Climate Change in 2012 revealed that every year since 2003 there has been an increase in the number of pensioners falling into fuel poverty.

In 2003, there were 0.5 million pensioners said to be in fuel poverty and this increased to 1.1 million in 2010. However, current trends demonstrate that relative to 2009 the number of those classified as being in fuel poverty in 2010 has dropped for all categories including pensioners. The type of household most affected by fuel poverty was that of single people aged 60 or over where around 35 per cent of these households in England were in fuel poverty (Figure 13.19).