The impact of poverty on wellbeing during midlife.
Abstract: In this paper poverty during midlife (ages 40-64) is explored through objective measures such as income, assets, housing tenure and poverty thresholds, and through subjective measures of financial adequacy. Statistical analyses were used to determine which of the objective measures were associated with satisfaction with economic living standards and satisfaction with life overall. The associations between these variables provide insights into the impacts of poverty on wellbeing during this stage of life. In particular, living on a low income, living below the poverty threshold and living with few assets in midlife were all negatively associated with wellbeing. In addition to the immediate impact on wellbeing, the lower level of asset accumulation and home ownership, compared with older cohorts, suggest key challenges for policy as these midlife cohorts move into old age in larger numbers. The preservation of the purchasing power of New Zealand's universal superannuation scheme presents a further policy challenge.
Article Type: Report
Subject: Poverty (New Zealand)
Poverty (Social aspects)
Middle age (Social aspects)
Middle age (Economic aspects)
Quality of life (Economic aspects)
Authors: Waldegrave, Charles
Cameron, Michael P.
Pub Date: 07/01/2010
Publication: Name: New Zealand Journal of Psychology Publisher: New Zealand Psychological Society Audience: Academic Format: Magazine/Journal Subject: Psychology and mental health Copyright: COPYRIGHT 2010 New Zealand Psychological Society ISSN: 0112-109X
Issue: Date: July, 2010 Source Volume: 39 Source Issue: 2
Topic: Event Code: 290 Public affairs
Geographic: Geographic Scope: New Zealand Geographic Code: 8NEWZ New Zealand
Accession Number: 243877213
Full Text: Midlife, defined here between the ages of 40 and 64, is the life stage where many people have become established in their family and employment roles. It is often characterised by the latter years of parenting and the early years of grandparenting. Income earning tends to peak during this period and asset accumulation usually reaches its zenith just prior to retirement. Key bodily functions and abilities that worked well during early midlife begin to manifest their limitations during the latter stages of midlife.

Midlife is typically the period prior to retirement transitions and as such sets the stage for wellbeing in later life. The current midlife group is studied here, because included within it is the large post-War 'baby boomer cohort' who are about to become eligible for National Superannuation (currently 65 years in New Zealand) in greatly increased numbers than qualifiers in previous years. The median age of New Zealanders is projected to rise from 36 years at present to 46 years in 2051, and the percentage of those aged 65 or over will increase from 12 percent to 26 percent in 2051 (Statistics New Zealand 2007). Understanding the characteristics of these cohorts and how well they are prepared for later life is important, if the country is to ensure their wellbeing and the wellbeing of subsequent cohorts. This paper focuses on the associations of the various components of economic poverty with wellbeing within the midlife group.

Poverty during midlife is explored through objective data, such as income, assets, housing tenure and poverty thresholds, and through subjective data such as personal assessments of financial adequacy, satisfaction with economic living standards and satisfaction with life overall. The associations between these data provide insights into the impacts of poverty on wellbeing during this life stage and provide evidence for a number of policy challenges.

Economic Living Standards and Wellbeing

Subjective wellbeing in the literature is often a synonymous term for quality of life and primarily concerns subjective assessments of satisfaction and/or happiness. Some studies focus primarily on the relationship between wellbeing and income, while others explore the relationships between wellbeing and asset accumulation, socio-economic status and health. As the studies described below show, positive relationships between income, socio-economic status and wellbeing are often, but not always, apparent.

Generally speaking, the midlife cohort is better off than all other cohorts in both cross-sectional and longitudinal studies. In an analysis of the New Zealand Household Economic Survey database, Perry (2008) showed that those aged 45-64 years made up the largest proportion (36%) of the top quintile of equivalised disposable household income (1). Further, they have the second largest proportion (28%) in the second highest quintile. These relative positions were maintained both before and after housing costs were taken into account. However, although the midlife cohort had a greater proportion living in households on higher incomes, there was still a substantial proportion living in households on lower incomes (17% in the lowest quintile and another 13% in the quintile above that).

These financial circumstances appear to translate into a similar distribution of living standards for midlife New Zealanders. The New Zealand Living Standards Report showed that 72% of those aged 45-64 years had a 'comfortable' or better living standard, second only to the 65 and over age group, with 81%. However, 17% lived in some form of hardship (Jensen et al., 2006).

A number of international studies have explored the relationship between income and wellbeing, and between social class and wellbeing. A study of more than 10,000 Swedish adults between the ages of 20 and 64 years found that those with a good financial position had a slightly higher wellbeing score than those with a less positive financial situation (Hansson, Hilleras & Forsell, 2005). They also found those aged 50-64 years had a higher level of wellbeing than those at younger ages. A similar 'clear gradient' in increases in income (both personally and nationally) and satisfaction with life has been demonstrated in an Australian study (Cummins et al., 2001). As income rose, satisfaction increased. The influence was much greater for those in households with incomes of less than $30,000 and especially those under $15,000, but still marked for those over $30,000.

The positive relationship between level of income and wellbeing has been found to be more complex in some studies. Easterlin (2006) used longitudinal data from the United States General Social Survey for the years 1973-1994 to follow happiness levels for different age cohorts. Average happiness was greatest in midlife (peaking at age 50). However, satisfaction with financial domains differed from the other domains in that it did not follow the life course pattern of people's actual financial situation. Instead it was noted that: "income rises throughout most of the working years and then levels off and declines, but satisfaction with one's financial situation moves almost inversely, starting to rise noticeably in midlife, and increasing most in late life when income, if anything, is typically declining" (p.475). This suggests satisfaction with one's financial situation may be separate from one's happiness overall as expectations lower in later life.

The relationship of wellbeing to socio-economic status has also been investigated in two of the larger longitudinal studies of ageing. Using the 2004 Survey of Health, Ageing and Retirement in Europe (SHARE) database, which contains data on people aged 50 years and over, Knesebeck et al. (2007) found that quality of life was higher for those aged 50-64 than those aged 65 or more. Those respondents aged 50-64 in the top three deciles of net worth (assets) were significantly more likely to have a higher quality of life in nine out of the ten countries (2). Regression analysis on respondents aged 50-64 showed that income, assets, education and car ownership were all associated with a higher quality of life score, while home ownership was not. The lack of association with homeownership is surprising; given it is usually one of the largest items of expenditure for most households.

Cross-sectional data drawn from three waves of the Swiss Household Panel (Vetter et al., 2006) was examined for associations between working poverty and two measures of psychological wellbeing among persons 20 to 59 years, taking into account both a low income threshold and a restricted standard of living. The findings showed that a restricted standard of living (i.e. poverty), was significantly negatively correlated with psychological wellbeing, and was also associated with increased risk of unmet mental health need. Women were affected more intensely than men. This latter finding is important, particularly because of the changing participation of midlife women in the labour market during their lifetime and their greater longevity.

A number of overseas studies have analysed the sub-populations most at risk of poverty. Applying data from the English Longitudinal Study (ELSA), Emmerson and Muriel (2008) examined changes in living standards among those aged 50 years and over. They found that single people were more likely to be in income poverty than couples. Women who were divorced, separated or widowed had the highest poverty risk. Those with low state or private pensions and those in midlife below the pension age had a much greater risk of being in poverty. Furthermore persons who moved out of the labour force, and those whose partners moved out of the labour force, were also at greater risk of being in poverty.

Vartanian and McNamara (2002) found that women's poverty in midlife (40-59 years) was strongly related to poor economic outcomes in old age, although it was only one predictor. For example, relative affluence in middle age did not necessarily preclude poverty in later life, as labour force involvement, education and marital status were also significantly related to old age economic outcomes. In a later study McNamara (2007) found that women with low incomes in midlife were unlikely to improve their income levels through their work effort unless there were additional advantages like unionisation, core sector status (as differentiated from periphery sector status) and pension plan availability.

Overall, the international research on the relationship between income, socio-economic status and wellbeing among people in midlife is fairly consistent. Higher incomes are generally positively associated with greater wellbeing and quality of life, while those on lower incomes generally report lower wellbeing. Significant associations are also found between age, gender, family type, labour market attachment and poverty. There is a higher likelihood of midlife householders being poor in the earlier midlife years if they are women, single and have low labour force attachment. Furthermore, it appears that low income or socio-economic status may have a greater impact on the wellbeing of women than of men.

Housing Tenure

The surprising lack of association between quality of life and homeownership noted above in the large scale European SHARE study (Knesebeck et al. 2007) raises the question of whether or not this is an effect of the different approaches to social housing and tenancy laws in Europe and New Zealand . In an associated study of a national random sample of 65 to 84 year olds (Waldegrave and Cameron 2009), the current authors found a modest but significantly positive association between homeownership and wellbeing. The question as to whether or not there is a relationship between tenure and wellbeing for the midlife cohort is important to our understanding of the components of poverty. Housing costs are, for most households, a major budget item and the studies below show the midlife cohort has experienced the largest percentage decrease in housing affordability when compared with other aged cohorts, and that homeownership in subsequent cohorts is declining.

The Ministry of Social Development found that the midlife group of 45-64 year olds have experienced the largest percentage increase in economic hardship in terms of housing affordability. Through the 1990s and the new millennium, the proportion of midlifers in the low affordability category increased steadily from 5% in 1988 to 19% in 2007 (Ministry of Social Development, 2008).

In an analysis of the housing data from the 2006 Census, it was found that home ownership rates for those in midlife have been steadily declining since the 1980s (DTZ New Zealand, 2007). The rate of decline in home ownership has been higher for midlife cohorts than for older cohorts, but lower than that among younger cohorts. Another New Zealand study of home ownership examined census data from 1991 to 2006, and found a trend of declining probability of ownership with age that was most pronounced for those in the lowest income quartile, and least pronounced for those in the top income quartile (Morrison, 2007).

The impact of decreasing housing affordability alongside decreasing homeownership may be pushing increasing numbers of lower income midlife householders below the poverty threshold. Furthermore a New Zealand study has shown that homeowners by late midlife (60 to 64 years) have on average have over 96% of equity in their dwellings (Statistics New Zealand and Retirement Commission, 2002), which may represent a significant advantage in lifetime wealth accumulation for homeowners compared with nonowners.

Poverty and Wellbeing and the Research Questions that arise

The literature identified above, at an international level, points to significant associations between wellbeing and key components of poverty among midlifers. These components include income, socio-economic status and assets. Associations have also been found between wellbeing and gender and family type that impact negatively on women and those not living in a partnership. Associations between wellbeing and housing tenure were not found.

This study addresses a number of research questions, with a national random New Zealand sample, that explore whether or not the same or different associations are found in a New Zealand population as for the international studies cited. Firstly it seeks to understand the overall association between poverty and wellbeing. Secondly, it seeks to explore the income, asset and poverty levels of midlifers and their housing tenure. Associations between them with gender, age and marital status are further explored. Thirdly, the relationship between each objective measure and subjective measures of 'satisfaction with economic standard of living' and overall wellbeing is investigated.

Data and Method

The data were collected as part of the Enhancing Wellbeing in an Ageing Society (EWAS) research programme (Waldegrave and Koopman-Boyden 2010). This research programme aimed to provide new knowledge that would assist policymakers and stakeholders to develop policies and interventions to achieve the best possible outcomes for all age, gender and ethnic groups in an ageing New Zealand society.

Data were collected from a national random sample of 1,958 New Zealand respondents aged between 40 and 64 years (48.9% men and 51.1% women) using computer assisted telephone interviewing (CATI). The survey took place between late January and June 2008 (Waldegrave and Cameron 2010).

The survey collected data on several objective indicators as well as subjective indicators of income and wealth. Objective indicators included income and wealth, including current personal and household income, total assets and housing tenure. Furthermore, the two most internationally used relative poverty thresholds were applied to the income data to derive measures of poverty, being: (1) the threshold applied by the OECD to compare countries, which is 50% of the median equivalent after tax household income; and (2) the European Union's Social Inclusion threshold and the United Kingdom's poverty benchmark, which are both set at the higher level of 60% of the median, equivalent, disposable, household income. The UK, along with New Zealand, also applies two further thresholds. The first is the 60% threshold after housing costs have been paid, and the second is a constant value threshold benchmarked to the 1998 median, adding the cost of living for each year thereafter. This latter threshold is the most commonly used poverty line in New Zealand, and is updated annually in the annual Ministry of Social Development's Social Report (Ministry of Social Development, 2009). To apply these measures in the current research, each respondent's total household income was first equivalised. Equivalisation is a procedure to adjust household incomes so that they are comparable between households of different size and composition. The revised Jensen Index was used, as it is the most commonly used method to equivalise incomes in New Zealand (Jensen, 1988).

For a subjective indicator of the adequacy of the respondents' income, respondents were asked to identify the adequacy of their current income with the question: "How well does your personal or household's total income meet your everyday needs?" Available responses included that the respondent had "not enough money", "just enough money", "enough money", or "more than enough money". A further subjective indicator of economic standard of living was the dichotomous response (satisfied/ not satisfied) to a question of whether the respondent was satisfied with their economic standard of living.

Finally, overall subjective wellbeing was assessed by asking about respondents' satisfaction with their life as a whole, with their responses

dissatisfied" continuum, measured on a 5 point scale. This question forms part of the World Values Survey questionnaire, first used in the 1990-1993 survey of 43 societies (excluding New Zealand), and more latterly in the 1998 and 2005 surveys in 88 countries (including New Zealand) (Inglehart et al., 1998).

The data were analysed for univariate associations of demographic factors with the objective indicators of income and wealth, i.e. income, asset ownership, housing tenure, and poverty. Furthermore, the relationships with subjective wellbeing for each objective indicator and poverty were tested using logit models (with satisfaction with economic standard of living as the dependent variable) and ordered logit models (with overall wellbeing as the dependent variable).

Results

Income

There was a 76.2% response rate (n=1491) to a question concerning total personal income before tax. This was a reasonably good response rate given that high response rates to personal questions about a person's income are difficult to achieve in any survey. The median personal income before tax was $45,000 reflecting the higher incomes of the midlife group when compared with other life stages. However, not all midlifers lived on high incomes. The data showed that 15.2% of respondents lived on pre-tax incomes of $15,000 or less, and that 22.0% lived on pre-tax incomes of $20,000 or less. There were also substantial gender differences as Table 1 shows. The median income for women at $31,137 was just 59% that of men at $56,912 (3).

There were differences in median income by age (4) and widows had a significantly lower median income than married people (5), also shown in Table 1, confirming the trends identified in the literature referred to earlier.

This objective indicator of personal income was then compared with a number of subjective indicators. Respondents were asked how well their total income met their everyday needs for accommodation, food, clothing and other necessities. Subjectively, over 40% of respondents considered they had 'enough money', and a further 16.6% said they had 'more than enough money'. However, 28.6% said they had 'just enough money', and 14.8% said that they had 'not enough money'. Despite the differences in absolute income noted above, there were no statistically significant differences in the subjective assessment of the adequacy of income between men and women.

When this subjective assessment of income was compared with responses to the actual personal income question, a significant association was found (6). Just 40.6% of those with less than $20,000 of personal income said they had enough or more than enough money, compared to 56.9% of those with personal income between $40,000 and $60,000, and 83.4% of those with personal income over $80,000. More people in the lowest income bracket (29.8%) stated they had not enough money, compared with just 3.1% of those in the highest income bracket.

Assets

An exploration of the asset wealth of midlife New Zealanders can further our understanding of the relationship between poverty and wellbeing. There was an 81.7% response rate (n=1599) in the EWAS survey to a question concerning the total value of household assets not including the value of the family home (7).

The data showed that 37.6% of respondents had no assets at all other than a family home (if they had one), and a further 18.4% possessed assets worth $100,000 or less. However, around a fifth of respondents had assets worth more than $500,000, which together with the lower levels of home ownership noted later demonstrate a significant level of asset inequality in this group.

As with personal income, men had significantly more total assets in value than women (8). Among women, 42.4% had no assets (other than the family home where one was owned) compared with 33% of men, while 19.4% of women had assets of $500,000 or more compared with 23.5% of men.

The association between total assets and age was much weaker than that between personal income and age (9). The tendency was for younger participants to possess fewer assets than the older ones. For example, 42.2% of the 40 to 44 year olds had no assets apart from the family home, compared with 30.6% of the 60 to 64 year olds. Conversely, 8.8% of 40 to 44 year olds possessed assets worth over $1 million, while 15.0% of 60 to 64 year olds did. This result is consistent with assets accumulating with age, despite the lower median income of the 60 to 64 year olds compared with the other age categories. This pattern has been noted in the literature (Statistics New Zealand, 2008; Department of Prime Minister and Cabinet, 2008).

There was also a significant association between asset wealth and current marital status, just as there was

for personal income (10). While 54.1% of single, 53.8% of widowed, and 66.7% of divorced or separated people had no assets beyond the family home, just 30.9% of married or partnered respondents possessed no assets. This pattern is similar for both men and women, although the proportions of women with no assets beyond the family home are slightly higher within each marital status category than the proportions of men--while 65.5% of single women, 56.8% of widowed, and 71.5% of divorced or separated women had no assets beyond the family home, just 33.7% of married or partnered men possessed no assets.

Housing Tenure

The EWAS research programme involved two surveys, the midlife survey addressed in this paper and a second survey of a national random sample of 1,680 older citizens aged 65 to 84 years (Waldegrave & Cameron, 2009). The increasing number of renters as opposed to home owners noted in the literature is born out in a comparison of the two studies. Whereas 7.6% of the older group rented, 16.1% of the midlife group did. Both groups demonstrated high levels of home ownership with mortgaged and freehold houses comprising 76.4% and 77.5% respectively, but when houses in retirement villages, family trusts and homes owned by other family members are added the midlife respondents showed 83.4% in some form of home ownership, considerably less than the 92% for the older group.

These results are consistent with the studies referred to in the literature section, which showed a steady decline in home ownership, particularly with the younger cohorts and those on low incomes (DTZ New Zealand, 2007; Morrison, 2007; Ministry of Social Development, 2008). As a consequence, the numbers renting have been increasing. Within the midlife sample considered here, no significant association was found between housing tenure and gender, but a significant association was found with age (11). Older participants were more likely to be owners, and the percentage of renters progressively decreased across each successively older cohort, which is consistent with the literature noted earlier (DTZ New Zealand, 2007; Ministry of Social Development, 2008).

Housing tenure was also significantly related to marital status (12). Around 88.0% of each category of married/partnered people and widowed people owned their own home, either by themselves, with other family members, or through a family trust, whereas only 60.6% of single people and 67.0% of divorced or separated people did.

As would be expected, housing tenure was significantly related to participants' income and the total value of their assets other than the family home (13). When the higher and lower income groups were compared, a greater proportion of those on higher incomes were owners and a greater proportion of those on lower incomes were renters. Among owners, 32.7% had no other assets, compared with 60.0% of renters and 37.2% of others. These findings are consistent with Morrison's study noted earlier (Morrison, 2007) and demonstrate the links between income, home ownership and protection against poverty.

These results are of concern because they suggest that inequality in wealth is likely to increase as those who are married and on higher incomes are more likely to be able to consolidate a substantial asset (housing), while those who are not partnered or on lower incomes are more likely to be left to the vicissitudes of the market.

Poverty

The standard relative poverty measures, noted in the Data and Methods section of this paper (i.e. the OECD 50% and European Union/UK 60% of median, equivalent, disposable, household income), were applied to the 921 respondents who answered both income and housing cost questions. The number of people between the ages of 40 and 64 years in these households was 1,669. A third measure was also applied, which was the constant value threshold

benchmarked to the 1998 median with the addition of the cost of living each year since then. As noted earlier, this is the threshold used in New Zealand's Social Report each year, and is the commonly used measure of poverty in New Zealand (Ministry of Social Development, 2008).

Table 2 provides a summary of the three measures and their comparison with the Ministry of Social Development results. It is important to note that the population for the EWAS study comprised a national random sample of people between the ages of 40 and 64 years, whereas the Ministry of Social Development calculations apply the Household Economic Survey figures, also a national random sample, to the whole population for the 50 and 60% of median income thresholds. Their constant value threshold was, however, grouped by age. The comparison in this discussion is made between MSD's 45-64 year old group and EWAS's 40 to 64 year old group.

The EWAS data showed 10.8% of the sample was below the 60% constant value poverty threshold, which is almost identical to the Ministry of Social Development's figure of 11% for a similar age group (Perry, 2008). The EWAS sample recorded lower levels of poverty than the Ministry of Social Development's results for the 50% and 60% poverty thresholds for the whole population, indicating the higher incomes of the midlife cohort when compared with the total population as noted in the literature section. Table 2 shows that these differences were minimal at the 50% of median income threshold, but more substantial at the 60% threshold.

The results from a multivariate logit model of poverty are presented in Table 3, using the 50% poverty threshold (14). These results show that within this midlife sample, older cohorts were significantly more likely to be in poverty, but gender and asset ownership were not consistently significantly related to poverty. However, marital status appears to have a significant impact--singles are 16 times more likely to be in poverty than married/partnered people, widowed people are more than 19 times more likely to be in poverty, and divorced or separated people are more than 10 times more likely to be in poverty. Housing tenure also has a significant effect, with home owners being 55% less likely to be in poverty than non-owners.

Income, Assets, Tenure, Poverty and Wellbeing

The results from a multivariate logit model of subjective satisfaction with economic standard of living are presented in Table 4. They demonstrate that those on higher incomes were more likely to be satisfied with their economic standard of living, women were more likely to be satisfied than men, and those with more assets and married or partnered people were more likely to be satisfied. However, holding other variables constant, home ownership does not appear to be associated with higher satisfaction with economic standard of living. If the income variable is replaced by a poverty measure, the poor are shown to be between 42% and 50% less likely to be satisfied with their economic standard of living than the non-poor.

The results from a multivariate ordered logit model of overall wellbeing are presented in Table 5. They demonstrate that overall wellbeing was significantly positively associated with income and was significantly higher for married people than for others. If all other variables are held constant, wellbeing is higher among the older age cohorts and is higher among women than men. The relationship between overall wellbeing and total assets appears to be in the shape of an inverted 'u', with the highest level of overall wellbeing occurring among those in the middle categories of total asset ownership. As with satisfaction with economic standard of living, home ownership does not appear to have a positive association with overall wellbeing holding other variables constant. Furthermore, if the income variable is replaced by a poverty measure, the poor are shown to have significantly lower overall wellbeing.

Conclusion and Discussion

The results of this survey of midlife New Zealanders are broadly consistent with other New Zealand and international studies. They demonstrate the higher incomes and wealth of this cohort when compared with other cohorts, as most people have become established in their careers and family life. Although they also showed relatively high levels of home ownership, the proportion renting was more than double that of older cohorts aged 65 to 84 years.

It is clear that there is a substantial

minority of people in midlife who live on low incomes, with few assets and often in rented accommodation. Around 15% lived on incomes below $15,000, nearly 3 8% had no assets apart from the family home (if they had one) and 16% lived in rental accommodation. Applying the 60% constant value poverty threshold most commonly used in New Zealand, 10% of people were poor. As the midlife population is expected to double the elderly population over the next 30 years (Statistics New Zealand, 2007), based on these results the numbers of older people living in considerable hardship could be expected to grow substantially.

Significant gender differences were demonstrated for income and assets but not for housing tenure. Men had higher personal incomes and greater asset wealth. However, interestingly, when all other variables were held constant, women and older midlifers showed greater overall wellbeing. The age result is consistent with Easterlin's (2006) findings that wellbeing increased almost inversely to income declining. The gender result suggests that material (financial) wellbeing is of greater subjective importance to men than to women.

The associations between age and income and assets were modest. However, age was significantly associated with housing tenure, as home ownership increased with age and renting decreased. It was also associated with poverty, with older cohorts significantly more likely to be below the threshold than their younger counterparts. Non-married or nonpartnered people were significantly more likely to be living in poverty, have lower incomes and assets and be renters rather than homeowners. Renters were also significantly more likely to be living in poverty.

Most subjective measures showed significant associations with the objective data. The subjective indicator of perceived 'adequacy of money' was significantly related to actual personal income. However, despite significant differences in actual personal income, there were no differences in subjective perceptions of the adequacy of income between men and women. Income and assets, but not housing tenure, were significantly associated with respondents' subjective satisfaction with their economic living standards. Overall wellbeing was positively associated with income, with assets (not including the family home) up to half a million dollars, with marriage or partnership status and was negatively associated with poverty. These results are consistent with the international literature cited earlier suggesting socioeconomic and marital/partnership statuses are substantial contributors to the wellbeing and quality of life of midlife New Zealanders. However, the socio-economic result was stronger for men than for women indicating a greater emphasis for women of non-material assets.

From the perspective of the impact of poverty on wellbeing, the converse is also true. Low incomes, few assets and, as already noted, poverty in midlife were negatively associated with wellbeing. Unexpectedly, housing tenure, though significantly associated with income and assets and negatively associated with poverty, showed no significant association with either a subjective satisfaction with economic standard of living or overall wellbeing. The non-significant association with overall wellbeing is consistent with the results of the SHARE study (Knesebeck et al 2007).

As prime goals of good social and economic policy consist of keeping people out of poverty and enhancing the wellbeing of citizens, the key policy issues arising from the results of this research centre on the low level of asset accumulation for the majority of participants and the declining rates of home ownership suggesting many older citizens will have substantially higher housing costs in the future. For instance, as many as 56% of respondents had assets of $100,000 or less, most of whom possessed no assets other than the family home, suggesting there will be serious fiscal challenges for policy makers as this cohort moves in large numbers into old age and retirement. Midlife is usually the time of peak income earning prior to retirement. However, the combination of nearly 11% of participants living below the Ministry of Social Development's low income threshold, low asset ownership, and the decline in home ownership together pose a serious challenge for future policy if levels of hardship among older people are not to increase.

As the combination of a reasonably generous (by international standards) universal superannuation scheme and high levels of home ownership largely protect the current elderly population from poverty, so the combination of adequate income and decent and affordable housing will be necessary as the current midlife cohorts move into older age.

Policy solutions will need to be found for containing the costs of home ownership for those who own their houses, and greater investment in social housing for most of those who don't. Market rents can be expected to be beyond the affordability threshold for older people living largely on superannuation and minimal assets, despite the assistance of the Accommodation Supplement. Sustainable solutions for the ongoing payment of superannuation in a way that preserves its purchasing power will also need to be devised if older people are to live free of poverty, particularly in the light of the recent Government decision to suspend contributions to the New Zealand Superannuation Fund (English, 2009).

Author Note

This paper reports on results from the Enhancing Wellbeing in an Ageing Society (EWAS) research programme which is jointly led by the Family Centre Social Policy Research Unit in Lower Hutt, Wellington, New Zealand and the Population Study Centre at the University of Waikato, Hamilton New Zealand. EWAS has been funded by the Foundation for Research, Science and Technology (FRST).

References

Cummins, R., Eckersley, R., Pallant, J., Davern, M., & Misajon, R.(2001). Special report on income and geographic location. Australian Unity Well-Being Index: Survey 2: Report 2. Melbourne: Australian Centre on Quality of Life, Deakin University.

Department of the Prime Minister and Cabinet. (2008). Families in Australia 2008. Canberra: Department of the Prime Minister and Cabinet.

DTZ New Zealand. (2007). Census 2006 and housing in New Zealand, Centre for Housing Research Aotearoa New Zealand and Building Research: Research Report. Wellington: Centre for Housing Research Aotearoa New Zealand and Building Research.

Easterlin, R. (2006). Life cycle happiness and its sources. Intersections of psychology, economics, and demography. Journal of Economic Psychology, 27, 463-482.

Emmerson, C., & Muriel, A. (2008). Financial resources & well-being. In J. Banks, E. Breeze, C. Lessof & J. Nazroo (Eds.) Living in the 21st century: Older people in England--The 2006 Longitudinal Study of Ageing (Wave 3) (pp 118-149). London: Institute of Fiscal Studies.

English, Hon B. (2009) 2009 Budget,, Wellington: New Zealand Government, Retrieved from http://www.treasury.govt. nz/budget/2009

Hansson, A., Hilleras, P., & Forsell, Y. (2005). Well-being in an adult Swedish population. Social Indicators Research, 74, 313-325.

Inglehart, R., Basenez, M., & Moreno, A. (1998). Human values and beliefs: A cross-cultural sourcebook. Michigan: Michigan University Press.

Jensen, J. (1988). Income equivalences and the estimation of family expenditures on children. Wellington: Department of Social Welfare.

Jensen, J., Krishnan, V., Hodgson, R., Sathiyandra, S., Templeton, R., Jones, D., Goldstein-Hawes, R., & Beynon, P (2006). Living standards report 2004 nga ahuatanga noho o Aotearoa. Wellington: Centre for Social Research and Evaluation, Ministry of Social Development.

Knesebeck, O. von dem, Wahrendorf, M., Hyde, M., & Siegrist, J. (2007). Socioeconomic position and quality of life among older people in 10 European countries: results of the SHARE study. Ageing & Society, 27, 269-284.

McNamara, J. (2007). Long-term disadvantage among elderly women: The effects of work history. Social Service Review, September 2007, 423-452.

Ministry of Social Development. (2009). The social report 2008: Social Indicators for New Zealand. Wellington: Ministry of Social Development.

Morrison, P. (2007). On the falling rate of home ownership in New Zealand. Wellington: Centre for Housing Research Aotearoa New Zealand

Perry, B. (2008). Household incomes in New Zealand: trends in indicators of inequality and hardship 1982 to 2007. Wellington: Ministry of Social Development

Statistics New Zealand. (2008). Family net worth in New Zealand. Wellington: Statistics New Zealand.

Statistics New Zealand. (2007). National Population Projections: 2006 (base)--2061. Wellington: Statistics New Zealand

Statistics New Zealand & Retirement Commission. (2002). The net worth of New Zealanders: A report on their assets and debts. Wellington: Statistics New Zealand and Retirement Commission.

Vartanian, T., & McNamara, J. (2002). Older women in poverty: The impact of midlife factors. Journal ofMarriage and Family, 64, 532-548.

Vetter, S., Endrass, J., Schweizer, I., Teng, H., Rossler, W., & Gallo, W. (2006). The effects of economic deprivation on psychological well-being among the working population of Switzerland. BMC Public Health, 6, 223. Retrieved from http://www.biomedcentral.com/1471 2458/6/223

Waldegrave, C., & Cameron M. (2009) Income, assets, living standards and housing. In P. Koopman-Boyden, & C. Waldegrave (Eds.) Enhancing Wellbeing in an Ageing Society--65-84 year old New Zealanders in 2007 (pp. 83-108). Hamilton: Population Studies Centre, University of Waikato and The Family Centre Social Policy Research Unit, Lower Hutt, Wellington.

Waldegrave, C., & Cameron M. (2010 in print) Income, assets, living standards and housing among midlife New Zealanders. In C. Waldegrave & P. Koopman-Boyden (Eds.) Midlife New Zealanders Aged 40 to 64 in 2008: Enhancing Wellbeing in an Ageing Society (pp. 75-99). Hamilton: Population Studies Centre, University of Waikato and The Family Centre Social Policy Research Unit, Lower Hutt, Wellington.

Waldegrave, C. & Koopman-Boyden, P. Midlife New Zealanders Aged 40 to 64 in 2008: Enhancing Wellbeing in an Ageing Society. Hamilton: Population Studies Centre, University of Waikato and The Family Central Social Policy Research Unit, Lower Hutt, Wellington.

Footnotes

(1) Equivalised income refers to income after applying equivalence scales. Equivalence scales enable a calculation of the income required for households of different sizes and ages to achieve an "equivalent" standard of living. They are often used in poverty measurement where it is recognised, for example, that the same standard of living for a one parent and one child family will require a different amount of income than a two parent and three children family. Disposable income is income after the payment of taxation. Household income refers to the sum of income received by household members.

(2) The nine countries were Austria, Denmark, France, Germany, Greece, Italy, Netherlands, Spain, and Sweden. The associations were particularly strong in France and Germany, but not significant in Switzerland.

(3) p-value of <0.001 based on quantile regression.

(4) p-value of 0.014 based on quantile regression.

(5) Based on quantile regression, widowed people have a lower median income (p=0.028) compared to married/ partnered people, but single people (p=0.340) and divorced or separated people (p=0.199) do not.

(6) Chi square statistic had a p-value of <0.001

(7) The family home was not included in this analysis because if it was sold it would have to be replaced in some form, i.e. another house, a retirement village, ongoing rental, etc.

(8) Chi square statistic had a p-value of 0.008

(9) Chi square statistic had a p-value of 0.040

(10) Chi square statistic had a p-value of <0.001

(11) Chi square statistic had a p-value of <0.001

(12) Chi square statistic had a p-value of <0.001. Housing tenure when compared with Marital Status was collapsed into 2 categories from 3. These were 'Owners' and 'Renters'. The third category entitled "Others' consisted of only 1.4% of participants and the cell size was too small for a reliable chi square calculation.

(13) Chi square statistic had a p-value of 0.001 for income and <0.001 for asset total. Housing tenure when compared with income was collapsed into 2 categories from 3. These were 'owners' and 'renters'. The third category entitled "others' consisted of only 1.4% of participants and the cell size was too small for a reliable chi square calculation.

(14) Qualitatively similar results are obtained when other poverty lines are used to define those who are poor.

Charles Waldegrave, Family Centre Social Policy Research Unit, Lower Hutt, Wellington

Michael P. Cameron, University of Waikato

Corresponding Author:

Charles Waldegrave

Family Centre Social Policy Research Unit

Lower Hutt

Wellington, New Zealand

waldegrave.c@fc.org.nz
Table 1. Average Total Personal Income, by Gender, Age and
Marital Status

                     Median Income   Mean Income

Gender                     $              $

Male                     56,912         79,292
Female                   31,137         49,507

Age Cohort

40-44                    45,000         65,370
45-49                   48, 511         64,142
50-54                    45,000         62,067
55-59                    45,000         68,635
60-64                    35,000         68,038

Marital Status

Single                   40,103         60,204
Married/ Partnered       45,000         66,230
Widowed                  45,000         53,025
Divorced                 40,000         60,862

Table 2. Numbers below Poverty Thresholds from the EWAS and MSD
Studies (%)

Poverty Threshold            EWAS: people aged   MSD: people
                             40-64

50% of median                       8.9          10 (whole population)
60% of median                      12.6          18 (whole population)
60% constant value used in         10.8          11 (ages 45-64)
  Social Report

Table 3. Multivariate Logit Model of Poverty, using the 50% of
Median Income Poverty Line

Variable                                  Odds Ratio      p-value

Age                                          1.0582        0.010 **
Gender (1 = male)                            1.0399        0.891
Non-Home Assets $0-$50,000 ([dagger])        0.5170        0.147
Non- Home Assets $50,000-$100,000            0.7866        0.593
Non-Home Assets $100,000-$250,000            0.3455       0.079 *
Non-Home Assets $250,000-$500,000            0.4054       0.084 *
Non-Home Assets $500,000-$1 million          0.0924       0.032 *
Non-Home Assets $1 million-$1.5 million      0.8668        0.845
Single ([double dagger])                    16.2139     <0.001 ***
Widowed                                     19.3618     <0.001 ***
Divorced or Separated                       10.8394     <0.001 ***
Home Owner (1 = yes)                         0.4470      0.007 ***

* significant at 10% level, ** significant at 5% level, ***
significant at 1% level, ([dagger]) comparator group is those
with no assets, no respondents with more than $1.5 million in
assets were poor with this poverty line, ([double dagger])
comparator group is married or partnered people

Table 4. Multivariate Logit Model of 'Satisfaction with Economic
Standard of Living'

Variable                                    Odds Ratio      p-value

Age                                         1.0218         0.097 *
Gender (1 = male)                           0.6625         0.020 **
Income                                      1.000006       0.068 *
Non-Home Assets $0-$50,000 ([dagger])       1.5496         0.089 *
Non- Home Assets $50,000-$100,000           2.2571         0.009 ***
Non-Home Assets $100,000-$250,000           3.2419         0.001 ***
Non-Home Assets $250,000-$500,000           4.3737        <0.001 ***
Non-Home Assets $500,000-$1 million         7.7721        <0.001 ***
Non-Home Assets $1 million--$1.5 million    5.2306         0.003 ***
Non-Home Assets $1.5 million--$2 million    7.3526         0.056 **
Non-Home Assets over $2 million             8.3967         0.005 ***
Single ([double dagger])                    0.6086         0.074 *
Widowed                                     0.2933         0.001 ***
Divorced or Separated                       0.4351        <0.001 ***
Home Owner (1 = yes)                        1.2831         0.251

* significant at 10% level, ** significant at 5% level, ***
significant at 1% level,

([dagger]) comparator group is those with no assets, ([dopuble dagger])
comparator group is married or partnered people

Table 5. Multivariate Ordered Logit Model of Overall Wellbeing

Variable                                    Coefficient      p-value

Age                                         0.2317          0.008 ***
Gender (1 = male)                          -0.4081          0.001 ***
Income                                      0.0000019       0.049 **
Non-Home Assets $0-$50,000 ([dagger])       0.4902          0.016 **
Non- Home Assets $50,000-$100,000           0.4894          0.023 **
Non-Home Assets $100,000-$250,000           0.6079          0.004 ***
Non-Home Assets $250,000-$500,000           0.7677         <0.001 ***
Non-Home Assets $500,000-$1 million         0.3099          0.115
Non-Home Assets $1 million-$1.5 million     0.5814          0.038 **
Non-Home Assets $1.5 million-$2 million     0.0350          0.924
Non-Home Assets over $2 million             0.5799          0.091 *
Single ([double dagger])                   -1.2670         <0.001 ***
Widowed                                    -1.1001          0.001 ***
Divorced or Separated                      -0.9535         <0.001 ***
Home Owner (1 = yes)                        0.1679          0.322

* significant at 10% level, ** significant at 5% level, ***
significant at 1% level,

([dagger]) comparator group is those with no assets, ([double
dagger]) comparator group is married or partnered people
Gale Copyright: Copyright 2010 Gale, Cengage Learning. All rights reserved.