aja_268.

fm Page 20 Monday, February 4, 2008 10:31 AM

DOI: 10.1111/j.1741-6612.2007.00268.x

Research
Blackwell Publishing Asia

Correlates of housing affordability stress among older Australians

Jeromey B Temple In the context of these important health consequences, an

Australian Demographic and Social Research Institute, Institute of examination of the ability of older Australians to meet shelter
Advanced Studies, The Australian National University, Acton, and utilities costs is particularly significant. The purpose of this
Australian Capital Territory, Australia
paper is two fold: first, to use a new variable available in the
Objectives: The purpose of this study was to examine the 2002 ABS General Social Survey (GSS) to measure the pre-
prevalence and correlates of housing affordability stress valence of ‘housing affordability stress’; and second, to examine
among community-dwelling older Australians. the potential social, demographic and economic correlates of
Methods: The 2002 ABS General Social Survey was used to housing affordability stress among community-dwelling older
measure the prevalence of housing affordability stress. Rare Australians.
event logistic regression was used to measure the potential
correlates of housing affordability stress. Measurement of housing affordability stress
Results: Almost 5% of Australians aged 55 years and older, A continuing debate in the housing cost literature is the
and 20% of those younger than 55 years, are estimated to measurement of housing affordability. Traditionally, housing
experience housing affordability stress. Men and women living affordability is measured using the share of housing costs out
alone are more likely to experience affordability stress when of current income. Once a normatively determined budget
compared to couples. Low-income earners, those with a share is reached, the household is considered to have ‘excessive’
consumer debt or who do not hold assets, are at a heightened housing costs. Different definitions of what is an excessive
risk of such stress. Home ownership, regardless of income, is budget share have been used: for example 50, 28, 25, and
the strongest buffer against housing affordability problems in 35%, but the most common is 30% [10–12].
old age.
Conclusion: Although the prevalence of housing affordability As pointed out by numerous authors, high budget shares for
stress is low among older Australians when compared to the housing may also reflect heterogeneity of preferences for
younger population, a definite social gradient exists in those at housing consumption, savings and wealth [13], that is, they
risk. have a preference to spend on housing to obtain better housing
quality or are prone to overspend [11,14–16]. Attempts to
Key words: aged, housing, poverty, socioeconomic factors. improve the measurement of affordability have been made by
censoring analysis to lower income earners [17–20]. Since the
early National Housing Strategy, the Australian government
Introduction
The ability to reside in housing that is affordable and of has tended to take a different stance on what constitutes an
sufficient quality contributes to ‘successful ageing’ [1]. In particular, appropriate measure of housing affordability or housing stress.
the ability to reside in affordable housing, independently in the For example, commenting on the measurement of housing
community is an important factor containing the additional affordability and Commonwealth Rent Assistance;
public expenditure required because of population ageing [2]. The adoption of an affordability benchmark would fail to
recognise the element of choice exercised by customers who
Research has found that housing affordability problems place a higher value on housing then others in comparable
influence subjective well-being, the number of general practitioner circumstances. p. 1672 [21]
consultations and the likelihood of poorer health status,
poorer mental status and stress [3 – 6]. Apart from shelter costs What then is an acceptable or affordable level of housing
generally, the ability to pay for utilities is particularly important costs? In Hancock’s (1993) key study, he argues that it is
for older people who are at greater risk of hypothermia when important to account for heterogeneity in preferences,
a dwelling is inadequately cooled or heated [7,8]. Similarly, economic resources and social structure [22]. He concludes, ‘In
during periods of hot or cold weather, older people living in spite of explaining that “affordability” and “affordable rents”
inadequately cooled or heated dwellings may be at a higher are only distinctly related, many may yet wonder how a
risk of mortality [9]. To cut back on general shelter costs, older definition of affordable housing can be used to determine an
people may reduce their expenditure on utilities [10]. affordable rent. The answer is, that any rent will be affordable
which leaves the consumer with a socially acceptable standard
of both housing and non-housing consumption after the rent
Correspondence to: Dr Jeromey B Temple, Australian Demographic is paid’ (p. 144). The ability to meet rent or mortgage payments
and Social Research Institute, Institute of Advanced Studies, The on time is one such indicator of a socially accepted standard of
Australian National University. Email: [email protected] consumption.

20 Australasian Journal on Ageing, Vol 27 No 1 March 2008, 20–25

© 2008 The Authors
Journal compilation © 2008 ACOTA
aja_268.fm Page 21 Monday, February 4, 2008 10:31 AM

A g e i n g a n d h o u s i n g a f f o r d a b i l i t y

Methods Results
Data for this study are drawn from the 2002 ABS GSS. The Using the measure available in the ABS GSS, the prevalence of
2002 ABS GSS was conducted by the Australian Bureau of housing affordability stress is significantly lower among older
Statistics between March and July 2002, on a multistage than younger Australians. About 20% of respondents younger
area sample of dwellings. An initial sample of 19 500 dwell- than 55 report that they have difficulty meeting their housing
ings was drawn, from which 17 000 dwelling were used due and utility bills on time, when compared to just 5% of those
to uninhabited dwellings or dwellings with residents outside aged 55 and older. Although the likelihood of housing afford-
of the scope of the survey. Of the 17 000 dwellings, 15 500 ability stress is low relative to younger Australians, results
fully responded. The confidentialised data file contains a from the descriptive analysis (Table 1) indicate that there is a
sample of people aged 18 years and older and who lived in definite social gradient in those at risk.
a private dwelling. This study is restricted to an analysis
of 4354 Australians aged 55 years and older. For com- Table 1 tabulates the propensity to be in housing affordability
pleteness, this prevalence of housing affordability stress stress by demographic, economic and housing characteristics
is also compared with those younger than 55. These data of the household. Significant differences in the propensity to
were made available to the author through an agreement have affordability problems are calculated using Wald tests of
between the ABS and the Australian Vice Chancellors proportions. Compared to women living alone (5.4%), couples
Committee. are at a lower risk of housing affordability problems (2.2%),
whereas members of ‘other’ household types are at a con-
An important component of older people’s household expend- siderably higher risk of having difficulties meeting housing and
iture is the ability to ‘afford’ adequate housing. Rather than utility payments on time (8.8%). Although about 7.5% of men
enter this debate of what contributes an affordable budget living alone report having affordability problems, this figure is
proportion, I utilise a unique variable in the 2002 ABS GSS. By only marginally higher than that for women living alone
directly asking whether the household has enough money to (P < 0.10). Among the other demographic factors, the risk of
pay for shelter or utility costs on time, the GSS provides a affordability problems appears to decline with age and is also
self-reported measure of housing affordability. One limitation lower among migrants from an English-speaking background
of this measure is that respondents may feel embarrassed about when compared to the Australian born.
reporting difficulty paying for housing and utilities and
therefore bias the prevalence downward. Apart from demographic factors, the propensity to have
housing affordability problems in old age differs by economic
The dichotomous dependent variable, housing affordability and housing characteristics. Not surprisingly, only about 2%
stress is coded ‘1’ if the household reported difficulty in paying of owner occupiers report difficulty meeting shelter costs,
shelter or utilities costs on time, within the last 12 months, and when compared to almost 16% of private renters and about
is coded ‘0’ if the household was able to pay bills on time over 10% of purchasers and public renters. High income too plays
this same period. In cases of rare dichotomous events, the a protective role in protecting older people from housing
dependent variable is unweighted, leading to the potential for affordability problems. About 7% of respondents in the lowest
misspecified logit coefficients [23]. I specify a rare event logit 20% of the income distribution reported affordability problems,
model to estimate the independent effect of demographic, eco- compared with less than half of 1% of those among the top
nomic and housing factors upon the probability of housing 20% of income earners. Similarly, holding investments in cash,
affordability stress among older Australians. The rare event business or bonds and not holding consumer debt is protective
logit model re-corrects the parameter coefficients and results in against having difficulties meeting shelter payments on time.
more accurate predicted probabilities [24]. The estimation
procedure provides accurate logit coefficients in the presence In contrast, there are no differences in the risk of housing
of rare events through a technique that yields a lower mean affordability stress by level of education, region of residence or
squared error. The full mathematical proof is given in detail in labour force status. Although these descriptive results are
King and Zeng [23]. informative, any conclusions drawn about the potential
correlates of housing affordability stress are misleading
With the full model specified, visual tests of the error distri- without controls for confounding factors. For example, once
bution are conducted. To ensure the robustness of the model controlling for income, does tenure remain an important factor
specification, the standard errors are re-estimated using a explaining housing affordability stress?
non-parametric bootstrap based on the biased corrected
method. The bootstrapped standard errors are highly Table 2 displays the results from the Rare Event Logit (REL)
comparable to those derived, lending support to the model regression model that enables an examination of the factors
specification. Moreover, a range of interaction terms were associated with housing affordability stress, after controlling
included in the regression to ensure correct specification of for all factors in the model. The rare event adjusted logit
the model. Prior to estimation of the regression model, coefficient β a is presented which measures the influence of each
descriptive analyses were conducted using Wald tests of covariate upon the probability of housing affordability stress
proportions. after controlling for all other factors in the model. In particular,

Australasian Journal on Ageing, Vol 27 No 1 March 2008, 20–25 21

© 2008 The Author
Journal compilation © 2008 ACOTA
aja_268.fm Page 22 Monday, February 4, 2008 10:31 AM

Te m p l e J B

Table 1: Prevalence of Housing Affordability Stress among Table 2: Rare Event Logistic regression model of factors
older Australians, 1998/1999 associated with housing affordability stress, Australians
aged 55 and older, 2002
n Percentage in stress
βa Odds ratio (OR) % z
Demographic factors
Living arrangements Demographic factors
Lone women 1129 5.40 – Living arrangements
Couple 2081 2.16 *** Lone woman –
Lone men 598 7.53 * Couple –0.815 0.443 –55.7 –3.24 ***
Other 546 8.79 ** Lone man –0.379 0.685 –31.5 –1.58
Age Other 0.214 1.239 23.9 0.78
55– 64 ‘young-old’ 1919 6.98 – Age
65 –74 ‘old’ 1344 3.27 *** 55– 64 ‘young-old’ –
75+ ‘oldest-old’ 1091 1.92 *** 65– 74 ‘old’ –0.663 0.515 –48.5 –3.22 ***
Country of birth 75+ ‘oldest-old’ –1.056 0.348 –65.2 –3.83 ***
Australia 2977 4.53 – Country of birth
Mainly English speaking 650 2.92 ** Australia –
Other countries 727 6.19 * Mainly English speaking –0.583 0.558 –44.2 –2.08 **
Education Other countries 0.127 1.135 13.5 0.60
University education – No 3856 4.69 – Education
University education – Yes 498 3.61 University education – No –
Region of residence University education – Yes 0.054 1.055 5.5 0.18
Major cities 2547 4.67 – Region of residence
Inner regional 982 4.07 Major cities –
Other areas 825 4.85 Inner regional –0.493 0.611 –38.9 –2.36 **
Disability or health condition Other areas –0.298 0.742 –25.8 –1.38
Yes 1599 2.56 – Disability or health condition
No 2755 5.74 *** Yes –
Self-reported health No –0.345 0.708 –29.2 –1.58
Excellent/very good 1634 2.33 – Self-reported health
Good/fair 2266 4.85 *** Excellent/very good –
Poor 454 11.23 *** Good/Fair 0.372 1.451 45.1 1.69 *
Housing factors Poor 0.861 2.366 136.6 3.08 ***
Owner 3074 1.79 – Housing factors
Purchaser 449 9.13 *** Owner –
Private renter 362 15.47 *** Purchaser 1.261 3.529 252.9 5.04 ***
Public renter 367 10.35 *** Public renter 0.937 2.552 155.2 3.97 ***
Other 102 8.82 ** Private renter 1.043 2.838 183.8 4.25 ***
Economic factors Other 1.404 4.071 307.1 3.31 ***
Equivalent income‡ Economic factors
Bottom 20% 1793 7.08 – Equivalent income†
20– 40% 975 4.00 *** Bottom 20%
40– 60% 612 2.94 *** 20– 40% –0.477 0.621 –37.9 –2.05 **
60– 80% 411 2.92 *** 40– 60% –1.108 0.330 –67.0 –3.29 ***
Top 20% 563 0.53 *** 60– 80% –1.040 0.353 –64.7 –2.66 ***
In the labour force? 80 –100% –2.705 0.067 –93.3 –4.17 ***
Yes 1189 5.21 – In the labour force
No 3165 4.33 No –
Investment in cash, business, bonds etc? Yes –0.510 0.600 –40.0 –2.10 **
Yes 3391 1.77 – Investment in cash, business, bonds, etc.?‡
No 963 14.43 *** Yes –
Household has consumer debt? No 1.406 4.080 308.0 7.51 ***
Yes 931 10.85 – Household has consumer debt?‡
No 3423 2.86 *** No –
Yes –1.017 0.362 –63.8 –5.57 ***
Source: 2002 ABS General Social Survey. Constant term –1.968 –4.71 ***
***P < 0.001; **P < 0.01; *P < 0.05; †P < 0.10; – comparison category; ‡Equivalent income
(using Organisation for Economic Co-operation and Development equivalent scale) collapsed into Source: 2002 ABS General Social Survey.
quintiles; §A respondent is recorded as having an asset if (i) holding over $Aus1000 in cash or *P < 0.10; **P < 0.05; ***P < 0.01; – comparison category; Likelihood ratio tests that all
deposited in financial institutions, or (ii) owning an incorporated business, or (iii) has shares, stocks covariates are jointly zero; †Equivalent income (using Organisation for Economic Co-operation
or bonds, or (iv) investment property. A household is coded as not holding assets if they do not and Development equivalent scale) collapsed into quintiles. ‡A respondent is recorded as having
hold any of the above. ¶A respondent is recorded as having consumer debt if they hold debt on an asset if (i) holding over $Aus1000 in cash or deposited in financial institutions, or (ii) owning
(i) credit cards or store cards not paid off by the due date, or (ii) car loans or personal loans, or an incorporated business, or (iii) has shares, stocks or bonds, or (iv) investment property. A
(iii) interest-free purchase, or (iv) hire purchase agreements, or (v) ‘other’ consumer debt. A household is coded as not holding assets if they do not hold any of the above. §A respondent is
respondent is coded as not having any consumer debt if they do not report holding debt in any of recorded as having consumer debt if they hold debt on (i) credit cards or store cards not paid off
the forms discussed above; significant differences in the propensity to be in housing affordability by the due date, or (ii) car loans or personal loans, or (iii) interest free purchase, or (iv) hire
stress are determined by using Wald tests of proportions. purchase agreements, or (v) ‘other’ consumer debt. A respondent is coded as not having any
consumer debt if they do not report holding debt in any of the forms discussed above.

results are controlled for the important effect of housing tenure Confirming results from Table 1, with controls for economic
upon affordability problems. I also refer to the rare event odds factors, couple households are about 56% less likely than
ratio (OR = expβ a) and the marginal percentage change women living alone to experience housing affordability stress
calculated from the OR. (OR = 0.443). Earlier studies have shown how women living

22 Australasian Journal on Ageing, Vol 27 No 1 March 2008, 20–25

© 2008 The Author
Journal compilation © 2008 ACOTA
aja_268.fm Page 23 Monday, February 4, 2008 10:31 AM

A g e i n g a n d h o u s i n g a f f o r d a b i l i t y

alone are at great risk of housing affordability problems when were purchasing a home. In comparison, 19% of those aged 54
compared to men living alone [24]. However, the coefficient and younger owned their own home and a further 43% were
measuring membership of a lone man household is not currently purchasing one. As shown by the results above,
significant (P > 0.10), indicating that there is no difference in owning a home is an important factor protecting households
the prevalence of housing affordability stress between lone from housing affordability stress. A further explanation is
men and lone women. Similarly, there is no difference in the offered by the range of concessions available to older Australians.
risk of housing affordability problems between lone women For example, the pensioner’s concession card provides
and ‘other’ household composition types. concessions on council rates, water and sewerage, electricity,
registration and public transport. Persons with a pensioner’s
Contrary to popular opinion, age is found to be strongly and concession card or commonwealth senior’s health-care card
negatively associated with the prevalence of housing afford- are also eligible for heavily subsidised pharmaceuticals
ability stress. Compared with the young-old (55– 64), those in through the Pharmaceutical Benefits Scheme (PBS). These con-
age group 65 – 74 are almost 49% less likely to experience cessions no doubt partially explain the lower prevalence of
affordability stress (OR = 0.067). Similarly, compared with the housing affordability among older Australians, by subsidising
young old, the oldest-old (75+) are about 65% less likely to household expenses.
experience housing affordability stress (OR = 0.348). Wald
tests of the hypothesis yield insignificant results (F = 1.98, Nonetheless, although only 5% of those aged 55 and older
P > 0.10), providing some support that the prevalence of reported difficulty in meeting housing and utility bills on time,
affordability stress flattens after age 64. Of course, longitudinal there is a definite social gradient explaining those at risk. In
data are needed to confirm the relationship between the risk of particular, living arrangements, age, tenure, income and asset
affordability stress and increasing age. and debt status were strong predictors of the prevalence of
housing affordability stress.
Apart from these demographic factors, housing and economic
factors are strongly associated with the risk of housing afford- Results here show how persons living alone, either man or
ability stress. Estimates show that when compared to home woman, are at far greater risk of experiencing housing
owners, purchasers are at a 3.5-fold risk of housing stress and affordability stress when compared to persons in couple
public and private renters are at a 2.5- and 2.8-fold risk, households. Rossiter has pointed to the ‘cushioning effect’ of
respectively (OR = 3.53, OR = 2.55, OR = 2.84). These find- belonging to a couple household in reducing the burden
ings are important, as it shows that once controls are included of housing costs because of the use of two potential sources of
for income, assets and debts, tenure continues to strongly income, combined with an economies of scale effect [25],
explain the incidence of housing affordability stress. which is strongly supported by the multivariate findings. For
example, compared to a woman living alone, a couple household
Not surprisingly, high income is strongly and negatively was 56% less likely to experience housing affordability stress.
associated with housing affordability stress. Older persons in
the top 20% of income earners were about 93% less likely to Apart from these important economic considerations, having
experience difficulty paying for their housing and utility bill on a spouse (or partner) also reduces the burden of housing
time (OR = 0.067). Reflecting this economic effect, older expenditure stress in other ways. For example, Kendig (1981)
Australians holding assets were four times more likely to argues that having a spouse can ‘. . . compensate for limited
experience housing stress than those with such investments personal capabilities, which enables them to share household
(OR = 4.08). In contrast, not holding any form of consumer responsibilities and help each other to overcome any personal
debt was associated with a 64% decline in the risk of afford- limitations’ (p. 87) [26].
ability stress (OR = 0.362).
It was interesting that there was no significant difference in the
With controls for other factors, neither education, nor having prevalence of housing affordability stress between lone men
a disability or health condition was associated with the risk of and lone women. Even without controls in the multivariate
housing affordability problems. Interaction effects between model, the unadjusted univariate findings show remarkable
tenure and age and tenure and living arrangements were found similarity in the housing expenditure of lone men and lone
to be statistically insignificant (P > 0.05). women. These findings for living arrangements differ from
earlier Australian findings. In Rossiter’s (1986) study of the
Discussion housing conditions of older women, it is argued ‘In terms of the
Results from this study indicate that relative to the younger distribution of housing problems, however, there are relatively
population, the risk of difficulty in paying for housing and few elderly men in these circumstances and the majority of men
utility bills is significantly lower among those aged 55 years are to be found in the comparative security (from a housing
and over. A likely explanation for this finding is the different viewpoint) of a couple’ (p. 6) [25]. This author also concluded
rates of home ownership among these two groups. Estimates that ‘It is possible to conclude that the relatively disadvantaged
from the 2002 GSS indicate that about 71% of respondents position of single elderly women in terms of housing costs
aged 55 and older owned their own home and a further 10% largely derives from their low incomes’ (p. 10) [25].

Australasian Journal on Ageing, Vol 27 No 1 March 2008, 20–25 23

© 2008 The Author
Journal compilation © 2008 ACOTA
aja_268.fm Page 24 Monday, February 4, 2008 10:31 AM

Te m p l e J B

There are several important reasons for the difference in Conclusion

findings by sex between this earlier study and the current study. A popular image about housing and older people is that they
Firstly, Rossiter’s (1986) and earlier studies were based on data are ‘free and clear’ of burdensome expenditures given that a
from the 1980s’ cross-sections of the income and housing costs large proportion of this group are home owners [30]. This view
survey. Using more recent data from the 1998/1999 House- ignores the heterogeneity in the economic and housing circum-
hold Expenditure Survey, it has been suggested that differences stances of older people and, indeed, the results from this study
in housing costs between men and women may have converged reject this outdated stereotype.
because of other factors, shifts in economic well-being and key
policy changes [27]. Of particular importance is the assistance Persons who live alone, do not own their own home, have low
with housing repairs offered through the Home and Community incomes and hold debt with few assets are at a greater risk of
Care (HACC) program. Since the early 1980s, substantial housing affordability stress, even when other economic,
funding has flowed into the HACC program, leading to a housing and demographic factors are controlled for. Policy
subsidisation of repair costs for many older people and initiatives that seek to improve the housing situation of these
subsequently lowering shelter expenses. Furthermore, more groups will increase individual well-being, and enable these
funding has also flowed into the Commonwealth Rental older people to remain living independently in the community.
Assistance program.
One limitation of this study is the focus on only one aspect of
In addition to the effect of living arrangements, this study finds housing affordability. The ability to meet housing and utility
that older age groups are at a lower risk of housing affordability expenses on time undoubtedly reflects just one dimension of
stress when compared to younger groups within the older housing affordability stress. Many older Australians, who meet
population itself. There are several potential explanations for these expenses on time, may nonetheless experience problems
this finding. First, there is the strong possibility of a selectivity with housing affordability. For example, they may reduce
effect in survival occurring in the data. Disney (1996), analysing expenditure on health care or food to ensure that their housing
saving and consumption behaviour, summarises ‘Disaggregated and utility bills are paid in a timely manner. Some older Australians
cross-section studies of saving behaviour among different age may also have their public or private rental income deducted
groups, however, contain two forms of bias, which work in automatically from Centrelink receipts thereby masking afford-
opposite directions. One is a form of attrition bias: wealthier ability and financial problems. Moreover, along with financial
individuals tend to survive longest, and it is their higher wealth barriers, housing stress may be induced by concerns with housing
that is observed in the data’ [28]. That is, given the higher quality or neighbourhood safety. The housing stress of older
wealth of those surviving in old age, the housing expenditure Australians may be higher according to these alternate definitions.
profile may flatten artificially because of death and institution- Further research may wish to examine these issues further.
alisation of the poorest residents. The finding that the risk of
housing affordability stress declines with age is partially the
result of this selectivity effect. The second important consideration Acknowledgement
The 2003 ABS General Social Survey was made available
is the cohort effect operating through cross-sectional data.
through the Australian Bureau of Statistics (ABS) and Australian
Technically, what is being measured is the difference in housing Vice Chancellors Committee (AVCC) agreement. The opinions
expenditure between the three age groups at one point in time. expressed herein do not reflect those of the ABS or AVCC. Com-
Longitudinal data are required to correctly measure age grading ments from the two reviewers helped improve the manuscript.
with respect to housing expenditure.

Apart from these demographic factors, owning a home in later

life was found to protect the household against affordability stress, Key Points
when compared to renting (public or private) or other tenure types. • Approximately 5% of persons aged 55 and older
This provides strong support for Kendig (1984) who argues had difficulty meeting housing and utility expenses
that the low housing costs afforded by home ownership on time in 2002. The corresponding figure for those
‘emerge as important as income’ in retirement (p. 249) [29]. aged younger than 55 was about 20%.
• Older people living alone, regardless of sex, are at
Finally, apart from these key demographic and housing
greater risk of housing affordability stress when
influences, economic factors, in particular, holding assets, high
compared to couple households.
income and not having consumer debt are important predictors
of being able to afford housing and utility costs when they • Those with low levels of economic resources (low
come because of old age. This further highlights the significance income, low levels of assets, having a consumer debt)
of financial planning early in the life course to help fund are at a greater risk of housing affordability stress.
retirement. However, many of the low income owners without • Home ownership, even with controls for economic
a home or any assets during retirement would most likely have resources, acts as an important buffer against
been on low income earlier in the life course, posing puzzling housing affordability stress in old age.
questions for policy makers.

24 Australasian Journal on Ageing, Vol 27 No 1 March 2008, 20–25

© 2008 The Author
Journal compilation © 2008 ACOTA
aja_268.fm Page 25 Monday, February 4, 2008 10:31 AM

A g e i n g a n d h o u s i n g a f f o r d a b i l i t y

References 14 Schwenk F. Housing expenditures. Family Economics Review 1988; 1: 1–

1 Kendig H, McCallum J. Greying Australia: Future Impacts of Population 7.
Ageing. Canberra, ACT: Australian Government Publishing Service, 15 Thalmann P. Identifying households which need housing assistance.
1986. Urban Studies 1999; 36: 1933 –1947.
2 Bishop B. ‘Healthy ageing discussion paper’. The National Strategy for an 16 Linneman P, Megbolugbe I. Housing affordability. Urban Studies 1992;
Ageing Australia. Canberra, ACT: Commonwealth of Australia, 1999. 29: 369–392.
3 Nettleton S, Burrows R. Mortgage debt, insecure home ownership and 17 Moore E, Skaburskis A. Canada’s increasing housing affordability
health: An exploratory analysis. Sociology of Health and Illness 1998; 20: burdens. Housing Studies 2004; 19: 395 – 413.
731– 753. 18 Harding A, Phillips B, Kelly S. Trends in housing stress. Paper presented
4 Dunn J, Hayes M. Social inequality, population health and housing: A at the National Summit on Housing Affordability. Canberra, 2004.
study of two Vancouver neighbourhoods. Social Science and Medicine 19 National Housing Strategy. The Affordability of Australian Housing .
2000; 51: 563 – 587. Canberra, ACT: Australian Government Publishing Service, 1991.
5 Nettleton S, Burrows R. When a capital investment becomes an emotional 20 Howe A. Housing for Older Australians: Affordability, Adjustments and
loss: The health experience of mortgage possession in England. Housing Care. Canberra, ACT: Commonwealth Office for the Aged, 1992.
Studies 2000; 15: 463 – 479. 21 Australian Government Comments in SCRGSP. Report on Government
6 Smith C, Smith C, Kearns R, Abbott M. Housing stressors, social support Services 2004. Canberra, ACT: Steering Committee for the Review of
and psychological stress. Social Science and Medicine 1993; 37: 603 – Government Service Provision, Productivity Commission, 2004.
612. 22 Hancock K. Can pay/won’t pay? Or economic principles of affordability.
7 National Consumer Law Center, in Golant S, La Greca A. The relative Urban Studies 1993; 30: 127 –145.
deprivation of U.S. elderly households as judged by their housing 23 King G, Zeng G. Logistic regression in rare events data. Political Analysis
problems. Journal of Gerontology B: Psychological Sciences and Social 2001; 9: 137 –163.
Sciences 1995; 50: S13 – S23. 24 Tomz M, King G, Zeng G. Relogit: Rare events logistic regression. Journal
8 Iams D, Royce V. Rising energy costs and their impact on household of Statistical Software 2003; 8.
budgets of older Americans. Journal of Housing for the Elderly 1984; 2: 25 Rossiter C. Housing tenure and costs of older Australians: Gender issues.
15 – 23. Australian Journal on Ageing 1986; 5: 4 –12.
9 Reddick J. The interdependence of health and housing for the elderly. 26 Kendig H. Housing and living arrangements of the aged. In: Howe E, ed.
Journal of Housing for the Elderly 1985; 2: 77 – 82. Towards an Older Australia. St Lucia, Queensland: University of Queensland
10 Golant S, La Greca A. The relative deprivation of U.S. elderly households Press 1981, 85 –101.
as judged by their housing problems. Journal of Gerontology B 1995; 50: 27 Temple J. Older Australians as consumers: The role of selected demo-
S13 – S23. graphic variables in explaining expenditure on necessities (PhD Thesis).
11 Guadagno M. Families with high housing expenditures: A comparison of Canberra, ACT: Graduate Program in Demography, The Australian
homeowners and renters. Family Economics and Nutrition Review 1992; National University, 2006.
5: 2 –11. 28 Disney R. Can We Afford to Grow Older? A Perspective on the Economics
12 Struyk R. The housing expense burden of households headed by the of Aging. London: MIT Press, 1996.
elderly. Gerontologist 1977; 17: 447 – 457. 29 Kendig H. Housing tenure and generational equity. Ageing and Society
13 Disney R, Henley A, Stears G. Housing costs, house price shocks and 1984; 4: 249 – 272.
savings behaviour among older households in Britain. Regional Science 30 Cook F, Lomax F, Setterstein R. Expenditure patterns by age and income
and Urban Economics 2002; 32: 607 – 625. among mature adults: Does age matter? Gerontologist 1995; 35: 10 –23.

Australasian Journal on Ageing, Vol 27 No 1 March 2008, 20–25 25

© 2008 The Author
Journal compilation © 2008 ACOTA