Kōtuitui
New Zealand Journal of Social Sciences Online
The housing fulcrum: balancing economic and social factors in
housing research and policy
Arthur Grimes
Suzi Kerr
Andrew Aitken
Robert Sourell*
Motu Economic & Public Policy Research
PO Box 24390
Wellington, New Zealand
*Present address: Institut für Sozialwissenschaften,
Bienroder Weg 97, 38106 Braunschweig, Germany.
Abstract We discuss the pivotal role that housing
plays for both social and economic outcomes. All people need to be
housed, and housing circumstances help determine social outcomes,
especially for vulnerable groups. We outline an analytical framework
that treats housing as a dynamic system, incorporating the life-cycles
of both individuals/households and houses. Each is long-lived;
decisions impact on housing for decades. This approach is relevant to
the development of housing research and housing policy. We illustrate
the issues with reference to the relationship between rental yields and
measures of deprivation across New Zealand. The surprising nature of
this relationship has consequences regarding potential poverty traps
and wealth disparities. While highlighting an important housing market
issue, any policy response to this issue is complicated by the need to
take account of the life-cycles of both individuals and of houses, and
by the long-lasting impacts of decisions.
Keywords Housing policy; housing research; life-cycle; rental
yields; deprivation
INTRODUCTION
Housing is the fulcrum around which important economic and social
factors are balanced. All people need to be housed. The nature of that
housing, and of the neighbourhood in which they are housed, affects
their well-being and those of others around them. Outcomes of certain
population segments, particularly children and other vulnerable groups,
are linked to their housing circumstances. We examine the centrality of
housing for both economic and social policy.
Economic forces have a major bearing on housing outcomes. Grimes
& Aitken (2004) found that regional developments in production,
employment, commodity prices of locally produced agricultural goods,
and population relative to the existing housing stock all affect
regional house prices. House rents, in turn, reflect house prices
(Grimes et al. 2005).
Housing costs (especially rents and mortgage repayments) form a
substantial portion of household expenditures. In 2001, 33% of New
Zealand households spent at least one-quarter of their income on
housing; 15% of households spent more than two-fifths of their income
on housing. Housing costs, therefore, have a significant impact on
living standards for many people. Given that shelter is a necessity,
high housing costs are particularly germane for lower income households
since they cannot choose simply to avoid purchase of housing services
(as they may with other high cost, but less necessary, services).
Housing market developments, in turn, have strong macroeconomic
effects, influencing inflation outcomes and the setting of interest
rates (Reserve Bank of New Zealand 2006). Housing issues also impinge
on environmental outcomes through urban expansion, in-fill, energy
usage, and transport requirements.
The materiality of these interactions underlies the importance of
conducting research that analyses the determinants, and the effects, of
housing developments in New Zealand. The breadth of the interactions
indicates the usefulness of adopting multi-disciplinary approaches when
researching the interaction of housing with economic, societal, and
environmental outcomes1. In this paper, we
take as given the importance of conducting housing related research
within New Zealand. Our focus is on discussing a framework for
analysing housing issues as they relate to social and economic
outcomes. The framework highlights the life-cycle nature of
individuals’ and households’ housing choices and also the life-cycle
nature of the housing stock itself. The complex nature of the issues is
illustrated by reference to a surprising feature of New Zealand’s
rental market relating to the relationship between deprivation measures
and the rental yield. This relationship suggests a number of hypotheses
to be explored both as to its social implications and to the economic
(investment) factors that underlie it.
Housing policy needs to incorporate these multi-disciplinary
complexities in order to provide coherent responses to housing-related
issues. As our framework makes explicit, these factors must be
considered in a dynamic, rather than static, context since housing (and
housing policy) decisions have ramifications that last for decades.
DYNAMIC LIFE-CYCLE FRAMEWORKS2
Housing choices are dynamic and have long-lived implications. In
order to understand many observed features of the housing sector, it is
important to have a framework for thinking about how individuals and
households may behave as they progress through the housing market over
their lifetime. This understanding is important for policy formulation
since it highlights the way that people might wish to behave, either
through conscious decision-making or through following social norms.
Policy can then work in tandem with people’s aspirations rather than at
cross-purposes to them.
The life-cycle pertains both to the life-cycles of individuals and
households, and to the life-cycle of the housing stock itself. Both are
long-lived and, for housing policy purposes, it is important to
consider both of these life-cycle aspects. An emphasis on the personal
and house life-cycles translates into research and policy that focuses
on transitions and on the long-lasting dynamic implications of
decisions taken with respect to housing and the factors affecting
housing. Naturally, there is a diversity in circumstances of
individuals and households, just as there is considerable diversity of
houses. Despite this heterogeneity of experiences (across people and
across time) each person undergoes a life-cycle and their housing needs
change over the course of that life-cycle. Understanding the diverse
circumstances of people across the life-cycle (especially those that
differ markedly from what might traditionally be regarded as more
“typical” situations) is important for highlighting and interpreting
particular housing issues that may warrant policy attention. Thus, the
life-cycle approach is as important for identifying atypical
developments as it is for tracing out the typical dynamics and
transitions over individuals’ lifetimes.
INDIVIDUAL AND HOUSEHOLD LIFE-CYCLE
Rossi (1955) pioneered the concept of housing transitions motivated
by life-cycle stages in his study Why Families Move. He
analysed a “typical” hierarchical decision tree relating to migration
and housing decision-making that is reflected in our discussion below.
We also build on approaches of those who have adapted the traditional
life-cycle model to suit more recent and diverse social trends (see
Dieleman 2001) and apply the insights to New Zealand examples. With
respect to the latter, the recent survey-based study by DTZ New Zealand
(2005) is useful in establishing “typical” New Zealand attitudes to
home ownership and other housing-related decisions. Their evidence
indicates that the traditional “housing career” path remains the
dominant desired pattern in New Zealand.
An individual starts life within some family environment which
determines their initial housing status. Most children make no explicit
housing choice; the housing situation of their parent(s) or
caregiver(s) determines their dwelling status. The early housing
situation may, however, be a crucial determinant of current and future
outcomes for the individual child. For instance, a considerable body of
New Zealand and international research across different disciplines
shows that educational, health, and anti-social outcomes may each be
affected by the child’s personal housing circumstances (Green &
White 1997; Crampton et al. 2000; Glaeser & Sacerdote 2000;
McNicholas et al. 2000; Galster et al. 2002; Haurin et al. 2002;
Howden-Chapman & Carroll 2004). These outcomes may also be affected
by community characteristics in which that housing is located, with
externalities stemming from the quality of other houses in the
neighbourhood (Furstenberg & Hughes 1995; Levanthal &
Brooks-Gunn 2000; Durlauf 2004).
Once children are ready to leave home, they face a number of
choices. They may decide to stay in the same town and pursue work or
educational (or benefit) opportunities in that town. They may decide to
leave for another location to pursue those opportunities. The choice of
regional location is a crucial decision at this stage of life. It will
be affected by many factors including: education opportunities in
different locations (for instance, only six New Zealand cities have a
university); work opportunities (employment growth has been much
stronger in Auckland than in the southern North Island or much of the
South Island); natural and social amenities; costs, availability, and
quality of housing; transport costs and transport links between housing
and work/education opportunities; and transport links with family and
friends, both in prospective locations and in the original home
location.
Normally, a young adult leaving home has relatively low income and
little or no capital, so begins their independent housing in a rented
property3. If there is a shortage of
suitable rental properties, the newcomer places pressure on the rental
market. This pressure may result in one or more of: increasing rents
and/or house prices, overcrowding, the purchase of an existing
owner-occupied house by a (private or state sector) landlord in order
to increase the rental stock, or building of additional housing stock
for a landlord to rent out. Alternatively, the “newcomer” may postpone
the proposed location shift.
Each of these responses has different effects on housing outcomes
over different timeframes. Overcrowding and/or postponing a shift has
immediate personal and social impacts on the people concerned. Purchase
of an owner-occupied house for rental takes some time and does not
directly increase the housing stock; thus, any housing shortage is
reflected elsewhere (e.g., in a shortage of owner-occupied dwellings).
Building an additional rental dwelling increases the stock of rentals
without diminishing the housing stock elsewhere. It has other effects,
however, including utilising land—so contributing (in urban situations)
either to in-fill or to sprawl—and increasing demand on local
infrastructure and local amenities. Thus, environmental, as well as
social and economic, impacts arise.
The circumstances of migrants to New Zealand vary considerably with
respect to housing choices. Some migrants will be on moderately high
incomes and/or may already have capital from the sale of a house in
another country. They may enter the housing market directly via the
owner-occupied category through purchase of an existing house or
building a new dwelling. Other migrants, particularly migrants with
little financial capital, can be considered similarly to young adults
starting out on their housing “careers”. Frequently, these migrants
will enter rental housing in their initial foray into the New Zealand
market. However, migrants may already be in a family grouping and so
the demand for initial house type may differ from that of a young
adult. There may be greater demand for housing located near work
opportunities rather than educational opportunities (except for
migrants with school-aged families, where school choice could be a
crucial factor in the presence of zoning restrictions). Alatini (2004)
provides an in-depth account of residential circumstances of the Tongan
community in Otara, with many living in substandard or crowded rental
accommodation or in accommodation owned by relatives (see also Cheer et
al. 2002). Refugees are also likely to have low income and lack of
capital. They may be in a family situation, so needing a large house,
despite being hampered by lack of capital and possibly by poor
income-earning opportunities. They may also face different types of
discrimination than faced by young adults born in New Zealand.
As in the labour market, young adults tend to shift house
relatively frequently as perceived opportunities and constraints change
(Dieleman 2001). They are more mobile than other population segments
and can experiment in a number of dimensions (living arrangements,
education, work, international migration, etc) resulting in greater
mobility4. During this period, they are
likely to choose to remain in rented accommodation at the lower end of
the housing market.
Two key life-cycle events may then impact on a person’s choice of
housing type. The first is employment in a reasonably secure job, or at
least employment in a location offering reasonably secure ongoing
employment opportunities in an occupation or industry seen as a
long-term prospect by the individual (Rossi 1955). For instance, a
young adult trained in marketing may shift to Auckland, where he/she
sees ongoing employment opportunities, even if not with the same firm.
Corresponding with this development is often a rise in current or
prospective income for the individual. In New Zealand, Maré
& Timmins (2003) find that job prospects (employment growth) is a
key variable influencing migrants’ choice of destination once they have
decided to emigrate from their home region.
The second life-cycle event is the choice to have a family and to
place roots in a location where they wish that family to become
established. The two events may be related.
As in the choice of educational opportunity, a regional component
is highly significant in determining the housing choice associated with
either or both of these events. The decision to shift to a particular
region is the highest tier in a hierarchy of decisions. Once that
decision is taken, the housing choice within that region follows
without overriding the original regional choice, except in extremis
(if housing is simply not available at any realistic price in the
desired location). This location decision will often involve shifting
to (or within) a city that is growing relatively strongly, so offering
a greater range of opportunities than other locations (Maré
& Timmins 2003). In some cases, the choice will be to move
(permanently or semi-permanently) offshore, especially if prospective
overseas locations are growing strongly relative to domestic
alternatives5. Where domestic migration is
involved, the supply of houses has to increase in the prospective
locations. Thus, pressures for greater in-fill and/or urban expansion,
and for more infrastructure and amenities, result in these locations.
Housing-related decisions taken around these life-cycle events will
often involve a conscious choice to change the nature of housing tenure
from rental to owner-occupation. This is particularly likely to be the
case with the onset of children, since owning a home provides a more
secure environment for raising a family and provides a “natural hedge”,
ensuring that desired housing characteristics for the family will be
maintained over time. However, the choice can only be activated if
affordability and finance constraints allow. The life-cycle approach is
useful for identifying atypical situations where a household with
secure work opportunities and a young family does not follow the
transition to owner-occupation. The causes of this atypical behaviour
can then be teased out. For instance, DTZ New Zealand (2005) found that
households perceive the inability to save a deposit out of current
income is a significant barrier to moving into home ownership. This
barrier is particularly germane in light of the reported preference of
younger households to move into modern and maintenance-free first homes
rather than into cheaper properties requiring renovation or in less
preferred areas.
Any change from one owner-occupied home to another involves
significant transactions costs. Thus, the choice to enter
owner-occupation is often made only once an individual or family
envisages settling in a particular community for a substantial period.
Factors that change the likelihood of moving from one location to
another in the near term are likely to discourage people from entering
the owner-occupied market. Over recent decades, young adults have
entered higher education in considerably greater numbers than
previously, when the norm was to enter the workforce straight from
school6. This move to greater tertiary
education has enlarged many people’s employment options across
different locations and so is a factor in delaying the decision to
enter owner-occupied housing. Another factor is the greater delay in
entering marriage and/or having children. Prolonged education and
greater overseas travel while young also means that people choose to
build less capital early in life than was the prior norm. This choice
increases the constraint they face when young in affording the deposit
for owner-occupied accommodation.
While the dynamics may have changed, the desire of people to become
owner-occupiers as they settle down in employment and family status
currently remains strong. DTZ New Zealand (2005: 5) provide
survey-based evidence that this is the case, stating: “The overwhelming
majority of focus group participants whether currently in either public
or private rental, families, singles or couples, located in Auckland or
Hawkes Bay, envisaged and aspired to being owner-occupiers in ten years
time.”
The mortgage market enables people to borrow against expected
future income and so purchase a house, to be repaid over their working
life. Table mortgage repayments will often be high as a proportion of
income in early years, with a lower ratio in later years as the
(nominal) household income rises. Banks normally place a ceiling on the
proportion of income which an individual or family can devote to debt
servicing. Households may therefore face a cashflow constraint which
prevents them acquiring their desired house, especially when they are
still on relatively low incomes. These constraints will be particularly
relevant in times of high interest rates (e.g., mid-1980s) or at times
when house prices have risen strongly relative to current incomes
(e.g., at present). In such circumstances, individuals may be forced to
accept a lower quality dwelling than desired and later “trade up” as
their incomes rise.
Typically, as a household’s income rises, they move from rental
housing to a “starter” owner-occupied dwelling, with the nature of that
dwelling determined by current needs (e.g., couple without children)
and by affordability, especially deposit availability (DTZ New Zealand
2005). Such dwellings may include an apartment, terrace, or
semi-detached house. As family size and household income rises, they
may then transition to larger family homes through their prime aged
work years. A household’s preferred house location within a region may
well change as their circumstances and requirements for various
amenities change. This places pressures on house prices in certain
areas, tending to segment the market between those able and willing to
pay a premium (e.g., for location within a desirable school zone) and
those unable and/or unwilling to pay the premium7.
Educational, health, and/or crime outcomes may become linked to certain
locations determined by factors such as school zones or proximity to
amenities and infrastructure.
People also move down the housing hierarchy due to personal
circumstances. In terms of life-cycle dynamics, this is most likely to
happen in three sets of circumstances.
The first set of circumstances is breakup of the family. In such
circumstances, each partner’s income is less able to support the
previous housing quality, and a shift downwards within the housing
market may result. For those without capital, the result may also be a
shift back to rental accommodation from owner-occupation; unclear
future personal options increases the likelihood of a shift to rental
housing.
The second set of circumstances is the onset of retirement. Once
children have left home and income from employment falls or disappears,
couples/individuals can reassess their housing needs, including choice
of region in which they wish to retire. The choice made in early
adulthood to locate near desired work options is now less relevant, or
even obsolete. Other factors—such as a warm climate, and proximity to
coast, family or iwi links—may predominate. Relocation based on climate
or coastal location may lead to growing numbers of older people
shifting to such locations. Some of these locations may be metropolitan
centres; some will be in non-metropolitan areas, possibly arising from
family or iwi links. In each case, there will be increased demand for
housing, infrastructure, and amenities. If the decision is to retire to
a non-metropolitan location that has experienced a population decline,
a major issue may be to find housing of suitable quality amongst the
vacant houses in the area. Vacancies may have arisen as a result of
young adults shifting to locations with greater educational and work
opportunities. This process may have freed up houses that are not
re-inhabited quickly and are left to deteriorate as maintenance is
foregone8.
A third set of circumstances, relevant especially to
non-metropolitan areas, may cause housing upheaval. These circumstances
arise where a major employer in a relatively remote location closes
down or cuts back its workforce. Employment options in the immediate
area are limited, and people who had earlier made the decision to
locate near the large employer now have to assess their next best
option. It may be to stay in the location and look for alternative
(probably lower paid) work or move onto a benefit, or to shift location
to an area where work or retraining opportunities are more prevalent.
The pressures of the latter course of action are similar to those
associated with young adults, except in one respect. It is quite
possible that the household which intends to shift in these
circumstances includes children. Their circumstances may resemble those
of low income immigrants, including refugees, settling into a new city.
The difficulty of resettling into a location with greater work
opportunities means that the costs which have to be met by the
relocating household can be very high. Also, the lifetime over which
such costs can be amortised will be shorter. As a result, the household
may have to relocate to the least desirable parts of a larger centre,
with potential negative effects for family, and especially for
children’s, outcomes.
The framework above has largely concentrated on “typical”
movements. Identification of these typical patterns importantly helps
to identify atypical patterns for certain people. A number of
groups of people may require alternative housing paths.
One such group are people who have to move into sheltered care once
they can no longer manage independent living owing to age-related
conditions. The housing needs of this group will become more important
in aggregate terms as they constitute a greater proportion of the
population over the next three decades. A second group comprises mental
health clients and people with mental disabilities. In past decades,
many such clients were housed in various forms of institution. The
closure of many institutions and a greater preference for independent
living has raised the need for suitable housing for these people
throughout their lifetime. Regulatory issues (e.g., requirements placed
on boarding houses), availability of suitable accommodation types, and
housing support for people unable to earn sufficient income to afford
normally priced housing, will help determine the quality and degree of
imbalance in this market segment.
HOUSING LIFE-CYCLE
Houses are long-lived assets, although different types of houses
have different life expectancies9. Because
of their longevity, houses also undergo a life-cycle that is relevant
to understanding housing issues. As with people, houses are diverse,
but there are key life-cycle characteristics that are relevant. In
particular, significant persistence is built into the supply of housing
since (other than mobile homes) each house—once built—remains specific
to a particular location. Further, each house will depreciate and
requires maintenance in order to sustain it as a habitable dwelling.
A new house may be purpose-built for a prospective owner, meeting
their needs at that time, or it may result from a more speculative
investment by a construction firm aimed at prospective purchasers.
Typically, the signal for builders to construct new houses is an
increase (or the prospect of an increase) in house prices within the
location relative to land and construction costs.
In order to be able to build new houses, the supply of people with
skills relevant to the house construction industry must be sufficient
within or near the relevant location. Where these skills are in short
supply relative to demand, the signal to attract extra construction
workers will be through higher wages/salaries for those involved in the
construction sector10. If construction
resources are in short supply nationally, the result may be a shortage
of construction workers elsewhere as the resources shift to locations
with the greatest imbalances.
Another key requirement for housing supply to keep up with demand
in growing areas is the availability of land with adequate
infrastructure that is suitable for supporting residential expansion.
This requires a supportive regulatory framework enabling either in-fill
housing construction or new subdivisions. If land availability is
squeezed because of topographical features or regulatory constraints or
lack of infrastructure, the result will be a bidding up of land values,
feeding through to house prices (and rents) in the location.
Once the new house is built and the first owner takes possession,
the house and its owners will age. The house may become less suitable
for that owner, and the owner will sell and move on. At the time of
sale, the house is no longer the same commodity that it was when it was
built. Its materials will have weathered, its style may no longer be
fashionable, and the desirability of its location may have changed as
its neighbourhood and other neighbourhoods have matured.
The new owner will therefore not necessarily be similar to the
original owner in terms of income or social status. In some cases, the
desirability of the neighbourhood may have increased if the area
becomes more fashionable. A new owner may decide to upgrade the quality
of the house, through renovation and/or extension. The status and
income of second or subsequent owners may then be higher than the
original owner. This represents a process of “gentrification” or
“filtering-up” (see Collins et al. 2002, and references therein, for a
discussion of the filtering processes in the United States).
Typically in many cities, houses tend to “filter down” towards the
lower parts of the housing hierarchy as they age (Grigsby 1963). In
this case, we might find that the initial owner-occupier is replaced by
a “lower status” and lower income owner-occupier. The house may then
change hands a number of times until owner-occupiers find the house
less desirable than does a landlord, who purchases the house as a
rental property. The life-cycle may then be such that the house follows
an ageing trajectory whereby the perceived quality of tenants (in the
eyes of the landlord; see Smith & Fraser 2004) and level of rents
declines over time, reflecting the state of the house and of the
neighbourhood.
Finally, the landlord may find that the value achieved by renting
out the property (to low income tenants) is less than the value
obtainable by demolishing the house and rebuilding on the same site. In
making this decision, the landlord has to factor in the cost of
maintenance, which may itself be related to the quality of tenants. The
landlord also has to factor in the dependability of rental payments
(i.e., probability of default), which may be related to tenant quality.
In locations with declining populations (especially some rural
areas), some houses will be left uninhabited and may become run-down or
derelict. Maintenance, even of inhabited houses in the location, may be
postponed owing to a lack of income, or through an explicit decision to
reduce the capital tied up in a house in an undesirable location. In
this case, the ageing process may be brought forward relative to the
process in a strongly growing community.
This housing life-cycle forms a crucial backdrop for understanding
the mix of housing quality and availability in particular locations
(noting that housing is virtually always location-specific; see Sweeney
1974 and O’Flaherty 1995, 1996). As a hypothetical example, if it is
the norm that houses generally become appropriate for low income
tenants say 40 years after construction, then the events of 40 years
ago have a major bearing on the availability of low income rental
housing today. In a rapidly growing location, there will be
comparatively few older houses available and hence a greater likelihood
of a housing shortage for low income households. This would be the case
even if new construction in the location were keeping up with the
housing demand of newly migrating higher income households. Further,
higher income households may pay a premium for houses in a desired
location, including lower quality houses, so shifting the income/status
of residents in a neighbourhood upwards. This further exacerbates the
housing shortage for those on low incomes in these locations. Given the
length of the housing life-cycle, it may take decades for the mix of
the housing stock to become appropriate for the needs of all
demographic groups in such a location. In the interim, a sustained
period of homelessness, overcrowding, or long-range commuting may be
the outcome.
In addressing policy issues, this housing life-cycle needs to be
borne in mind. Neglect of its importance could lead to superficial
responses that ignore the effects of long-lived dynamics in the housing
market11. The housing life-cycle, and its
interaction with the individual/household life-cycle, also provides a
rich basis for designing an agenda of housing research (see Grimes et
al. 2004). The framework highlights the interdependence of individuals’
and households’ dynamic housing decisions with their decisions over
locational choice, and with the whole history of the housing stock in
each location.
It also highlights the importance of the housing supply response to
changes in demand. Factors affecting housing supply, such as
infrastructure provision, regulations, and availability of construction
labour, have impacts on the speed of the supply response. Hence, they
affect the price and availability of houses facing people in different
locations at various stages of their life-cycle. While supply is
important, no clear distinction has been made according to the source
of the supply response. Thus, we have not singled out housing response
from private suppliers, the state, or third sector organisations. Each
of these responses may be important, but the division between them is a
subsidiary issue for the housing sector compared with the overall
determinants of housing demand and housing supply12.
Another key insight from the life-cycle framework is that there is
a process of switching houses and accommodation types throughout one’s
lifetime. Consequently, as in the labour market, individuals (whether
buyers, sellers, landlords, or tenants) will at times have to undergo a
process of search to find suitable accommodation and/or suitable
tenants or purchasers (Wheaton 1990). In any market involving search,
there is a flow of new commodities (houses) coming onto the market and
a flow of new people searching for them. A snapshot taken at any
specific time will reveal vacant properties and unhoused
individuals/families, sometimes in the same location. In the housing
context, there will be a flow of new construction in train to meet new
demand. The existence of vacancies, new construction, and unmet housing
demand at the same time should not be taken as a sign of inefficiency.
It is an inevitable part of a well-functioning dynamic market.
Instead, policy issues are likely to arise where individuals (and
particularly families) have prolonged spells of unmet
accommodation needs, especially at times when others flow in and out of
the market at normal speeds. In these cases, research is required into
the specific problems faced by these individuals, families, or wider
groups of people. The issue may not be a housing problem per se;
the unmet housing need may be a manifestation of a separate problem
that is highlighted by their housing situation.
Policy issues also arise when certain properties (possibly in less
desired non-metropolitan areas) remain vacant (and possibly
deteriorating) for long spells. While this may reflect a rational
individual response by an owner to disinvest from housing in that
location, it may not necessarily be socially optimal in the presence of
unmet demand elsewhere. This is especially the case where that demand
is supported by other policy actions. As an example, the availability
of income support for those living in highly pressured housing markets
may encourage people to locate to such markets and away from declining
regions, with consequent declines in housing quality in the latter
regions.
Policy-makers (and researchers) may be particularly interested in
cases in which life-cycle housing choices result in
externalities—positive or negative. For instance, positive
externalities from housing may arise where housing choices (e.g.,
stable house ownership in an area) foster social stability, social
involvement, and socially desirable behaviours amongst both youth and
adults (Rohe et al. 2002). Conversely, the formation of ghettos, or of
relatively disadvantaged suburbs, may lead to social (educational,
health, crime) problems that necessitate broader policy responses. If
this is the case, policy-makers need to understand the role of
individual decisions and housing policies in contributing to these
externalities in order to devise ways of addressing the underlying
issues.
ILLUSTRATION: THE RENTAL MARKET
Consistent with the life-cycle framework, one marker of deprivation
in New Zealand is “not living in own home” (Crampton et al. 2000).
People in this group tend to comprise either young adults or older
adults who face circumstances preventing them from making the “typical”
transition into owner-occupation. Because of the housing life-cycle,
many renters occupy dwellings that are older and in poorer repair than
the average owner-occupied house.
House prices are a summary statistic for the way purchasers value
the location, amenity, and housing characteristics in each locality
(Meen 2001; Can 1992; Dubin 1992); rents perform a similar role for
renters. Given similar valuations of these characteristics, it might
therefore be reasonable to expect that rentals will be proportionate to
house prices across locations. Alternatively, since renting is
positively correlated with other deprivation markers, the incomes and
other socio-economic characteristics of this group will tend to be
lower than average. On affordability grounds, we might then expect
rents in more deprived areas to be comparatively low relative to house
prices in those areas as compared to the ratio in up-market areas. In
the latter areas, the choice to rent may be related more to lifestyle
and life-cycle factors; for instance, the desire for mobility amongst
individuals and childless couples, even where incomes are high.
Surprisingly, the rent:price ratio (i.e., the rental yield) is
neither proportionate across locations nor higher in more affluent
areas. Instead, rental yields tend to increase with the level of
deprivation across localities. To demonstrate this relationship, we use
tenancy bond (rental) data from the Department of Building and Housing
and house price data from Quotable Value New Zealand, plus data from
the 2001 census. We combine all houses into one group and all
flats/apartments into another. We use area units (AUs, i.e., “suburbs”
or towns) as our measure of locality. After data cleaning, we have 798
cross-sectional observations for houses and 415 observations for
apartments (see Grimes et al. 2005 for details).
We test the cross-sectional correlation (in 2001) between the gross
rental yield and socio-economic variables across AUs. The gross rental
yield is the ratio of rents to house prices in an area. The net rental
yield deducts maintenance, depreciation, default, and other related
ownership costs from the gross rental in calculating the rental yield.
Socio-economic variables influence the desirability of an area and may
influence capital gains expectations (which, theoretically, should
influence the required rental yield; see Grimes et al. 2005 for details
of the theoretical background). Specifically, we examine the
relationship between the rental yield and each of a deprivation index
(see Crampton et al. 2000), median income, the employment rate, and two
separate education variables: the proportion of people with no
education, and the proportion of people with a university degree. Table 1 summarises the correlations of the rental
yield with each of these factors.
In each case, the cross-sectional correlations are significantly
different from zero at the 1% level, and indicate that higher
deprivation is associated with higher rental yields. Figure
1 provides an illustration of the cross-sectional relationship,
graphing the median AU rental yield for houses against the proportion
of the AU population that has no educational qualification.
Relationships in the opposite direction hold between the 1996
values of the same socio-economic variables and rates of capital gain
between 1996 and 2001 (reported in detail in Grimes et al. 2005). All
cases (other than one) are again significant at the 1% level. Thus,
areas with greater deprivation in 1996 tended to have low rates of
capital gain in the succeeding five years. As a finance relationship,
this outcome between capital gains and rental yields is as expected
since the total return to the landlord is the rate of capital gain plus
the net rental yield, which should be equalised across localities.
Nevertheless, the findings raise major issues about: (1) the
implications of the positive relationship between gross rental yields
and deprivation; and (2) the nature of localities experiencing capital
gains and losses.
Taking the first issue, individuals towards the extreme end of the
deprivation scale are not only highly likely to rent, but also appear
to face high rents relative to the standard of housing that they
receive (if house prices are a reasonable measure of housing quality).
This result may possibly be due to three factors in addition to the
finance relationships described above. First, on average, rental
accommodation in low socio-economic areas may be heavily weighted
towards high maintenance properties, possibly as a result of the
filtering-down process described earlier. Second, tenants at this level
of deprivation may be more prone to actions resulting in repairs and
maintenance bills for the landlord. Third, these tenants may have
higher than average default rates. Each of these circumstances leads to
higher rents since the landlord must price the effects into the
required rents. Another factor, related to the housing life-cycle,
could be that the supply of houses suitable for renting to more highly
deprived renters is in short supply, perhaps as a result of a
relatively high number of such renters relative to the supply of old
(unrenovated) houses that might form the bulk of this market segment.
This shortage may force up rental yields for this segment13.
While these reasons may be justifiable in an economic (investment
return) sense, the life-cycle framework indicates that they may result
in negative social outcomes. In particular, people in this
category—whether young adults or more mature adults who are still
renting—may be caught in a poverty trap. With low incomes and high
relative rents, they have little income left for other expenditures,
potentially including investment expenditures on education and
training. Families who have moved from a “closed” mill-town to the city
may find it particularly difficult to meet the family’s necessities in
the new surroundings.
The second issue is relevant to the issue of disparities. Between
1996 and 2001, capital gains (of both houses and apartments) in less
deprived localities tended to be higher than in more deprived
localities. Thus, house owners who were already able to buy property in
highly priced locations increased their wealth to a greater extent than
did those who purchased in lower priced areas. In some cases, capital
gains were negative, especially in more deprived locations. Whatever
the reasons for this outcome (possibly related to changing tastes about
desirable housing locations), the social implications are important.
The increase in wealth disparity will tend to have favoured older,
wealthier households already located in richer suburbs. The hurdle
facing younger adults without houses wishing to purchase their first
house (or those with houses in low priced areas wishing to trade up)
increased over this period.
CONCLUSIONS
The purpose of this paper was firstly to emphasise the pivotal role
that housing plays in relation to social outcomes as well as economic
outcomes. Accepting the importance of this role, the paper has sought
to outline an analytical framework that treats the housing market as a
dynamic system incorporating life-cycles of both individuals/households
and of houses themselves. While the purpose is not to promote a
particular set of housing policies, this way of interpreting the sector
is highly relevant to the development of housing policy. Policy options
must be viewed within a system context that emphasises both the dynamic
choices made by individuals and the long-range implications of housing
investment decisions.
The nature of housing as a fulcrum between economic and social
factors is illustrated by the rental yield results. Economic factors
may explain why it is that more deprived areas tend to have high rents
relative to house prices than do less deprived areas. While potentially
explicable, the social consequences of this relationship require
careful analysis as to their dynamic implications regarding potential
poverty traps and widening wealth disparity.
Finally, even if these social consequences are known, policy
responses are far from automatic. The systems nature of housing, and
the slow response of the overall housing stock to price and policy
changes, means that housing assistance targeted at one population
segment is often at the expense of the housing fortunes of another
sector. A long-term systems view, which takes account of the
life-cycles of both individuals and of houses, is required in
formulating the resulting policies.
ACKNOWLEDGMENTS
We thank two referees for their insightful comments on an earlier
draft. We also thank the Centre for Housing Research Aotearoa New
Zealand (CHRANZ), and the Foundation for Research, Science and
Technology (programme on Understanding Adjustment and Inequality)
for funding assistance. Neither these organisations nor the referees
are responsible for the views expressed.
REFERENCES
Alatini MN 2004. Housing and related
social conditions of the Tongan community living in Otara. Unpublished
MA thesis, University of Auckland, Auckland, New Zealand.
Bernheim BD,
Schleifer A, Summers LH 1985. The strategic bequest motive. Journal of
Political Economy 93: 1045–1076.
Can A 1992. Specification and estimation of
hedonic housing price models. Regional Science and Urban Economics
22(3): 453–474.
Cheer T, Kearns R, Murphy L
2002. Housing policy, poverty and culture: “discounting” decisions
among Pacific peoples in Auckland, New Zealand. Environment and
Planning C: Government and Policy 20: 497–516.
Collins M, Crowe D,
Carliner M 2002. Supply-side constraints on low-income homeownership.
In: Retsinas NP, Belsky ES ed. Low-income homeownership: examining the
unexamined goal. Washington DC, Brookings Institution. Pp. 175–199.
Crampton
P, Salmond C, Kirkpatrick R, Scarborough R, Skelly C 2000. Degrees of
deprivation in New Zealand: an atlas of socioeconomic difference.
Auckland, David Bateman.
Dieleman FM 2001. Modelling residential
mobility: a review of recent trends in research. Journal of Housing and
the Built Environment 16(3–4): 249–265.
DTZ New Zealand 2005. Housing tenure
aspirations and attainment. Research report for Centre for Housing
Research Aotearoa New Zealand and Building Research, Wellington.
Dubin RA 1992. Spatial autocorrelation and
neighborhood quality. Regional Science and Urban Economics 22(3):
433–452.
Durlauf S 2004. Neighbourhood effects.
In: Henderson V, Thisse J-F ed. Handbook of regional and urban
economics Vol. 4. Amsterdam, North-Holland. Pp. 1–102.
Furstenberg F, Hughes M 1995.
Social capital and successful development among at-risk youth. Journal
of Marriage and the Family 57: 580–592.
Galster G, Pettit
K, Santiago A, Tatian P 2002. The impact of supportive housing on
neighborhood crime rates. Journal of Urban Affairs 24(3): 289–315.
Glaeser E, Sacerdote B 2000.
The social consequences of housing. NBER Working Paper 8034. Cambridge
MA, National Bureau of Economic Research.
Green RK, White MJ 1997. Measuring
the benefits of homeowning: the effects on children. Journal of Urban
Economics 41(3): 441–461.
Grigsby W 1963. Housing markets and
public policy. Pennsylvania, University of Pennsylvania Press.
Grimes A, Aitken A 2004. What’s the
beef with house prices: economic shocks and local housing markets. Motu
Working Paper 04-08. Wellington, Motu Economic and Public Policy
Research.
Grimes A, Holmes F, Bowden R
2000. An ANZAC dollar? Currency union and business development.
Wellington, Institute of Policy Studies.
Grimes A, Kerr S, Aitken A
2004. Bi-directional impacts of economic, social and environmental
changes and the New Zealand housing market. Motu Working Paper 2004-09.
Wellington, Motu Economic and Public Policy Research.
Grimes A, Sourell R, Aitken
A 2005. Regional variation in rental costs for larger households. Motu
Working Paper 2005-02. Wellington, Motu Economic and Public Policy
Research.
Haurin D, Parcel T, Haurin J
2002. Does homeownership affect child outcomes. Real Estate Economics
30(4): 635–666.
Howden-Chapman P, Carroll P
ed. 2004. Housing and health: research, policy and innovation.
Wellington, Steele Roberts.
Jencks C 1994. The homeless. Cambridge MA,
Harvard University Press.
Leventhal T, Brooks-Gunn J
2000. The neighborhoods they live in: the effects of neighborhood
residence on child and adolescent outcomes. Psychological Bulletin 126:
309–337.
McClay S, Harrison R 2003. The
impact of school zoning on residential house prices in New Zealand.
Paper presented to New Zealand Association of Economists conference,
University of Canterbury Economics Department, Christchurch.
McNicholas
A, Lennon D, Crampton P, Howden-Chapman P 2000. Overcrowding and
infectious diseases—when will we learn the lessons of our past? New
Zealand Medical Journal 113: 453–454.
Maré D, Timmins J 2003.
Moving to jobs. Motu Working Paper 2003-07. Wellington, Motu Economic
and Public Policy Research.
Meen G 2001. Modelling spatial housing
markets: theory, analysis and policy. Boston, Kluwer Academic
Publishers.
O’Flaherty B 1995. An economic
theory of homelessness and housing. Journal of Housing Economics 4:
13–49.
O’Flaherty B 1996. Making room: the
economics of homelessness. Cambridge MA, Harvard University Press.
Reserve Bank of New
Zealand 2006. Monetary policy statement, March, Wellington.
Rohe WM, Van Zandt S,
McCarthy G 2002. Social benefits and costs of homeownership. In:
Retsinas NP, Belsky ES ed. Low-income homeownership: examining the
unexamined goal. Washington DC, Brookings Institution. Pp. 381–406.
Rossi PH 1955. Why families move. Beverley
Hills & London, Sage Publications. [2nd ed. 1980]
Smith KS, Fraser R 2004. National
landlords survey: preliminary analysis of the data. Wellington, Centre
for Research, Evaluation and Social Assessment.
Sweeney J 1974. A commodity hierarchy
model of the rental housing market. Journal of Urban Economics 1:
288–323.
Wheaton W 1990. Vacancy, search and
prices in a housing market matching model. Journal of Political Economy
98: 1270–1292.
Table 1 Correlations between
socio-economic factors and the rental yield.
| |
House |
Apartment |
| Median income (2001) |
Negative |
Negative |
| Deprivation index score (2001) |
Positive |
Positive |
| Proportion of people employed (2001) |
Negative |
Negative |
| Proportion of people with no education (2001) |
Positive |
Positive |
| Proportion of people with university degree (2001) |
Negative |
Negative |
All correlation coefficients are significant at 1% level.
Fig. 1 Relationship between rental yield
and education.
1One example is the multi-disciplinary
research programme currently studying housing affordability issues in
Nelson, Tasman, and Marlborough that involves researchers with
sociology, economics, law, and property services backgrounds. The study
is being conducted under the auspices of the Centre for Housing
Research Aotearoa New Zealand (CHRANZ); see link (last
accessed: 7 April 2006).
2Some studies, especially within the
sociology and geography literature, refer to this approach as following
the life-course rather than the life-cycle. We are agnostic as to the
terminology and follow the latter, more traditional, usage. Our purpose
in using this construct is primarily descriptive, without implying any
normative content.
3This is often the case even when there
is considerable family wealth, since wealth is often retained by
parents and passed on substantially only at death (Bernheim et al.
1985).
4After analysing the international
literature, Dieleman (2001) concluded that “there is a strong
correlation between the rate of mobility and the stage in the life
cycle of a person. In all developed societies, young adults between the
ages of 20 and 35 are by far the most mobile segments of the
population.”
5For example, in the 1990/91 New Zealand
recession, almost 2% of New Zealand’s labour force migrated to
Australia (Grimes et al. 2000).
6In 2001, 38% of 25–29 year olds had
post-school qualifications compared with just 26% of 55–59 year olds
(source: New Zealand Census, Statistics New Zealand), illustrating the
substantial increase in post-school study over the intervening 30
years.
7See McClay & Harrison (2003) for
preliminary evidence on the school zone premium in Christchurch.
8This issue was reported in oral
interviews conducted in preparing Grimes et al. (2004), particularly by
Maori considering a return to areas of tribal affiliation.
9We use the word “houses” to refer to any
type of dwelling, not just to stand-alone residences. At times, we make
specific reference to certain types of dwelling, such as apartments or
mobile homes.
10There may also be an effect of drawing
in unqualified (and sometimes untrained) labour to the construction
industry, which may result in a deterioration of construction quality.
This latter effect will be most severe in areas in which building codes
are not strictly enforced.
11For instance, O’Flaherty’s (1995,
1996) analysis of the phenomenon of homelessness in the United States
uses this framework. By doing so, it adds considerable depth to the
treatment of this subject relative to the treatment by Jencks (1994)
which does not include explicit reference to the housing life-cycle.
12Specific supply responses will
nevertheless be important for meeting needs of particular groups. For
example, mental health clients may face particular needs that are not
being met by the private housing market.
13Housing assistance policies may also
have some impact on rents if landlords incorporate the effects of
housing assistance paid to tenants into the rents that they charge.
There is little reliable evidence in New Zealand, however, as to
whether this issue has empirical relevance or not.
This year's abstracts | Journal home page | All abstracts | Publishing home page
