1 INTRODUCTION

1.1 Research scope and purpose

Housing policies are implemented for different reasons, one of which is to improve
housing outcomes for lower income households. To achieve that objective, countries
adopt different strategies, suggesting that differences in housing outcomes for low
income households might be expected. The broad concern of this study is whether the
housing outcomes for low income households are different when different housing
policies apply.

Many factors, in addition to housing policy, impinge on housing outcomes in the short
and the longer term. Among these, the influence of economic conditions, political and
governance systems, demographic factors, welfare regimes, the urbanisation process and
cultural traditions has been highlighted in the diverse field of housing-related research. In
this study, it is the relative importance of housing policy strategy - defined by Kemeny
(1995a) as an enduring set of principles (and assumptions about how housing markets
function) underpinning specific policy interventions - that provides the analytical focus.

To constitute its empirical and historical evidence, this study adopts a comparative
approach using two case studies, Australia and the Netherlands. An extended in-depth
study of two cases, rather than a (necessarily) more superficial examination of a larger
number of cases, is considered essential to isolate housing policy effects and to interpret
their significance among the multitude of factors and social relations contributing to the
housing system of any country.

This study contributes to two prominent contemporary areas of research and policy
debate. One debate concerns understanding the role of government (or, more abstractly,
the state) in housing. That debate manifests itself in both research and policy arenas
through strongly contested theoretical views and differing empirical evidence about the
scale, duration and type of subsidy, regulation and substitution of housing markets that is,
depending on the viewpoint chosen, either likely, necessary, appropriate or effective. In
that context, a study of housing outcomes under different national policy strategies is
seen as a means of elucidating their influence and significance.

A second debate to which this study contributes concerns discussion of the effect of
economic and social adjustment that has been occurring across advanced capitalist
societies under the processes of globalisation and demographic change on the
sustainability of particular national housing policy approaches. In that context, the study
is designed to contribute to an assessment of the way in which distinctive national
approaches to housing policy are promoting or constraining housing outcomes in the
current conjuncture.

The emphasis on comparing housing outcomes under contrasting policy regimes over an
extended period is central to the purpose and design of the study. Placing emphasis on
the broader and extended consequences of different policy strategies, rather than on their
more specific and short-term effects, shifts the comparative perspective from one that is
19
primarily concerned with differences in the form, level and delivery of housing assistance
(that is, with specific outputs of a housing system) to one that probes the more
fundamental social and economic consequences of different approaches to the provision
of housing (that is, the broad social outcomes) and, thereby, helps to reveal underlying
similarities and differences in the structure of society and the way that policy choices are
made (Golland, 1998).

The following outline of the cases chosen, the time period of analysis and the proposed
empirical focus provides more detail on the scope and design of the study.

1.1.1 The cases

The case studies chosen have many similar general features: they are advanced capitalist
countries with dependent, export oriented economies, they have small populations and
comparable 20th century demographic trajectories and they have experienced dramatic
economic restructuring and social change in the last three decades. Their value to this
study is that, while their economic, demographic and social development is comparable
in many respects, they differ in their long-term pursuit of highly distinctive national
housing (and welfare) policy strategies.

Australia has a housing and urban policy practice of enabling mass home ownership and
low density suburbanisation through publicly supported market allocations,
supplemented by a small, administratively managed public housing system. In contrast,
housing and urban policies in the Netherlands have emphasised medium and high density
mass social housing provision orchestrated through comprehensive government
regulation and a high rate of public investment and, since the 1980s, delivered mainly
through an independent not-for-profit sector. Overall, reliance on market forces has been
much more limited than in the Australian case.

1.1.2 The period of analysis

Primarily, the study assesses housing outcomes at the end of the 20th century. However,
an extended period of analysis is required to interpret the impact of long-term housing
policy strategy on these outcomes. The housing policy development of both countries
after 1945 is considered in two broad phases.1 The first phase (extending until roughly
1980) marks the major growth in, and consolidation of, each country’s characteristic
regime of housing provision, as part of the extensive development of a particular welfare
state model.2 In each case, the general context for the evolution of the definitive housing
system is a period of rapidly increasing population and urbanisation, sustained economic
growth and rising community affluence.

The subsequent phase (extending over the last two decades of the 20th century) is
distinguished by the characteristics of the second demographic transition and globally
driven economic change. In particular: flexible labour market arrangements; the rapid
growth in smaller households, an increasing number of whom are outside the workforce;
the polarisation of household incomes and an ageing population have been identified as
significant interacting factors contributing to more diversity in housing demand and to
new housing market and tenure dynamics. (See Berry, 1999 and Yates, 2002a, 2000b on
20
the manifestation of some of these changes in Australia and, for the Netherlands, see
Marcuse and van Kempen, 2000 and van Kempen, 2001.) These extensive societal
changes have been accompanied by clearly discernible shifts in previous welfare and
housing policy strategies, involving attempts to reduce total government expenditure on
welfare and, in the case of housing policy, placing more reliance on market provision and
on consumer (rather than producer) subsidies.

1.1.3 The empirical focus

Housing policies have multiple social objectives and complex impacts and outcomes.
However, there is a need to define a manageable task for this research. For that reason,
one major social outcome area - housing affordability (that is, the relationship between
housing costs and income) for low income households - has been selected as the
empirical focus for the study.

Using affordability to evaluate the significance of different housing policy strategies can
be justified on a number of grounds. House prices and rents and their impact on
affordability for consumers are fundamental to the operation of contemporary market-
based housing systems. At an individual level, housing costs are one of the most
significant elements of a household budget across the life course. The affordability of
housing also influences the quality and quantity of housing that can be obtained and
maintained as household income changes. As well, the location of housing that is
affordable to a particular household influences the broader economic and social
opportunities and personal life choices of each member of that household.

The restriction of the empirical evidence to affordability outcomes has the disadvantage
that other differences (or similarities) between the two cases may be overlooked.
However, affordability is a multi-dimensional issue. Measures can be developed to
indicate how much housing is affordable, at what quality, in which locations and under
what legal rights. It has been demonstrated (though with varying rigour) that a wide
range of indirect effects (sometimes referred to as non-housing or non-shelter outcomes)
arise from unaffordable housing or from inadequate, but more affordable, housing.
Individual households experience these effects and they may also flow through to the
neighbourhood and to the broader community. Some of the main effects of poor
affordability or inappropriate housing on individuals that have been recognised and
researched include: family instability and breakdown, health impacts, the propensity for
criminal activity, labour force participation and several dimensions of what is referred to,
broadly, as social exclusion.3 The concept and dimensions of affordability used in this
study are described in chapter 3. Chapter 6 considers affordability conditions for low
income households in the two cases.

Housing policies directly and indirectly affect the housing circumstances of all
households. However, the need to ensure that lower income households have access to
appropriate and affordable housing has been a core rationale for, and significant
component of, state intervention in the housing systems of both the Netherlands and
Australia over the period of this study.4 The precise target group for such assistance has
varied across each case and over time but, generally, has fallen within the lower half of
the income distribution. A consistent definition of low income households is required for
21
comparative analysis to be meaningful. Accordingly, households in the lowest forty per
cent of the income distribution have been chosen as the standard against which the study
evaluates the success of the national objectives and related strategies to assist low income
households. In each case, this is a reasonably broad group, comprising over 2½ million
households at the end of the 1990s. (See chapter 6.) It encompasses the range of
households to whom specific assistance may be targeted and those who could be
considered, because of their limited income, to be at risk of having difficulty obtaining
affordable and appropriate housing without assistance, in some sub-markets at least.
Where appropriate, a further distinction is drawn between households in the lower half
(quintile one) and those in the upper half of the group (quintile two).

1.2 Research questions

Given the broad research interest and the context for the research just outlined, the three
core research questions that have been chosen to guide the study are:

1. How have the Dutch and Australian governments intervened in their respective
housing systems over the second half of the 20th century to help provide more
affordable housing?

2. To what extent are the housing policy strategies of Australia and the Netherlands
providing affordable housing for lower income households?

3. Which aspects of past and present housing policy strategies, mediated through
their local economic, political and social context, are judged to be significant in
sustaining or constraining the provision of affordable housing for lower income
households?

In subsequent chapters, theoretical, historical and empirical perspectives will be brought

to bear on these central questions as appropriate. A general overview of the method and
logical framework for the study is provided next. This chapter concludes by outlining
how the rest of the study is structured and presented.

1.3 Comparative analysis

The central logic of the comparative method is to use comparison as a device to identify
and explain the similarities and differences in a matter of interest across more than one
case (Pickvance, 2001). This study uses the comparative method as a tool for
investigating the dynamic relationship between housing policy strategies and housing
outcomes in the context of the particular economic, political and social environment of
two purposively chosen cases.

Critical reviews by Harloe and Martens (1984) and Oxley (1989, 1991, 2001) highlight
the strengths and weaknesses of established approaches to comparative policy research in
the housing field. In support of the continuation of comparative analysis, Oxley (2001)
argues that many of the problems associated with previous research could have been
overcome by clear enunciation of the purpose of adopting comparison as a means of
explanatory analysis and by better aligning the research method with the chosen aims. In
22
response to those propositions, the following sub-sections set out: the rationale for
choosing Australia and the Netherlands for comparison in this study, the case for the
focus on policy, the method of comparative analysis and explanation adopted and,
finally, some general problems that have been encountered during the study and the
strategies used to deal with them.

1.3.1 The choice of case studies

A review of the history and logic of comparative housing and welfare studies research
shows why the cases of Australia and the Netherlands suit the central purpose of this
study.

The most general comparative housing studies typically involve a large and diverse
group of nations or cities. These studies are predominantly numerical or statistical in
style, developing and/or applying a large, standardised set of macro-indicators of housing
conditions to describe and compare global housing systems. One of their main
advantages is that they can give a preliminary indication of issues worthy of more in-
depth analysis. One of their main problems, however, is that they ignore, or have no way
of dealing with, underlying structural or cultural differences in apparently similar
measures. (See Priemus (1992) for a critical overview of this kind of study.)

Many comparative researchers have restricted their analysis to a cluster of ‘like

countries’ to limit the variability in the analysis or, more arbitrarily, because policy
interests may be shared or the research effort made more convenient. Countries may be
chosen because: they are considered similar in terms of their stage of economic
development, they are members of a regional political grouping (for example, the
European Union) or they have cultural and/or linguistic links. The usual approach of
studies of this type has been to compare and to analyse issues or trends in housing policy
and its effects using a historical narrative method, drawing on evidence of both a
quantitative and qualitative kind.5 In many of these multi-nation studies, the comparison
has been weakened by discussing each case separately, using, at best, a common
analytical framework to achieve a comparative perspective.

Kemeny (1995a) also argues that there has been a tendency in many studies towards
cultural bias in the choice of countries for comparative investigation. In particular, he
considers there has been an over-concentration by English language researchers on
comparative research across Anglo-Saxon countries only. This has led to his own interest
in explicitly comparing European and Anglo models of housing provision and in
theorising the role of cultural factors in the determination of national housing policy
directions.

Where fewer cases are examined, awareness of the significance of the issue of case
selection may be heightened. Accordingly, intensive case studies are often chosen
because the particular nature of the distinction between the cases is theorised to be
fruitful for isolating new principles or developing what Kemeny and Lowe (1998) have
called ‘middle range’ theory - something between the explanatory notions of individual
differences or universal laws.

23
In particular, over the last decade, the thesis that advanced capitalist countries have
different welfare regimes, forged by different political power relations and institutional
arrangements, has come to the fore as a rationale for the selection of cases in comparative
welfare research (Esping-Andersen, 1990). The theory of welfare regimes will be
considered in more detail in chapter 2. However, because of its strong influence over the
design of comparative research in the welfare and housing fields, an overview of the
argument for linking case study selection to this framework is included in this section.

Esping-Andersen’s typology distinguishes three welfare regime types among nations

with advanced welfare states. A social democratic welfare regime is found in countries
where there has been a powerful labour movement and a long period of ‘left-leaning’
governments. Under this regime, social services are the most decommodified and are
provided universally to a high standard. Sweden is the archetypal case. In corporatist
societies, traditional groups and structures retain a powerful role in political institutions
and processes, resulting in coalitions of the main power groupings forging social policy
that, in turn, reflects the relative power of the parties involved. Corporatist processes tend
to result in social programs where church and family institutions play a primary role,
thereby preserving and reinforcing pre-existing stratifications in society, whether by
religion, class or ethnicity. Much of continental Europe, including the Netherlands, falls
within this regime type. Liberal traditions are dominant in more recently settled countries
without any significant pre-capitalist history or structure, especially where there are weak
or disunited labour movements. A market approach to service provision dominates and
state services are residualised and only provided to those for whom there is no market
provision. The United States of America, Canada, Australia and New Zealand are
considered to be archetypal, though different, examples (Esping-Andersen, 1990).6

A number of housing researchers have used Esping-Andersen’s classification to describe

and differentiate particular national housing policy approaches and to compare their
performance. (See, for example, Barlow and Duncan, 1994; Kleinman, 1996 and
Golland, 1998.) The view about the broad relationship between types of housing policies
and different welfare state regimes that has emerged from these studies is that:
• a liberal regime is typically aligned with residual social housing provision and
market-based home ownership policies;
• a social democratic regime typically involves mass social housing provision and a
strong state regulatory framework; and
• a corporatist regime normally exhibits widespread state support for a range of
housing options provided through both market and non-market institutions.

Of direct relevance to this study, several applications of welfare regimes analysis have
suggested that Australia and the Netherlands may be atypical or ‘puzzling’ cases, at least
with respect to some aspects of Esping-Andersen’s typology and underlying theory
(Castles, 1989a). For instance, the central and substantial role of social housing in the
Dutch welfare state is similar to the Scandinavian countries. However, the explanatory
factors which Esping-Andersen and others have associated with the development of the
social democratic welfare state exemplified by the Scandinavian cases, such as a close
alignment of political power and organised labour, do not pertain to the Netherlands
(Therborn, 1989; Barlow and Duncan, 1994; Harloe, 1995).

24
Using Australia as his test case, Castles (1989b, 1994) has remarked particularly on the
disparity between policy characteristics associated with a small and residual welfare state
and social outcomes that (at least until the 1970s) have reflected a high level of equality
and social protection. His observations have led to his claim for recognition of a
distinctive ‘wage earners’ welfare state’ where, he argues, higher wages (and their
corollary, lower taxes) achieved by the labour movement have obviated the need for (and
subsequent possibility of) extensive social services (Castles and Mitchell, 1993).

Adding to his argument more recently, Castles (1997a, 1998) has suggested that housing,
which was not included in the data for the original development of the welfare states’
typology, may be a key variable that has contributed to the ‘better than predicted’ welfare
outcomes in Australia. Using government support for home ownership as an example,
Castles has shown that older people (who are home owners) in Australia receive a similar
level of benefit to their European counterparts when what is compared is net income
(inclusive of housing benefits), rather than welfare expenditures or pension levels. From
this observation, he concludes that many previous studies may have exaggerated the
difference between welfare regimes because they have failed to recognise that different
countries achieve similar outcomes through diverse institutional and policy tools (ibid.).

Through the development of the comparative research field, there is now general
recognition that past propositions about differences in welfare states have been too
simplistic. Similarities at one level - the form of policies or institutions - may have
masked differences at another level - the processes operating to produce those results.
Alternatively, a focus on the differences in policies may hide similarities in their impacts.
Consequently, more in-depth studies of the diversity of approaches to government
intervention have been called for to develop more historically grounded and realistic
hypotheses about the connections between different types and levels of intervention and
social outcomes. “Surveying the diversity and identifying its outcomes is, arguably, the
best definition of the proper scope of the comparative analysis of public policy” (Castles,
1997a, p. 118). One suggested way forward, which is adopted in this study, is to use
hitherto unexplained, atypical or extreme cases that, by exaggerating particular
conditions, may allow new principles to emerge (Castles, 1989a; Barlow and Duncan,
1994; Pickvance, 2001).

Previous instances of comparative housing research involving Australia and/or the

Netherlands reflect the general pattern of comparative studies outlined above. Typically,
both countries are included in statistically based global studies and studies of OECD
members, which use standardized data sets. Australia, a geographically isolated country,
has not been included generally in more in-depth comparative studies of a cluster of
countries. The inclusion of the Netherlands in these studies is more frequent, especially
in Western Europe. Perhaps in keeping with Kemeny’s (1995a) assessment of cultural
bias, it has been rare for both countries to be included in one intensive study. To the
author’s knowledge, they had not been compared directly previously in any major
housing study.7 Nevertheless, as illustrated above, studies of either Australia or the
Netherlands that have used a housing perspective have made an important contribution to
the development of comparative welfare research.

25
Overall, the choice of Australia and the Netherlands suits the purpose of the present study
and is in keeping with the current state of comparative research. They are similarly
developed countries but with distinctive housing policy (and welfare) regimes.
Conceptually, this means that it should be possible to better isolate the effects of housing
policy, than if other factors considered to influence housing outcomes, such as population
pressure or the level of economic prosperity, also varied greatly. They have been
considered atypical or puzzling cases because existing theories do not account adequately
for the relationship between their institutional structures, policy choices and their welfare
performance. Added to this is the evidence from previous research that the role of
housing policy in their respective welfare systems may be a critical factor in explaining
their broad welfare outcomes. Finally, evidence from these cases will contribute to
informing the relatively underdeveloped cross-cultural perspective in housing studies.

1.3.2 Policy evaluation as the focus

This study is concerned with policy evaluation to inform policy making. Therefore, it has
to take note of the success of similar studies and recognise and address practical,
methodological and theoretical impediments to policy evaluation.

At a theoretical level, the traditional political science notion of the state as a benevolent
and ‘independent’ or neutral agency in social action is now generally discredited.
Nevertheless, grappling with precisely why and how state action is formulated remains a
fundamental question for policy research. A more specific conceptual and practical
problem for policy evaluation arises from the lack of clarity and specificity in the goals
of much state policy (van Vliet, 1990; Lundqvist; 1991). A definitional issue surrounds
the determination of the scope of the evaluation - for instance, non-housing policies, such
as employment policy, may have a greater impact on housing outcomes than housing
policies. Differences between the ‘intent’ and implementation of policy, and the
complexity and diffuse nature of its impacts, mean that validating the claim for policy
impacts becomes a significant methodological issue. For instance, a number of other
housing studies have demonstrated how policy-centred analyses may have led to the
influence of housing policy being overstated, by drawing out the strong influence of
macro-economic and demographic factors on a housing system. (See, for example,
Feddes, 1995; Bourassa et al., 1995 and van Kempen et al., 2000.)

A strong case can be made for continuing to pursue research based in policy evaluation,
if, at the same time, the difficulties, both perceived and real, are acknowledged. Ambrose
(1992) considers the means to more robust research of this kind is through the
development of a better understanding and definition of housing problems and a more
rigorous assessment of the performance of housing systems. However, as he points out,
there is little agreement about how to proceed. For example, there has been only limited
development and discussion of criteria that may be appropriate for assessing the
performance of housing systems or for comparing policy impacts.

Lundqvist (1991) argues a clear case for why policy-centred analyses are a valid,
practical and necessary focus for trying to explain how national differences arise in the
context of broader societal trends. “Policies …are the crucial link between theories of
context-content connections on the one hand, and propositions about content-
26
consequence relationships on the other” (ibid. p. 88). In practical terms, “policies do exist
in most countries. They can be identified without much difficulty…”(ibid.). “What is
required is independent analysis of policy and testing and theorising about the relative
influence of policy among other determinants (such as institutional forces, demographic
trends, economic developments) on an arena of impact, such as housing conditions”
(ibid. p. 86).

As discussed earlier, the starting point for undertaking comparative analysis is to

determine what is being compared and for what purpose. The debate about the quality
and validity of policy studies in housing research points to some constituents of a robust
conceptual framework for this study. First, the approach must recognise the historically
variable nature of the relationship between housing policies and their outcomes. Second,
the policy variables of interest need to go beyond those of direct housing policy itself to
incorporate the potential impact of a wider range of state actions. Third, previous studies
underscore the need for the context of housing policy to be included in the analytical
framework. Finally, appropriate criteria for judging policy impacts are required. The
comparative methodology of this study is now considered in more detail.

1.3.3 The method of analysis

Taking into account the record of policy-related research referred to above, three core
elements of comparative analysis are proposed as an initial conceptual framework (figure
1.1). This framework is further developed around concepts of the structure and
determinants of the attributes of a housing system in chapters 2 and 3 of the study.

POLICY CONTEXT CONSEQUENCES

Government The housing system and Housing affordability

interventions in the its economic, political, conditions of low income
housing system social and demographic households
context

Variables of influence Mediating variables Variables of interest

Figure 1.1 Logical relationships of research variables in a dynamic system

Source: adapted from Lundqvist (1991) and Rudestam and Newton (1992).

The variables of interest are similarities and differences in the housing conditions of low
income households. The variables of influence are differences in the national housing
policy strategies of the chosen cases. In this research, housing policy is defined as public
policy in relation to housing (Somerville, 1994). Both explicit (formal) and implicit
(indirect) forms of policy impacting on housing are recognised and discussed. In the case
27
of indirect policy, housing impacts arising from macro-economic, welfare, fiscal,
monetary and urban planning policy settings are included in the assessment.

The recognised intervening variables are the historically specific contexts within which
these policies form and are continually transformed in a dynamic housing system. The
local context for each case study includes the operations of the housing market, the
institutional arrangements through which housing is produced and consumed and factors
exogenous to the housing system itself that influence the conditions of housing provision.
The analysis of the housing policy strategies of the two case studies and the context of
their development is analysed in two parts: the period from 1945 to the 1970s is
considered in chapter 4 and the period of the 1980s and 1990s is discussed in chapter 5.

The separation of housing policy from the system of housing provision in the framework
above is intended as an analytical and presentational device, rather than to imply a ‘real’
or sustainable distinction. As discussed in chapter 2, housing policy is logically and
historically embedded in a wider system of housing provision and its local political
economy. The interpretation of the relationship between specific policy interventions, the
total housing system and the broader context is built up progressively through the
development of this study to highlight both policy differences and the particular ways
they have operated alongside other important political, economic, socio demographic and
cultural/institutional factors to affect the housing consequences under consideration.
Narrowing the empirical focus to one aspect of the ‘dependent’ variable - the
affordability of housing and its consequences - has provided another means of managing
the potentially enormous array of policy strategies and critical relationships that have to
be considered.

Another step in tackling the research questions involves establishing an adequate basis
for explaining the relationships expressed in the model above.

This study is concerned with explaining observed similarities and differences in the
variables of interest (in this case, affordability outcomes for low income households) in
terms of variations in logically related and inter-related variables (in this case, different
national housing policy strategies and the context within which they have operated). The
causal relations between these variables are complex and dynamic. Following Pickvance
(2001), this study uses a schema depicting the range of possible explanations of
differences and similarities in housing outcomes as an analytical framework to guide the
interpretation of the historical and empirical data later in the study (table 1.1). The
framework exposes, at the outset, the logical possibilities for explaining the relationship
between the policy and contextual analysis, on the one hand, and the affordability
outcomes that have occurred, on the other. It is anticipated that different causal models
may pertain to different aspects of the findings of the study.

The final aspect of the method involves considering the evidence to be used in the
analysis. The variety and complexity of relationships being considered cannot be
addressed adequately by a singular set of data or method of analysis (such as drawing
inferences from a statistical analysis). Instead, the study relies on a careful and logical
approach to synthesising and interpreting theoretical, historical and empirical (both
quantitative and qualitative) evidence.
28
Table 1.1 A framework for considering the relationship between housing policy and housing
outcomes

DIFFERENT POLICIES
DIVERGING a B: Multiple causation C: Plural causation with
A: Simple causation
EFFECTS differentiating outcomes
(Outcomes Housing policy and Other factors as well as There is over time/over
specific to each outcomes are directly housing policy explain place variability in the
case) related. significant differences. relationship between a
particular policy and its
For example, there is a For example, the
effects.
direct relationship differences in
between the level of affordability can be Outcomes diverge not only
expenditure on housing attributed to both as a result of policies but
subsidies and differences in economic also because of the dynamic
affordability. and demographic factors interaction between factors
and to differences in that mediate the association
housing subsidy levels. between policy and
outcomes. This may be
cumulative.
For example, home
ownership policies succeed
in improving access to
affordable housing in the
context of rising affluence
but not in the context of
polarising incomes and
unstable labour markets.
SIMILAR OR D: No causal E: Masked Relationship F: Plural causation with
CONVERGING relationship homogenising outcomes
EFFECTS
Housing policy has a Different policies have There is a greater variation
(Outcomes are marginal impact on similar effects. in policies than in outcomes.
cross-national) outcomes. Exogenous Policies and exogenous
For example, each
factors dominate and factors can have
country chooses a
outcomes converge. countermanding effects.
different policy path but
For example, rising the end effects on the For example, subsidy levels
affluence before the provision of affordable are different but
1980s and growing housing are similar. affordability outcomes are
income inequality since similar because market
then are the major forces moderate the impact
determinants of the of the differences in
form of, and changes in, subsidy.
affordability patterns.

a
Categories developed from Pickvance (2001)

1.3.4 Difficulties in making cross-national/cross-cultural comparisons

In addition to conceptual issues already referred to, all comparative analysis presents a
range of methodological and practical difficulties for research. While some of these
problems can be addressed in the research design, others have to be managed through the
research process. In some instances, they may qualify the findings that can be claimed for
29
the study. Particular problems encountered in this study (and the means adopted to deal
with them) are identified and discussed where they arise. Some of the more general
problems that have influenced the overall design and execution of the study are discussed
below.

The choice of cross-cultural case studies presents a particular challenge to a solo

researcher. At a practical level, much primary source material and data in a foreign
language are not accessible. As there is a greater potential for misinterpretation or
distortion of secondary sources, local housing experts are needed to check and validate
the interpretation of the foreign case material (see appendix C). In terms of the research
process, the comparative researcher is, at the same time, an ‘insider’ in his/her own
milieu and an ‘outsider’ in the foreign setting. Of course, this brings a benefit, as the
unfamiliar environment gives rise to a challenge to ‘taken for granted stances’ and new
questions emerge. Ultimately, as Oxley (2001) notes, the extent to which a thorough and
nuanced understanding is established will influence whether a comparative study can
contribute credible views about the validity and transferability of policy and practice.
Overall, however, it is unavoidable that a different level of appreciation and
understanding of a system in another linguistic and cultural milieu will occur and this
situation is acknowledged in this study.8

Many data problems are encountered in comparing housing affordability both within and
across national systems. Within countries, definitions of key variables, such as household
income or housing outlays, may not be consistent across surveys or time periods. Surveys
of housing expenditure themselves differ in their purpose, scope and coverage. These
types of problems are compounded in cross-national comparison because survey
instruments and data definitions are rarely standardised across countries. In time series
data, there are usually a limited number of data points available, making it difficult to
draw valid conclusions. Where an assessment of long-term trends is required, it is
particularly important to recognise the context of each data point included, so that
allowance can be made for short-term deviations caused by cyclical or idiosyncratic
factors. Cross nationally, survey timing rarely coincides, adding a confounding factor.

Key concepts underpinning the structure of housing data sets are themselves culturally
and institutionally defined and bounded, with the effect that what may appear comparable
is not, when the historically and culturally specific form that a concept or construct takes
is recognised. The problems associated with the cross-national comparison of generically
named tenure categories, such as social housing or home ownership, provide one
example of this difficulty that has been widely acknowledged and discussed (Barlow and
Duncan, 1988; Ruonavaara, 1993; Marcuse, 1994; Priemus, 1997a).9

Different national research traditions also influence the kind and quality of information
available. For example, Australia has a long tradition of poverty research including a
unique emphasis on housing-related poverty (Fincher and Nieuwenhuysen, 1998). The
Netherlands by contrast has given more prominence to ‘housing-costs-to-income ratio’
approaches (see chapters 3 and 6) in assessing housing affordability and has not linked
trends in poverty to changing housing costs so strongly (SCP, personal communication).
Such differences of emphasis in national data collections are themselves revealing about

30
policy disposition, housing problems (both perceived and real) and sensitive issues
within national domains.
Because of the problems outlined above, this study has had to rely on a limited
comparative data set to obtain the main empirical evidence. To overcome some of the
constraints presented by that situation, the approach that has been taken has involved
backing that data up with previous research findings, looking for consistent evidence
across similar studies and, when possible, assessing the sensitivity of the results to
different constructs and measures. Because statistical inference or estimation has not
been attempted, precise values on single variables have not been required. Overall, it is
considered that the level of comparability of data that has been achieved for the two cases
warrants the generality of the conclusions that have been drawn.

Because of the complexity of comparing different measures of housing costs and housing
affordability, the basis for the comparison of the findings is explained in each section
where data are included. Chapter 3 provides more information on conceptual and
measurement issues affecting the assessment of housing affordability and appendix D
compares the features of the survey instruments and the databases drawn on for the
empirical analysis.

1.4 Structure of the report

As described above, the study uses three main analytical perspectives - conceptual,
historical and empirical - to address its three core questions. Six chapters follow this
introductory chapter. The purpose and contribution of each chapter is outlined below.

Theoretical propositions about the general impact of housing policy strategies on housing
outcomes are considered first. Chapter 2 investigates previous comparative housing
studies and relevant research from the related fields of urban and welfare studies. The
purpose is twofold. First, to select tools and concepts that can be used to identify
differences in national housing policy strategies and, second, to theorise the potential for
there to be a differential impact on housing outcomes of predominantly social and private
models of housing provision, both historically and in the current economic and political
conjuncture.

The concept of housing affordability, what is known about its determinants and
established ways of measuring and evaluating affordability outcomes form the subject
matter of chapter 3. The main concern of that chapter is to identify the logical
relationships between different forms of policy intervention and housing affordability
conditions and to propose a method for determining these conditions empirically.

Chapters 4 and 5 give a detailed account of the development, operation and

transformation of housing policy strategies in Australia and the Netherlands under
changing political and economic conditions. In answer to the first research question,
these chapters provide the historical evidence of the nature and extent of intervention
associated with housing affordability that has occurred, drawing out the extent of
similarity and dissimilarity between the two cases. An extended period of analysis from
1945 to the turn of the 20th century is used to establish a solid basis for assessing the
extent of consistency in any apparent relationship between housing policy and
31
affordability outcomes. Unlike in most comparative studies, the analysis of the two cases
is presented systematically, policy by policy, to facilitate a clear picture of where and to
what extent there are similarities and differences.

Two levels of analysis are entwined in the approach taken to documenting the policy
history. The first level considers the policy interventions that (in view of previous
research and theoretical propositions about the determinants of housing affordability) are
themselves likely to have most influenced (either directly or indirectly) housing costs and
housing affordability for low income households. The second level considers how such
policy interventions have interacted with each other, and with other major factors that
have influenced the evolution of the housing regimes of the respective cases, to give a
historically grounded basis for interpreting the relative and changing significance of these
factors.

As its third component, the study presents a quantitative and qualitative account of recent
affordability outcomes and, where available on a comparable basis, trends over the last
two decades. This material, presented in chapter 6, provides the response to question two,
concerning the affordability of housing for low income households. One aim is to
establish how far and in what ways the differences between the two contrasting systems
extend in terms of the chosen outcome area. The other aim is to consider the pattern of
changes within each case over time, so that these can be linked to the development of,
and adjustments in, policy regimes and to changes in external conditions - that is, to
establish the relative significance of policy factors versus other forces.

The three areas of analysis - logical arguments, historical analysis and empirical trends in
housing affordability - together provide the basis for answering question three about what
are the plausible connections between the observed outcomes in each case and the
different housing policy strategies, as they have evolved under the influence of local and
cross-national political and economic conditions. These considerations are the subject of
the concluding chapter.

Given the complex nature of the causality involved, it is not possible to explain fully the
outcomes that have occurred. Moreover, in contrast to many other housing studies
(reviewed in chapter 2), the purpose is not to explain why or how the identified policies
have been introduced. Instead, the main intent of the concluding chapter is to provide an
assessment of the broad influence and relative significance of different policy strategies
and to contribute to the debate about the efficacy and importance of particular housing
policy approaches in the present conjuncture. However, the analysis of outcomes also
stands alone (that is, whether they are ‘explained’ or not) in terms of providing a record -
albeit limited to the indicators that are used - of the comparative quality of the current
housing circumstances of low income households in two markedly different types of
housing systems.

Specifically, chapter 7 first considers to what extent the policy regimes in each place
have worked (and continue to work) to provide affordable and appropriate housing for
low income households. The comparison shows there are both strong similarities and
notable differences in affordability conditions across the two cases. The chapter then
draws on the extended comparative analysis to show how the respective housing policy
32
strategies have contributed to both the similarities and differences in the results. The key
elements of housing policy strategy that emerge as crucial in the current conjuncture are:
• the short and long run impact of housing policies favouring home ownership;
• the type and extent of urban policies that are geared to the supply and protection
of affordable housing;
• the role, organisation and capacity of social rental housing; and
• the effectiveness of demand-side subsidies (in particular, housing allowances).

Chapter 7 shows how each of these policy factors has contributed to the pattern of
affordability found today in each chosen country. It also includes a brief discussion of the
theoretical and policy implications of the findings.

Notes

1
Before 1945, national housing policy was piecemeal and less developed. The antecedents of housing
policy after 1945 are considered briefly in chapter 4.
2
More discussion of welfare states is provided later in the chapter. Following Esping-Andersen (1990),
Australia is usually classified as a residual (or targeted) welfare regime and the Netherlands as a corporatist
(or comprehensive) regime.
3
Overviews of the conceptual approach to assessing non-housing outcomes and/or recent research findings
are found in van Kempen (2001), Phibbs (2000), Berry (2001) and Mullins et al. (2001).
4
What constitutes affordable and appropriate housing is defined normatively in each society and is
evidenced by policy rules, minimum housing standards, community norms and expectations etc.
5
Recent examples of in-depth, cross-national housing studies include Harloe (1985, 1995); Boelhouwer
and van der Heijden (1992); McCrone and Stephens (1995); Kleinman (1996); Oxley and Smith (1996) and
Donner (2000).
6
Other applications of the typology have led to different views that are discussed further in chapter 2.
7
Concurrent with this research, Julie Lawson (University of Amsterdam) has used the historical
development of housing policy in Australia and the Netherlands to demonstrate the application of critical
realist methodology in comparative housing research (Lawson, 2003).
8
Having supervisors in each country and close contact with other researchers have been used to minimise
the impact of differential knowledge of each system.
9
Appendix A explains the use of tenure and other terms in this study.

33