WorkingPapers No. 12 SOCIUM SFB 1342 - Heinz Rothgang Johanna Fischer Meika Sternkopf Lorraine Frisina Doetter
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
SOCIUM SFB 1342
•
WorkingPapers No. 12
Heinz Rothgang
Johanna Fischer
Meika Sternkopf
Lorraine Frisina Doetter
The classification of
distinct long-term care
systems worldwide: the
empirical application of
an actor-centered multi-
dimensional typologyHeinz Rothgang, Johanna Fischer, Meika Sternkopf, Lorraine Frisina Doetter The classification of distinct long-term care systems worldwide: the empirical application of an actor-centered multi-dimensional typology SOCIUM SFB 1342 WorkingPapers, 12 Bremen: SOCIUM, SFB 1342, 2021 SOCIUM Forschungszentrum Ungleichheit und Sozialpolitik / Research Center on Inequality and Social Policy SFB 1342 Globale Entwicklungsdynamiken von Sozialpolitik / CRC 1342 Global Dynamics of Social Policy Postadresse / Postaddress: Postfach 33 04 40, D - 28334 Bremen Websites: https://www.socium.uni-bremen.de https://www.socialpolicydynamics.de [ISSN (Print) 2629-5733] [ISSN (Online) 2629-5741] Gefördert durch die Deutsche Forschungsgemeinschaft (DFG) Projektnummer 374666841 – SFB 1342
Heinz Rothgang
Johanna Fischer
Meika Sternkopf
Lorraine Frisina Doetter
The classification of distinct
long-term care systems worldwide:
the empirical application of an actor-
centered multi-dimensional typology
SOCIUM • SFB 1342
No. 12
Heinz Rothgang (rothgang@uni-bremen.de),
Johanna Fischer (johanna.fischer@uni-bremen.de),
Meika Sternkopf (meika.sternkopf@uni-bremen.de),
Lorraine Frisina Doetter (frisina@uni-bremen.de)
Collaborative Research Centre 1342 ‘Global Dynamics of
Social Policy’ and SOCIUM Research Center on Inequality and
Social Policy, University of Bremen.
Acknowledgements:
Funded by the Deutsche Forschungsgemeinschaft
(DFG, German Research Foundation) –
Projektnummer: 374666841 – SFB 1342.
We would like to thank Antonio Basilicata for his excellent
contributions to data collection and Ojarmyrat Gandymov for
assistance with formatting.Abstract
Long-term care (LTC) systems vary between countries in several ways. One im-
portant difference exists with regard to the question of who, that is which type
of corporate actor, takes over the main responsibility in providing, financing and
regulating LTC. In this article, we employ a multi-dimensional, actor-centered ty-
pology of LTC systems to classify all distinct LTC systems existing worldwide at the
point in time when they were first established. In doing so, the article contributes
to comparative LTC research by including novel cases and adding a historical
perspective. Our 18 cases fall into eight types, which we combine tentatively into
three distinct clusters: A predominantly state regulated and financed cluster, a
state regulated cluster with mixed financing and provision, and a cluster with pri-
vate regulation and provision plus societal financing. We find that the state plays
the major role in regulation (dominant in 16 countries) and financing (dominant
in 11 countries), while in provision we see a broader distribution with societal
and private for-profit actors taking a major role. Interestingly, and in contrast to
healthcare systems, no societal pure type emerges, not even among social insur-
ance countries.
[ii]Zusammenfassung
Die weltweit existierenden Pflegesicherungssysteme unterscheiden sich von Land
zu Land in vielfacher Hinsicht. Ein wichtiger Unterschied besteht darin, welche
Akteure die Hauptverantwortung für die Leistungserbringung, Finanzierung und
Regulierung der Langzeitpflege (LZP) inne haben. In diesem Beitrag verwenden
wir eine mehrdimensionale, akteurszentrierte Typologie, um alle weltweit vorhan-
denen eigenständigen LZP-Systeme zum Zeitpunkt ihrer Einführung zu klassifizie-
ren. Damit wird die vergleichende LZP-Forschung in zwei Richtungen erweitert:
Zum einen werden Fälle einbezogen, die in vergleichenden Darstellungen bis-
lang häufig nicht berücksichtigt werden und zum anderen wird eine historische
Perspektive hinzufügt. Die 18 Länder mit eigenständigen Sicherungssystemen ge-
hören zu acht Typen, die wir zu drei verschiedenen Clustern zusammenfassen:
Ein staatlich reguliertes und finanziertes Cluster, ein staatlich reguliertes Cluster
mit unterschiedlichen Akteuren in Finanzierung und Leistungserbringung und ein
Cluster mit privater Regulierung und Erbringung plus gesellschaftlicher Finanzie-
rung. Der Staat ist dabei der dominante Akteur bei der Regulierung (16 Länder)
und der Finanzierung (11 Länder), während wir bei der Leistungserbringung eine
breitere Verteilung sehen, bei der gesellschaftliche und private, gewinnorientierte
Akteure eine große Rolle spielen. Interessanterweise gibt es im Gegensatz zu
Gesundheitssystemen kein System mit der Dominanz gesellschaftlicher Akteure
in allen drei Dimensionen – auch nicht in Ländern mit einer sozialen Pflegeversi-
cherung.
SOCIUM • SFB 1342 WorkingPapers No. 12 [iii]Contents
1. Introduction ................................................................................................ 1
2. LTC systems throughout the world .................................................................... 2
3. Theoretical background ................................................................................ 3
4. Methods and data ....................................................................................... 6
4.1 Operationalization ...................................................................................... 7
4.1.1 Classified unit ........................................................................................... 7
4.1.2 Provision dimension ................................................................................... 9
4.1.3 Financing dimension ................................................................................ 10
4.1.4 Regulation dimension ............................................................................... 11
4.2 Data ....................................................................................................... 13
5. Classification results ................................................................................... 13
6. Discussion ................................................................................................ 15
7. Conclusion ............................................................................................... 17
References ......................................................................................................... 19
Appendix .......................................................................................................... 25
SOCIUM • SFB 1342 WorkingPapers No. 12 [v]1. Introduction sibility for LTC provision, financing and reg-
ulation (see e.g. Lyon & Glucksmann, 2008;
Ochiai, 2009; Rodrigues & Nies, 2013). In
In the last decades, long-term care (LTC) is analyzing the resultant ‘care-mix’ of LTC sys-
increasingly developing into a distinct social tems, we can, for instance, gain important
policy field. While the need for long-term insights into the role of the state and of public
assistance with daily living due to physical versus private actors. Furthermore, this focus
and/or mental impairments is not a novel sheds light on interaction logics present in
phenomenon per se, developments such as LTC systems and their associated outcomes
global demographic aging, changing family (Fischer, Frisina Doetter, & Rothgang, 2021;
structures and the emergence of a (human) Rothgang & Fischer, 2019).
rights perspective on disability and aging Adopting an actor-centered perspective,
(see e.g., Birtha, Rodrigues, Zólyomi, San- the present article compares distinct LTC sys-
du, & Schulmann, 2019) have contributed to tems throughout the world, identifying clus-
the recognition of LTC as a ‘new social risk’ ters or types of countries. We ask the follow-
necessitating public attention (Greve, 2018; ing question: How do distinct LTC systems
Österle & Rothgang, 2021). This develop- differ with respect to actor types dominant
ment is visible both on the inter- and transna- in service provision, financing and regula-
tional level – where international and region- tion? To systematically analyze the variation,
al organizations have increasingly come to we make use of a multi-dimensional, ac-
address LTC (e.g. Esquivel, 2017; European tor-centered typology of LTC systems recently
Commission [EC], 2013; Organisation for developed by Fischer et al. (2021). Typolo-
Economic Cooperation and Development gies constitute useful instruments for com-
[OECD], 2005; Scheil-Adlung, 2015; World parative research, helping to transparently
Health Organization [WHO], 2017) – and conceptualize categories for comparison
in individual countries worldwide. Concern- and sort complex empirical cases accord-
ing the latter, LTC is still a more salient topic ing to their similarities and differences. The
in the richer and older welfare states in the field of (country) comparative social policy
Global North, but is increasingly becoming has extensively engaged in identifying types
a field of political concern in countries and of welfare regimes and policies during the
regions in the Global South such as Latin last decades (see e.g. Lalioti, 2021; Powell,
America, China, and Southeast Asia as well Yörük, & Bargu, 2020) and classifications
(Loichinger & Pothisiri, 2018; Luo & Zhan, focusing on LTC in particular have also been
2018; Nieves Rico & Robles, 2019). put forward since the 1990s (see Section 3).
Irrespective of a growing, yet tentative The present article aims to add to this litera-
trend in LTC as a field of social protection, ture by taking a rigorous multi-dimensional
societies differ in the question of who takes approach towards classifying LTC systems
over responsibility for caring for LTC depen- as well as incorporating both a more global
dent people. This issue becomes of partic- and historical perspective by focusing on the
ular interest to social policy scholars once complete population of distinct LTC systems
care is no longer a mainly ‘private’ matter at the time point of system introduction.
and welfare states take over formal, legal The paper is structured as follows. In Sec-
obligations for LTC, establishing LTC systems tion 2, we briefly present the definition and
under public responsibility. With the (partial) empirical instances of what we have termed
‘socialization’ of LTC, different types of actors distinct LTC systems, which constitute our
such as the state, corporate societal-based population of subsequently classified cas-
organizations, commercial entities or fami- es. Section 3 provides the theoretical back-
lies can take over varying degrees of respon- ground of typological research in the field
SOCIUM • SFB 1342 WorkingPapers No. 12 [1]of LTC policy, with a particular focus on out- (ii) and the elements of the LTC system are
lining the multi-dimensions, actor centered some-what integrated, i.e. managed by one/
typology which we use as the classificatory several designated agencies (iii) (De Carval-
framework for comparing LTC systems. Sub- ho & Fischer, 2020, p. 13). Moreover, when-
sequently, the method of classification, op- ever the LTC system/policy differs between
erationalization of the typology’s dimensions age groups, we focus on LTC for the elderly
and data used are described in detail. We as the population group with highest levels
then move on to present and interpret the re- of care dependency (Colombo, Llena-Nozal,
sults of our classification in Section 5, while Mercier, & Tjadens, 2011; WHO, 2015, pp.
Section 6 continues to put them into per- 67–68).
spective with existing research and discusses When applying this definition, approxi-
limitations of our approach. Finally, we con- mately 50 countries worldwide have so far
clude by reflecting on the insights and further established public LTC systems (Fischer, Pol-
use of the typology. te, & Sternkopf, 2021; Fischer & Sternkopf,
forthcoming). However, some of these first
LTC related laws represent rather incipient
2. LTC systems throughout and rudimentary forms of LTC systems. While,
the world per definition, LTC benefits for at least some
share of the population have been formal-
ly introduced in all these cases, LTC benefits
LTC systems can be defined in different terms, may be granted as part of another welfare
for instance by stressing normative aspects state program as LTC is not (yet) conceived
of “appropriate, affordable, accessible” care of as specific social risk in its own right and
(WHO, 2017) or a focus on public funding a separate field of social policy making. In
(Spasova et al., 2018). The concept used in consequence, it is useful to distinguish yet
this article builds on an extensive discussion another form of systems to capture more in-
of health and LTC systems by De Carval- dependent and mature developments in the
ho and Fischer (2020). Accordingly, a LTC field. We therefore introduced the concept of
system can in general be described as the distinct LTC systems (under public responsi-
sum of provision, financing and regulatory bility) adding to the public system definition
arrangements in a society. In line with our outlined above the criterion of LTC being ac-
research focus on social policy and state re- knowledged as a distinct social risk that is in-
sponsibility, we limit our analytical focus by stitutionally treated as a social policy field of
studying LTC systems under public responsi- its own and has achieved a certain degree of
bility. These, in turn, can – according to a independence for other programs (cf. Fisch-
statutory, formal understanding – be seen to er et al., 2021). These more fullfledged sys-
exist in a country if country-wide legislation tems lend themselves much more to a com-
(i) establishes entitlements for LTC benefits prehensive comparative analysis than single
Figure 1.
Timeline of introducing distinct LTC systems
Netherlands Sweden Israel Germany South Korea Uruguay
Luxembourg
1967 1980 1986 1994 1998 2007 2015
1974 1982 1993 1997 2006 2014 2019
Denmark Finland Austria Australia Czech Republic United Singapore
Norway Japan Portugal Kingdom
Spain
Source: own illustration.
[2]LTC benefits integrated in different parts of (Austria, Czech Republic) and Southern Eu-
the health and/or social care systems. ropean (Spain, Portugal) cases as well as
Up to now, our research has identified Australia. In 2014 and 2019, respectively,
a population of 18 distinct LTC systems ex- the United Kingdom (UK)2 as the pioneer
isting worldwide.1 The timeline of adoption having introduced first elder care provisions
listing all countries is presented in Figure 1. in 1948, and its former colony Singapore
Accordingly, the first distinct LTC system was updated and unified their legal LTC-regu-
the introduction of the Algemene Wet Bijzon- lated frameworks, establishing distinct sys-
dere Ziektekosten (AWBZ, Exceptional Medi- tems. Furthermore, with Uruguay’s Sistema
cal Expenses Act) in the Netherlands in 1967 National Integrado de Cuidados (SNIC, Na-
(Companje, 2014), followed by the Den- tional System of Care), the first country from
mark, Finland, Norway and Sweden. While the American continent joined in recognizing
modern stateled development of institution- LTC as a distinct area for social protection in
al and home care services for the elderly in 2015 (Nieves Rico & Robles, 2019). In the
Scandinavian countries can even be dated remainder of the article, these 18 systems
back to the middle of the 20th century (Sipilä will be classified at the point of their respec-
et al., 2000), the incremental development tive introduction point.
of LTC policies seems to culminate in the
adoption of unifying, universal acts passed
in the 1970 and early 1980s, respectively. 3. Theoretical background
Subsequently, in the late 1980s, Israel es-
tablished a social insurance scheme dealing
specifically with the risk of LTC dependency The use of classifications to order and make
as the second country worldwide (H. Schmid, sense of our empirical world is by no means
2005), passing (to our knowledge) the first an exclusive characteristic of the social sci-
law which focused solely on the social pro- ences or sciences in general. It is, first and
tection for LTC (the previous introductions foremost, a fundamentally human and intrin-
all include other elements of social and/or sic aspect of cognition, which automatically
healthcare into their foundational laws as engages in the joint processes of compari-
well). In later years, only few countries have son and categorization (Freeman & Frisina,
chosen to follow this path of introducing dis- 2010). This regularly entails the grouping
tinct social LTC insurance schemes: Germa- together of similar types of a given category
ny in 1994, Japan in 1997, Luxembourg in or phenomenon to create typologies, which
1998 and South Korea in 2007 (Campbell, helps further reduce the cognitive workload
Ikegami, & Kwon, 2009; Companje, 2014). otherwise involved in the generation of al-
Furthermore, in the 1990s and 2000s, sev- ways new classificatory labels. Not only are
eral more countries which previously had typologies useful in grouping together in-
decentralized systems or single, non-dis- stances bearing a shared set of attributes,
tinct programs, introduced distinct LTC sys- they also facilitate the drawing of expecta-
tems. Among them were Central European tions related to those attributes. They are
therefore a highly useful tool in comparative
research.
1 For a detailed description of the procedure and
data sources used for identifying system introduc-
tions, see Fischer and Sternkopf (forthcoming).
The introduction dates (both date of adoption and 2 More specifically, we refer here to the Care Act
dejure implementation as well as a brief descrip- regulating LTC in England, the largest nation of
tion of the system and a justification for counting the UK (see also Section 4). However, both Scot-
the case as a distinct LTC system are provided in land and Wales also passed novel LTC acts in
the country data tables in the Appendix. 2013 and 2014 Snell (2015), respectively.
SOCIUM • SFB 1342 WorkingPapers No. 12 [3]While no shortage of critical attention also have an explicit focus on comparing
on (specific) typologies exists (see e.g. Arts public schemes specifically (e.g. Colom-
& Gelissen, 2010; Collier, Laporte, & Sea- bo et al., 2011; Joshua, 2017; Pacolet et
wright, 2012), a number of well-constructed al., 1999; Rothgang, 2009). To our knowl-
classificatory systems have come to domi- edge, all countries whose LTC systems have
nate the field of comparative social policy, been included in published typologies so
not least of all that of Esping-Andersen’s far are situated in Europe and/or are mem-
(1990) seminal welfare state regimes. Typol- ber states of the Organisation of Economic
ogies are particularly abundant in the study Cooperation and Development (OECD). A
of healthcare systems – a field of scholarship diverse set of criteria is used in extant typo-
spanning roughly six decades since Roemer’s logical research for sorting empirical cases.
classification of health departments and Most commonly, LTC financing is addressed,
medical care in the 1960s (cf. Ariaans, Lin- followed by aspects of coverage and regula-
den, & Wendt, 2021; De Carvalho, Schmid, tion, service provision and the integration of
& Fischer, 2020). Despite its relative infancy schemes/systems. Among the most frequent-
as a policy field, since the mid-1990s LTC ly used criteria is the distinction between tax
has also seen the emergence of classificatory and contribution based-financing schemes
work. Most notably, in the research of Ant- (e.g. Colombo et al., 2011, Pacolet et al.,
tonen and Sipilä (1996) and Bettio and Plan- 1999; Simonazzi, 2008), population cover-
tenga (2004) that takes a comprehensive age (Colombo et al., 2011; Kraus, Riedel,
(social) care perspective to LTC, integrating Mot, Willemé, & Röhrling, 2010; Ranci &
both child and elder care arrangements into Pavolini, 2013a), and the prominence of for-
one framework. This approach has its merits mal vs. informal care (Roit & Le Bihan, 2010;
and is particularly useful for broad and gen- Kraus et al., 2010; Nies et al., 2013).
dered understandings of the welfare state. Taken together, existing classificatory ap-
However, it falls short in capturing key dif- proaches have strongly contributed to the
ferences in the nature of benefits and degree conceptual and empirical understanding of
of familialism distinguishing the two policy the variety of LTC systems. As established in a
fields in many countries. review of 17 classifications (see Fischer et al.,
Not until the work of Pacolet, Bouten, Hil- 2021), however, these typologies are subject
de Lanoye, and Versieck (1999) and Timo- to number of important limitations. First, the
nen (2005) did typologies with an exclusive specification of criteria and/or underlying
analytical focus on LTC start to populate the procedure/methods for typology construc-
field of comparative social policy. Since then, tion is not always clear; second, the appli-
several typologies that have sorted countries cability of classifications to regions beyond
according to their LTC arrangements, both Europe is hardly discussed; third, they show
with and without an agerelated focus. More a paucity of information on the multi-dimen-
recently, multiple quantitatively-derived clas- sional aspects of LTC systems.
sifications of LTC systems using clustering Bearing these issues in mind, Fischer et al.
methods and standardized data have been (2021) put forth a deductively derived, ac-
put forward as well (Ariaans et al., 2021; Da- tor-centered typology that incorporates three
miani et al., 2011; Halásková et al., 2017; dimensions of the LTC system that have also
Kraus et al., 2010), adding yet another layer been used in healthcare typologies (Böhm,
to the typological study of LTC systems. Schmid, Götze, Landwehr, & Rothgang,
While many studies classify whole coun- 2013; Wendt, Frisina, & Rothgang, 2009):
tries’ LTC regimes (e.g. Halásková, Bednář, The first, service provision, refers to the most
& Halásková, 2017; Nies, Leichsenring, & elementary function of the system involving
Mak, 2013; Ranci & Pavolini, 2013a), some the actual task of caring. Care can consist
[4]of medically-related tasks, such as adminis- dimension they mostly take the form of social
tering medicines and maintaining hygiene, insurance bodies. Societal actors (self-)reg-
household-related tasks such as washing or ulate mainly through collective negotiations
cooking, as well as strengthening societal (Rothgang et al., 2010, p. 14).
participation and providing emotional sup- Moving on to private actors, thirdly, there
port. The second dimension, financing, re- are private for-profit actors, e.g. nursing
fers to the resources necessary for ‘produc- homes or home care services, which can de-
ing’ care, either in the form of monetary re- liver care, and financing agencies in the form
sources or, in case of informal, unpaid care of private insurances collecting premiums. It
provision, through time and foregone earn- is important to note that private for-profit ac-
ings (WHO, 2015, p. 131). Finally, the third tors in the provision dimension comprise a
dimension, regulation, that is the “interven- spectrum of providers reaching from domes-
tion in the behavior or activities of individual tic care workers, which often work (and live)
and/or corporate actors” (Koop & Lodge, in the care recipient’s household to large
2017, p. 97), influences and modifies the formalized corporations. Fourthly, private in-
production structure of care and crucially dividual actors, defined as persons from the
shapes the system (Mayntz & Scharpf, 1995). care recipient’s network, i.e. family mem-
For each of these dimensions different cri- bers, neighbors or friends (Timonen, 2009),
teria can be analyzed. We concentrate on the are crucial in many LTC systems in providing
question of who bears responsibility because (informal) care. Through out-of-pocket pay-
this is one crucial category in the analysis of ments, care-recipients and their relatives are
care and social policy, providing insights, for also an important financing source, even in
instance, into redistribution processes, legiti- LTC systems under public responsibility (e.g.
macy, social structures, and norms. To some Colombo et al., 2011; Rodrigues & Nies,
extent, this focus can also inform us about 2013). It is important to note that while both
how and what happens within each dimen- forms of private actors have limited means
sion, especially the associated interaction by setting general, external standards, they
logics (Fischer, Frisina Doetter, & Rothgang, can (self-)regulate (Rothgang et al., 2010;
2021; Rothgang & Fischer, 2019). In a sec- see also Black, 2001; Braithwaite, Makkai,
ond step, therefore, Fischer et al.’s (2021) & Braithwaite, 2007). In the regulation di-
LTC typology conceptualizes (up to) five types mension, we capture this mode of regulation
of (quasi-)corporate actors which take over by private actors jointly. Lastly, global actors
responsibility for provision, financing and/ such as foreign state, international govern-
or regulation of the LTC system: State, soci- mental or non-governmental organizations
etal actors, private for-profit actors, private might be involved in LTC systems in any of
individual actors and global actors. Firstly, the three dimensions. However, this is not
the state is defined as the public institutions the case for the population of distinct LTC
in the political-administrative system of a systems under public responsibility analyzed
country (Johnson, 1999), comprising differ- in this article, which is why we abstain from
ent – central, regional, local – state levels discussing this actor group further.
and as such is a relevant actor in all three Fischer et al.´s (2021) typology endeav-
dimensions. Secondly, societal actors are ors to deliver a widely applicable classifica-
characterized by their formal, non-profit, tory framework to identify the role of specific
non-governmental status and collective sel- actors across the multi-dimensional universe
forganization (Johnson, 1999; Wendt et al., of the LTC system. It is an ambitious response
2009). Societal actors appear as providers, to the aforementioned limitations of extant
for example in the form of charitable or mu- typological approaches – one which re-
tual aid organizations, while in the financing sults in a total of 100 LTC system types (see
SOCIUM • SFB 1342 WorkingPapers No. 12 [5]Figure 2.
Typological attribute space of the multi-dimensional, actor-centered typology
P R OV IS I O N
REGULATION FINANCING Private for-profit Private individu- Global
State Societal actors
actors al actors actors
State Type 1 Type 2 Type 3 Type 4 Type 5
Societal actors Type 6 Type 7 Type 8 Type 9 Type 10
State Private for-profit actors Type 11 Type 12 Type 13 Type 14 Type 15
Private individual actors Type 16 Type 17 Type 18 Type 19 Type 20
Global actors Type 21 Type 22 Type 23 Type 24 Type 25
State Type 26 Type 27 Type 28 Type 29 Type 30
Societal actors Type 31 Type 32 Type 33 Type 34 Type 35
Societal actors Private for-profit actors Type 36 Type 37 Type 38 Type 39 Type 40
Private individual actors Type 41 Type 42 Type 43 Type 44 Type 45
Global actors Type 46 Type 47 Type 48 Type 49 Type 50
State Type 51 Type 52 Type 53 Type 54 Type 55
Societal actors Type 56 Type 57 Type 58 Type 59 Type 60
Private actors Private for-profit actors Type 61 Type 62 Type 63 Type 64 Type 65
Private individual actors Type 66 Type 67 Type 68 Type 69 Type 70
Global actors Type 71 Type 72 Type 73 Type 74 Type 75
State Type 76 Type 77 Type 78 Type 79 Type 80
Societal actors Type 81 Type 82 Type 83 Type 84 Type 85
Global actors Private for-profit actors Type 86 Type 87 Type 88 Type 89 Type 90
Private individual actors Type 91 Type 92 Type 93 Type 94 Type 95
Global actors Type 96 Type 97 Type 98 Type 99 Type 100
*Note: Bold highlighted types are pure types with one dominant actor only; grey highlighted types are presumably unlikely/implausible.
Source: Fischer et al., 2021.
Figure 2). Of these, five emerge as ‘pure’ 4. Methods and data
types consisting of one actor dominating all
three dimensions. As outlined above, the typology we use for
Thus far, this typology has yet to be ap- classifying countries’ LTC systems in this ar-
plied with empirical rigor to verify its applica- ticle consists of predefined types created by
bility and utility as a classificatory framework intersecting the three dimensions and five/
for LTC systems worldwide. The present con- four actor types systematically (see Figure
tribution sets out to do just that, traversing 2). Consequently, each of the resulting types
the globe for empirical instances of distinct can be depicted as a configuration, that is
LTC systems and classifying them in line with as a combination of its properties which
Fischer et al.’s typology. together define the type as a whole (Kvist,
2006). Similarly, each empirical case of a
LTC system can be conceived of as a config-
uration of attributes in different dimensions
[6](Rihoux & Ragin, 2009; Wagemann, 2015). ployed for classifying health care systems by
Following this logic, we can classify an em- Böhm et al. (2013) – is exactly the approach
pirical case – that is, put it into a ‘cell’ – by we follow. Our sorting is based on the logic
identifying which type’s configuration has the that the homogeneity of both cases – one
highest overlap with the properties of a case. with 100 % and one with 60 % societal ac-
This can be done most easily when regard- tor based LTC provision – is higher than with
ing each dimension – service provision, fi- other cases where there is no or a minor role
nancing and regulation – separately during of societal actors in care provision (cf. Kelle &
the initial stage of the classification process. Kluge, 2010, pp. 100–101). Therefore, the
However, it is important to note that, as classification process marks these two cas-
LTC systems are very complex, cases do of- es as similar by assigning them to the same
ten not completely conform to any type. That type. Subsequently, when the dominant actor
is, adherence of real cases to the deductive- in each dimension has been determined for
ly constructed types of the typology can be a certain case, the country is classified ac-
stronger or weaker (Kvist, 2006; Schneider cording to the resulting configurational set-
& Wagemann, 2012, pp. 97–98). For in- ting and assigned to the respectively type in
stance, if care in a LTC system is exclusive- typology matrix. In the remainder of this sec-
ly provided by societal actors, the country tion we discuss the operationalization of the
strongly confirms to the ‘extreme’ expression coding process (Section 4.1) and the data
in the provision dimension; if there is a mix of basis for classifying (Section 4.2) in some
providing actors with societal actors making detail.
up the majority but not as the sole actor type
(e.g. a mix of 60 % societal actors, 30 % pri- 4.1 Operationalization
vate for-profit actors, and 10 % by state-run
facilities), societal actors are still dominant
in the provision dimension but to a small- For each of the three dimensions – regula-
er degree. While both of these exemplary tion, financing, and service provision – op-
cases differ to some extent, they can still be erationalization rules have to be determined
assigned unambiguously to a cell in Figure (Section 4.1.2 to 4.1.4). Before diving into
2, indicating that societal actors dominate this, however, we have to clarify what consti-
the provision dimension. It should be noted tutes a case in the subsequent analysis (Sec-
that any classification of metric data, as e.g. tion 4.1.1).
the share of financing that different actors
provide, leads to a loss of information. As a 4.1.1 Classified unit
consequence, even small changes may lead
to a reclassification of a system, if the metric The aim of this article is to systematical-
value is close to the threshold. The classifi- ly compare and, hence fore, classify cases
cation of a system is, therefore, not a suffi- of distinct LTC systems in various countries.
cient substitute for an in-depth study of the But what constitutes a ‘case’ in our study?
respective case, but is suitable for providing In general, a case can be described as “an
an overview on how cases compare to each instance of a class of events” (George &
other. Bennett, 2005, p. 17), with the event being
The above route of classifying cases by defined by spatial, topical, and/or temporal
identifying the dominant actor type per di- boundaries (Bennett & Checkel, 2015). First-
mensions3 – which has previously been em- ly, regarding the spatial confinements, cases
3 In some cases, only a relative dominance, i.e. be- can be identified. If this was the case, it is noted
ing the strongest actor but below a share of 50 % in the data table in the appendix.
SOCIUM • SFB 1342 WorkingPapers No. 12 [7]are equated with countries, meaning that the dominant in different schemes. If this was the
LTC system needs to be institutionalized by case, we took the major LTC scheme for iden-
nationwide legislation and be applicable – tifying dominant actors only. For instance,
albeit with potential regional modifications with the distinct LTC system introduction in
– to the whole country’s territory.4 Secondly, Germany in 1994, both a social LTC insur-
the topical focus is on classifying LTC sys- ance (LTCI) scheme and a mandatory pri-
tems. We define long-term care as being vate LTC insurance schemes were introduced
“concerned with a range of services and as- (Rothgang, 2010). As the social LTCI at that
sistance provided to care dependent persons time (and also later on) covered approxi-
who need support with daily living activities mately 90 % of the population (Rothgang,
over an extended time period due to physical 2009b), we have chosen to use this scheme
and/or mental impairments” (De Carvalho & for classifying Germany in the regulatory di-
Fischer, 2020, p. 8). The concept of a LTC mension. For countries where the regulatory
system, refers to the provision, financing and dimensions are based on parts of the overall
regulatory arrangements in a society deal- public LTC system only, this is documented
ing specifically with LTC as an area of social in the Appendix (row ‘Dominant scheme for
protection for (at least) (parts of) the elderly classification (if applicable)’). Similarly, for
population. some countries statistical data on financing
If a system does not cover the whole and service provision is only available for the
country, we need a further specification. On country level, but not for the public LTC sys-
the one hand, the LTC arrangement of the tem. This is, for instance, the case with data
whole country can be classified, including following the System of Health Accounts
both the public scheme(s) and all other (e.g. (SHA) standard (OECD, WHO, & Eurostat,
privately paid, informally provided) LTC. On 2011), the most important internationally
the other hand, the analysis can be limited comparative data on financing shares. Using
to the LTC system under public responsibility such data for classification can be regard-
(see Section 2). Conceptually, we follow the ed as a conservative estimate of public LTC
latter approach, not least as only systems un- system financing shares because typically the
der public responsibility may guarantee ac- share of private financing and service provi-
cess to care for the whole population, which sion in the system under public responsibility
is crucial from a human rights perspective. is lower than in the rest of the country’s LTC
Nevertheless, due to data availability in provision.
some cases we have to use countrywide data Besides the spatial and topical definition
instead. In countries with more than one LTC of cases, the temporal boundary is also im-
scheme simultaneous focus on the whole portant. Temporally, we focus on the intro-
public arrangement is sometimes not feasi- duction point of each distinct LTC system.
ble, especially when analyzing the regulation Empirical data about provider and financ-
dimension where diverging actors might be ing shares in particular are only telling after
the system has been implemented. There-
fore, we have used, if available, data for
4 One partial exception is the United Kingdom (UK), the time span of (approximately) three years
where, since the inception of the devolution pro-
after the de jure implementation of the law
cess in 1999, policies for social care/long-term
care are (partly) the political competence of the to stay both close to the introduction date
individual nations, i.e. England, Scotland, Wales and the (initial) design of the introduced sys-
and Northern Ireland, respectively (Bell, 2010; tem. However, there are also cases where
Glendinning, 2013). Therefore, the current legal the dominant actor type has switched within
acts do not necessarily cover the United Kingdom
the first years after system introduction, for
as a whole. Whenever necessary we focus on En-
gland as the by far largest part of the country. instance in the financing dimension in the
[8]Netherlands (state −> societal actors) or the well as regulatory competencies do often
provision dimension in Israel (societal actors differ between both settings. Residential care
−> private for-profit actors). In these cases, is provided continuously around the clock
we have taken the initially dominant actor for care dependent persons living jointly in
to characterize the system at its introduction a specific institution, for example a nursing
point.5 In short, a case in this study can be home or assisted living facility (Rothgang &
described as the complete distinct LTC system Fischer, 2019; WHO, 2015, p. 129). In con-
under public responsibility within a country at trast, the terms ‘home and community care’
the point of its introduction. or ‘community-based care’ summarize “all
forms of care that do not require an older
4.1.2 Provision dimension person to reside permanently in an institu-
tional care setting” (WHO, 2015, p. 129). It
LTC provision as one of our dimensions for comprises both assistance with personal care
classifying systems can take the form of for- and household activities in the care recipi-
mal care, i.e. paid, (semi-)professional care ent’s home as well as facilities like day care
provided in an organized setting, and infor- centers (Timonen, 2008, p. 142).
mal care, which is provided in unregulated In order to determine the dominant actor
‘private’ settings, often by family members, type in service provision we follow a three-
or fall between the poles of this ideal-typ- step approach (see Figure 3). First, we re-
ical formal-informal distinction (Pfau-Effin- cord the share of the three main LTC benefit
ger & Rostgaard, 2011; Timonen, 2009; types i.e., in-kind residential care services,
WHO, 2015, pp. 129–130). The form of in-kind home and community care services,
care crucially depends on the benefits avail- and monetary benefits. In doing so, we use,
able within the LTC system: While benefits in where possible, data on their respective pro-
the form of in-kind services generally trans- portion in the overall care-mix based on the
late into formal care provision conducted number of care recipients under each benefit
by state, societal or private for-profit actors, type.6 Second, the shares of actor types are
monetary transfers in the form of vouchers recorded for each relevant benefit type sepa-
or cash benefits can – often depending also rately. In the case of (unregulated) monetary
on the regulation for their use – result in a benefits there is often no data available on
spectrum between informal and formal care where the money goes. In conjunction with
arrangements provided (e.g. Da Roit & Le Bi- evidence from secondary literature, howev-
han, 2010; Le Bihan, Da Roit, & Sopadzhi- er, we can normally assume that most un-
yan, 2019). In the population of distinct LTC regulated cash benefits translate into care
systems classified in this paper, there is only provision by private individual actors, that is
one country offering exclusively cash benefits mostly family members and/or domestic care
(Singapore), while most countries offer only workers (e.g. Da Roit & Le Bihan, 2010; Rie-
in-kind benefits or a combination of in-kind del & Kraus, 2016). If we have evidence that
services and cash benefits. Furthermore, in cash benefits are used to finance live-ins, i.e.
the category of in-kind benefits/formal care, mostly migrants living in the household of a
it is important to distinguish between resi- care-dependent person in order to assist him
dential/institutional care versus home and or her, we subsume this arrangement under
community care as providing actor types as
6 While there are other measures such as expen-
diture, or, for formal care, granted hours of care
5 In the appendix, the political adoption date of the or number of employees in each sector, which
law as well as the de jure implementation date at could be used alternatively, data on the number
which the law formally enters into force are spec- of recipients is most often available and counts all
ified for each country. care recipients equally.
SOCIUM • SFB 1342 WorkingPapers No. 12 [9]Figure 3.
Calculation of dominant actor in the provision dimension
Source: own illustration.
private for-profit care-giving. Third, with the both funding from social and/or private in-
information from step two and three, we cal- surance depending on the concrete design
culate (if necessary) the total share of each of the scheme. This is further discussed be-
providing actor type in the whole LTC system. low. For 12 out of 18 countries, respective
data for a year close to the LTC system intro-
4.1.3 Financing dimension duction point can be found in the OECD sta-
tistics.7 Even though the statistics refer to the
The operationalization of the financing di- whole country and not the public LTC system
mension is mostly straight forward as we can only, they provide – especially if triangulated
equate financing sources with actor types (cf. with national sources and case descriptions
Böhm et al., 2013). Generally, four types – valuable standardized and comparable in-
of domestic financing sources which corre- formation on dominant financing schemes.
spond to the four domestic actor types out- However, it has to be noted that the cor-
lined in the typology (Fischer et al., 2021) relation between private for-profit actors and
can be distinguished: Tax revenues (state), the SHA classification poses some problems.
social insurance contributions (societal ac- In general, private insurance schemes can
tors), private insurance premiums (private take a compulsory or voluntary form, which
for-profit actors), and household out-of- comes with different implications regarding
pocket expenditure (OOP) (private individual the role of the state and the social protec-
actors) (Rothgang & Fischer, 2019). To reap tion of the schemes (OECD et al., 2011).
the benefit of using comparable data across Therefore, the SHA methodology (OECD
countries, whenever possible we relied on et al., 2011) classifies compulsory private
SHA-based (see OECD et al., 2011) inter- insurance (HF.1.2.2) and voluntary health
national comparative data from the health insurance schemes (HF.2.1) in two different
expenditure and financing database provid- categories which stresses the – undisputable
ed by the OECD. When doing so, we used strong – relation of the former with social in-
the following SHA categories to determine surance schemes. However, as in our analyt-
actor shares: Government schemes (HF.1.1) ical framework regulation is considered also
for state financing, social health insurance separately from financing we maintain that
(HF.1.2.1) for societal actors, voluntary both mandatory and voluntary private insur-
health care payment schemes (HF.2) for pri-
vate for-profit actors, and household out-of- 7 Respective data were unavailable for Israel, the
pocket payments (HF.3) for private individual Netherlands, Norway and Sweden (due to later
actors. The categories compulsory contribu- start of the time series) and the non-OECD mem-
bers Singapore and Uruguay. For these cases
tory health insurance (HF.1.2) and compul-
national data and secondary sources were used
sory private insurance (HF.1.2.2) can contain instead
[10]ance provide – differently regulated – hints Figure 4.
on the relevance of private for-profit actors in Regulatory relationships and objects
the financing dimension. This is why we clas- in the LTC system
sify financing in Singapore – the only classi-
fied case where actuarial private insurance
premiums are a major financing source – as
dominated by private for-profit actors even
though the scheme is (partly) mandatory.
4.1.4 Regulation dimension
Regulation is a particularly broad category.
Regarding the field of LTC, for instance the
available ‘benefit package’, quality, care Source: own illustration based on Rothgang et al., 2010.
providers’ standards or the extent of choice
of care recipients can be centrally regulated To get a more detailed picture of what
– e.g. by the state or an (social) insurance goes on in the regulation dimension of each
body – or left to selfregulation of involved case, we use a differentiated approach of
actors (e.g. Braithwaite et al., 2007; Da categorizing several regulatory subdimen-
Roit & Le Bihan, 2010; Kraus et al., 2010; sions or relations and combine the informa-
Murakami & Colombo, 2013; Rothgang tion to arrive at a final classification. In doing
& Fischer, 2019). Consequently, there are so, we draw on earlier works on health care
multiple ways of operationalizing the regu- systems which have conceptualized regulato-
lation dimension of the typology. The easi- ry relations and objects (Böhm et al., 2013,
est – and probably most limited – possibility 2012; Wendt et al., 2009), adapting them
is to record who – that is which organiza- to the context of LTC systems. The concep-
tion/agency, and, based on this, which actor tualization departs from the point that in any
type – is generally responsible for regulat- healthcare/LTC system there are three groups
ing the LTC system without formally consid- of actors involved which form a triangular re-
ering any specific aspects of regulation.8 In lationship: care providers, financing bodies,
LTC systems under public responsibility, this and (potential) care recipients (Rothgang et
will generally be the state directly, in the al., 2010, p. 11). The content of these re-
form of the central government, provinces/ lationships – visualized as the sides of the
federal states and/or municipalities or other triangle in Figure 4 – can be regulated.
public bodies such as LTC or health insur- For the relation between financing bod-
ance funds. Therefore, this form of opera- ies and (potential) care recipients (side A
tionalization automatically limits the kinds in Figure 4), there are two main objectives:
of actors which can achieve dominance in The entitlement/eligibility (1) describes which
the regulation dimension to public actors, (potential) care recipients have access to LTC
i.e. state or societal actors. Furthermore, benefits. Following the ‘who’ question, here,
identifying the main regulator by looking at we can either focus on who decides the enti-
the generally responsible organization(s) in tlement and eligibility criteria (e.g. citizenship
the LTC system is a quite crude way of mea- status, formal employment, dependency lev-
surement. els, age thresholds) defining inclusion (1a),
or ask which actor is responsible for execut-
ing eligibility assessment procedures (mostly
8 For each case, this information is recorded in the
care dependency assessment) (1b). As noted
data table in the appendix (row ‘dominant actor
agency’). by Böhm et al. (2013) for healthcare systems
SOCIUM • SFB 1342 WorkingPapers No. 12 [11]already, there is no variation in the former or determined by providers themselves (for
point (1a); in all (studied) systems defining instance in a so-called “Pork Barrel Market”
entitlement/eligibility criteria is exclusively as termed by Gingrich (2011)).
the responsibility of the state (see data ta- Lastly, there is relation C connecting (po-
bles in Appendix). Therefore, we exclude this tential) care recipients and care providers.
category for classifying and focus solely on This relationship is, on the one hand, about
question 1b, that is who is responsible for looking at the regulation of care recipients to
eligibility assessment? In this relation we also choose a concrete provider (5), that is who
look at a second question: Who decides if decides which provider will deliver care to the
and how much to pay/contribute to the sys- benefit recipient? If beneficiaries can choose
tem (2)? The question can be applied both a provider themselves, the category is classi-
to co-payments – i.e. who decides if and fied as private (individual), if care managers
what sum of co-payments the care recipient (or similar) take over the decision depending
needs to pay – and/or contribution or premi- on who employs the care manager the cate-
um rates – i.e. who decided if contributions/ gory is classified as state, societal or private
premiums need to be payed and what their (collective). On the other hand, the decision
level is. Interestingly, this regulatory relation which care benefits – that is, in-kind residen-
is also strongly– albeit not exclusively – pop- tial or home/community care or cash ben-
ulated – by the state. efits (see above) – a care recipient gets can
Moving on to the relationship between fi- also be decided by different actor types (6).
nancing bodies and care providers (side B There are two steps to consider here: Firstly,
of the triangle), the access of providers to if there is only one benefit type offered by
the public LTC system (3) and the system of law, the state regulates the choice of benefits.
remuneration of providers (4) are relevant Secondly, if there are several benefit types on
here. It is important to note that in systems offer (e.g. residential care and home care),
with cash transfers and in-kind services, for- the care recipient might be free to choose
mal as well as informal providers might need (‘private’), or care managers (or similar) em-
to be considered (separately). Regarding the ployed by other actor types might determine
provider access, we are looking at who de- the benefit for each care recipient.
fines if and under which conditions providers Summing up, based on previous concep-
can offer services in the public LTC system. tualizations of health care system regulation
While in most countries there is the necessity we have identified six relevant regulatory
to get a general license to operate a care categories which we used for classifying the
facility/service, we are specifically interested regulation dimension of the LTC systems. In
in who controls provider access to provide doing so, we have adhered to the following
publicly regulated/financed benefits. If there rules/steps:
is no specific entry requirement, provider ac-
cess is classified as ‘private’, otherwise as 1. If necessary, the principal LTC scheme in
‘state’ or ‘societal’ depending on the (domi- the country for classifying regulation is de-
nant) regulator. Furthermore, concerning the fined (see above).
remuneration, we ask who decides or nego-
tiates the payments/fees provider receive for 2. Data for each of the six regulatory sub-
offering (certain) care services? Remunera- dimensions (1-6) is collected. In case the-
tion levels can for instance be determined by re are regulatory differences for several
the state – which is mostly the case for the benefit types (e.g. for residential care and
level of cash benefits, but sometimes also for cash benefits), if possible information on
formal in-kind care provision –, negotiated both is recorded.
between providers and financing agencies
[12]3. Additionally, the organization/agency ing). All data sources used per (sub-)dimen-
which is generally responsible for regula- sion and country are specified a country data
ting the LTC system is recorded (7). tables in the appendix, with a reference list
provider for each country separately below
4. Based on the raw data, the 1-2 dominant- each country data table. For reporting reli-
ly involved actors in regulating the respec- ability of the data/results, for the actor clas-
tive sub-dimension (1-7) are identified. If sification in each (sub-)dimension, the confi-
two actors are identified and data allows dence in the data/actor rating was recorded
for it, one actor is marked as most do- following a three-point scale: High confi-
minant (in bold letters). If the two actors dence is achieved if the data is confirmed ei-
derive from the fact that benefit types are ther by a law or reliable primary data source
regulated differently, the dominant bene- (e.g. official statistics) directly or by at least
fit type according to the data collected two independent secondary sources and
for the provision dimension is marked as retrieved information is non-contradictory.
most dominant. Results are rated with medium confidence if
there is only one reliable secondary source
5. Based on the actors identified for each of providing the necessary information or there
the six relations, the overall dominant ac- is some ambiguity/unclarity about domi-
tor is determined. Each of our six relations nant actors from the available information.
is weighted equally. If there is one actor All data that were extremely ambiguous or
in a sub-dimension, this counts with a va- uncertain, or based on sources that are not
lue of 1. If a sub-dimension is populated deemed reliable by the researcher, are rated
by two actors, each of them count with as low confidence. Overall, due to lack of
a value of 0.5. The actor type achieving data in two cases, i.e. Luxembourg and Sin-
the highest value is rated as the dominant gapore, it was not possible to determine one
actor type. single dominant provider type. In these cas-
es, we resorted to combining two actor types
6. In case two actor types are equally strong in the provision dimensions to classify these
according to step (5), sub-dimension 7 cases. Furthermore, data for the regulatory
capturing the general regulatory agency sub-dimension of benefit choice in Portugal
is used as a tie-breaker. was missing.
4.2 Data
5. Classification results
Multiple data sources were used for identify-
ing dominant actors in each dimension: the Figure 5 shows the results of the classifica-
laws introducing the LTC system, academic tion exercise: The 18 countries with a distinct
publications and reports and grey literature, LTC system can be classified into altogeth-
both on single countries or with a compar- er 8 types. When introduced, the systems
ative focus, statistics (as a primary statistical of the Nordic countries Denmark, Finland,
source mostly the OECD health expenditure Norway and Sweden fell under Type 1, rep-
and financing database as outlined above), resenting state-domination in regulation, fi-
national online newspaper articles and (of- nancing and provision. Eight other countries
ficial) websites about the LTC schemes, as also show state-domination in regulation
well as primary data collected through the and financing, however, with service provi-
project’s Expert Survey on Long-Term Care in sion dominated by societal actors (Australia,
2020/21 (see Fischer & Sternkopf, forthcom- Netherlands, and Portugal, Type 2), private
SOCIUM • SFB 1342 WorkingPapers No. 12 [13]Figure 5.
Multi-dimensional actor-centered distinct LTC system classification
Source: own illustration based on data sources and dimension-specific classification results specified in the Appendix. Cluster A is high-
lighted red; Cluster B green; and Cluster C blue.
for-profit actors (Spain, United Kingdom, Finally, a third cluster (Cluster C, high-
Uruguay, Type 3), and private individual ac- lighted in blue, see Figure 5) can be iden-
tors (Austria, Czech Republic, Type 4) respec- tified with dominant regulation by private
tively. At the point of introduction, the LTC actors, societal financing and care provision
systems of 12 out of the 18 countries classi- through private actors (South Korea and
fied thus belonged to a cluster with predom- Germany, Type 58 and 59). The dominance
inant state regulation and financing (Cluster of private actors in the regulation dimension
A, highlighted in red, see Figure 5). comes as a surprise: In an initial theoretical
A second cluster combing state regulation assessment of the plausibility of types, private
with different actors dominating financing regulation paired with societal financing was
and care provision (Cluster B, highlighted in deemed as implausible following the ‘hierar-
green, see Figure 5). can be found in an- chy rule’ hypothesized by Böhm et al. (2013)
other four countries. While the combina- (see Figure 2).
tion of societal financing and societal and These results are remarkable as stateled
private for-profit provision is populated by systems with state regulation and financing
Japan, Luxembourg and Israel (Type 7/8), are by far the most common, comprising
state regulation, private for-profit financing two thirds of all systems under scrutiny, while
and private (for-profit and individual) care there is no counterpart to this in form of so-
provision can be found in Singapore (Type cietal-dominated systems as can be found in
13/14). While both state financing (eleven the field of healthcare.
countries) and societal financing (five coun- Although in the Netherlands, Israel, Ger-
tries) are quite common, Singapore occupies many, Luxembourg, Japan, and South Korea
a unique position among the classified LTC social insurance systems were introduced,
systems being the only country with a domi- they don’t appear as such in Figure 5. While
nance in private (for-profit) financing. financing – as the central definition criteri-
on of a social insurance system – is indeed
[14]You can also read
