EUROPEAN UNIVERSITY INSTITUTE, FLORENCE
DEPARTMENT O F POLITICAL AND SOCIAL SCIENCESER No. 87/281
FORMATION REGIMES MY APPROACH
Gosta ESPDMG-ANDERSEN
BADIA FIESOLANA, SAN DOMENICO (F I)
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
© Gosta Esping-Andersen
Printed in Italy in April 1987
European University Institute
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
S T A T E A N D M A R K E T IN T H E F O R M A T I O N O F S O C I A L S E C U R I T Y R E G I M E S A P o l i t i c a l E c o n o m y A p p r o a c h b y G o s t a E s p i n g - A n d e r s e n © The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
INTRODUCTION.
The Political Economy of the Welfare State does not refer to a
theory as much as to an analytical focus on the interaction of
polity and economy. Much, if not most, of welfare state research
has concentrated solely on public sector social policy. As a
consequence, it has helped uphold the artificial duality of state
and economy that has patterned 20th Century thought. Our aim is
to decompose this partition; markets cannot be understood without states, and vice-versa.
Political economy has come to mean two opposing approaches. One
is the attempt to understand political phenomena through the
analytical foundations of modern economics, micro-economics in
particular. Its counterpart is the investigation of economic
phenomena with the employment of sociological or politological
conceptual apparata. This study is of the latter kind. My aim is
not to add yet another political theory of the economy to the
many that exist, but to explore how, in the first place, the
kinds of division of responsibilities between state and market we find today came to be.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
Neither state nor market as such were predestined to become the dominant loci of welfare provision- In fact, my argument is that
states created marketsjmarkets created states. The study that
will serve to explore the argument centers on the evolution of
pensions in the presently advanced industrial democracies. If
pensions may seem too narrow or obscure a phenomenon of
application, one should keep in mind two circumstances: one,
pensions account for more than 10 percent of GDP in many
contemporary nations; two, pensions constitute a central link
between work and leisure, between earned income and
redistribution, between individualism and solidarity, between the
cash-nexus and social rights. Pensions, therefore, help
illucidate a set of perennially conflictual principles of
capitali sm.
Most comparative research on welfare states rests on linear
conceptions of the world: some spend more, some less, but they
are otherwise basically the same welfare states. A greater
sensitivity to institutional variability will, in contrast, yield
a picture of distinct regime clusters. It is to this issue that
we first turn. Following an overview of the principal
institutional variations in the provision of pensions today, the
study will analyse the historical origins and developmental
patterns of private sector provision across the western world.The
study will then turn to an analytic effort to explain observed
international diversity and convergence.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
1. THE ROLE OF THE PRIVATE SECTOR IN WELFARE STATE REGIMES.
The concept of welfare state regimes serves to denote the
institutional parameters that guide and shape concurrent social
policy decisions, expenditure developments, problem definitions,
and even the response and demand structure of citizens and
welfare consumers. The existence of policy regimes reflects the
circumstance that short-term policies, reforms, debates and
decision making takes place within a framework of historical
institutionalization. The boundaries of rights and claims that are attached to social citizenship constitute an example of such , relatively historically stable institutional parameters. Thus,
the scope of human needs that are given the status of a social
right is a central definitional issue with regard to the
identi ficat ion of welfare state regimes. An " i nst it ut i onal "
welfare state i n Titmuss” scheme is, in contrast to the
"marginalist"y one that recognizes no pre-ordained boundaries for
social rights CTitmuss, 1974; Korpi, 1980). A particularly
important element in the identification of welfare state regimes
will, accordingly, be related to the blend of publicly provided social rights and private initiative. In other words, regimes can
be compared with respect to which essential human needs are
relegated to public versus private responsibility. To be
concrete, are housing needs, employment rights, health or human
capital development considered a matter of citizenship or an area of self-reliance? © The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
A second crucial dimension of welfare state regimes pertains to social stratification and distributional logic. As elaborated in
greater detail elsewhere <Esping-Andersen, 1982;1986),
comparative research typically ignores the fact that welfare
states are inherently systems of stratification, that not only
define citizens in terms of distinct strata, but also develop and
cultivate lasting loyalties, identities and status demarcations.
Three distinct clusters of regimes demarcate the cross-national
variation in this respect. First, the universalistic model
strives towards comprehensive inclusion and equal status of all
citizens. This is a welfare state that aims to undo the
stratification that results from markets and hierarchies, by
establishing classlessness in the world of social citizenship.
The Beveridge plan of universal flat-rate pensions, the
Scandinavian "peoples’ pensions", or the principle of universal
national health care belong to this category. Without doubt, the
Nordic countries constitute model examples of universalistic
citizenship in terms of status and rights.
A second distinct cluster is composed of nations which developed
social security along traditional corporativist lines; that is,
regimes that are founded on, and which seek to perpetuate,
status- or group- differentials. Identities, loyalties and
community is thus defined in exclusive terms with reference to
socially recognized class/group/status demarcations. The French
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
distinction between "cadres", civil servants, manual workers; the
German separation between "Arbeiter", "Angestelte", and Beamten;
or the proliferation of more than a hundred separate occupational
pension funds in Italy, all reflect the dedication to status-
defined citizenship. The corporativist tradition is powerfully
present in the structuration of Continental European welfare
states. A special case may be seen in the Dutch model, where
status is not defined on the basis of occupation or economic position, but with reference to religion.
The third cluster, finally, can be identified as regimes where
the stratificational principle pursued is that of the market;
that is, welfare states are designed to consolidate and
perpetuate market pre-emminence. This is the "residual ist11 model
defined by Titmuss;one in which the state extends citizenship
rights is only under conditions conditions where the market
mechanism Cor the family) fails. Social security will seek to
conform to the standards of market exchange, by stressing
voluntarism, actuarial ism, a flavor of "less eligibility", and
target-efficiency in social provision. A hallmark of this regime
is the pursuit of maximum private sector provision, the
accentuation of means-tested social assistance, a pronounced
actuarial-type relation between individual contribution and
claims. The resulting regime will cultivate a system of
stratification that is quite at variance with the alternative
regimes mentioned; on one hand, it is likely to forge a dualism
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
between state dependents and market actors; on the other hand, it
is prone to replicate the stratificational results of market
competition (and segmentation). The Anglo—saxon "New Nations" and
countries with a particularly flourishing liberalist tradition
are, comparatively, clear examples.
There are two obvious conclusions that emerge in this portrait of
welfare state regimes. One, that present-day welfare state
charactersitics are powerfully pre-defined by the history of
nationbuilding, mode of industrialization, class structural and
political development. This conforms closely to the arguments
presented in leading historical works on comparative welfare
state evolution (Briggs, 1361; Rimlinger, 1971; Flora et. al. ,
1981). A distinct welfare state regime may certainly owe its emergence to political conflicts and reforms within the fairly
recent past, as is clearly the case with the American (the New
Deal) or the Scandinavian (the breakthrough of social democracy
in the 193Q's). But, we would probably have to explain the
features of the New Deal, or the social democratic realignment, with reference to the patterns of nationbuilding that preceded
them. Nonetheless, what occupies us here are not theories of
ultimate historical causation, as much as the identification of regimes.
The second conclusion is that the place of private provision
within the total welfare mix is a crucial factor in
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
distinguishing the stratificational and distributional nexus of
welfare state regimes. Private welfare must , ironically, be
analyzed in order to identify the welfare state; and as we shall
discover, the opposite is equally true. But, besides its
relevance for c 1 assi fi eatery endevours, the role of private
provision is an analytical precondition for any serious testing
of causal theories of welfare state growth, precisely because state provision is so powerfully intermingled with private
provision (Rein and Rainwater, 1986). As a conceptual and
theoretical problem, it is also an issue of methodology.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
2. METHODOLOGICAL PROBLEMS IN WELFARE STATE REGIME RESEARCH
The lion’s share of comparative welfare state research is notably
one-dimensional. First, it examines by and large only
governmental activity and thus ignores the private sector.
Secondly, it is overwhelmingly linear in its conception of
historical development and cross—national differences; on both
axes, analyses tend to focus on linear relations such as
expenditure levels and levels of economic development, working
class power mobilization, population ageing, or even
democratization. Thirdly, it is heavily wedded to mono-causal
arguments:"politics" versus "economic development", for example.
Reorienting research in the direction of a policy regime approach
invites new methodological considerations. The policy-and
sectoral mix that a regime approach demands, means that the
object to be studied must be defined relationally and
institutionally. This does certainly not mean that, a priori, any
linear-analytic approach must be discarded, but it seriously
raises the possibility that cross-national differences must be
identified via variables and criteria that are categorical and thus non-linear. If nations cluster, normal regression techniques must be applied with caution.
The great theoretical debates in welfare state research have
focused on a limited array of causal forces, the principal ones
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
being economic development, demographic change, democracy, and
political mobilization. The two first, in particular, strongly
argue cross-national convergence; the case for the latter rests
pretty much on divergent welfare state development. In any case, a falsification of either argument can hardly rest on a study of state activity alone. To give one concrete example, actual public
pension expenditure may be increasingly divergent across
similarly developed nations, but this does not necessarily
invalidate a demographic or economic explanation, since
(government) mandated private employer pensions , if included in
an expenditure measure, might very well equalize the variation
otherwise observed.
But, even more important is the need to specify more precisely
what causal forces operate at what explanatory level. An example
from t he power mob i1i z at i on approach will illustrate the point.
Political parties operate at the level of parliamentary
decisions; trade unions at the level of industrial relations.
Both are manifestations of wage earner mobilization. Yet, nations
differ dramatically with respect to the mix of these two power
resources. Some score very high on both (Scandinavia, Austria), some are characterized by weak or non—existant labor parties, but
trade unionism of some significance (the United States or
Canada), others by the reverse (France). It would be a more
sophisticated theory indeed, could it specify the conditions of
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
power mobilization that would produce any given combination of strong state or labor market based social provision.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
THE PRIVATE-PUBLIC MIX OF SOCIAL PROVISION IN ADVANCED CAPITALIST DEMOCRACIES.
Any study of the public-private mix faces formidable obstacles.
One is the paucity of reliable data, especially going back in
time; another is the difficulty of defining what exactly should
be considered private or public.
The definition issue must be first resolved. With regard to
pensions, the residual that emerges after having isolated
legislated social security programs, proper, is complex. Each
nation has experienced its own peculiar evolution in the arrival
of any given "package" of pension plans so that, what is termed
private in one may not necessarily enjoy a private status in
another. How, for example, does one treat private employer plans
that have been mandated by government?
The definitional question must be r esolved i n light of our
t h eor e t i c sa 1 c on cerns. Wit h this in rn i n d ,. we must begin with the
basic criteria of legislation, based on the argument that any
test of political power mobi1i zat ion theori es must i dent if y
outcomes that flow from parliamentary-level conflict and power.
This implies that we must include all pensions under the rubric of public if they are: a) directly legislated and/or administered
by the state, or b) if there exists a clear and explicit
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
government mandate that the private sector provide a given type
of pension. This criteria implies that the Finnish, British and
Dutch second-tier industrial pensions must be considered public,
but given the absence of an explicit statutory government dictum in France Cor Switzerland, until 1982), the labor market pensions there are to be classified as private.The same argument pertains
to the Dutch "Company pensions". The year in which government
mandating occurred defines the shift from private to public.
A second category that must be isolated are government civil
service pensions. These mirror the government's role as employer
and are thus occupational in nature ; despite being financed by, and paid out of, government budgets, they have very little to do with legislated social rights but much to do with a particular status privilege. They reflect, in a sense, corporative privilege in the classical sense.
In the field of pensions there remain two classes of "purely"
private provisions occupational pension plans, and individual
annuities. The two types are important to keep seperate due to
the differential logic that they espouse. In a strict sense, it
is difficult to consider occupational plans as simply market-
conforming. They often mirror employer paternalism (in the form
of the traditional gratuity pension) , collective insurance (in
the form of group-plans), or labor market collective bargaining.
Today, the first type is relatively marginal (excepting Japan),
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
and we can therefore view private occupational pensions primarily
as an indicator of trade union mobilization. Theoretically, we
would anticipate that private occupational pensions are
positively related to trade union power, in particular under
conditions of weak labor movement legislative capacity. Finally,
the category of individual insurance, such as life insurance
plans, reflect the traditional model of self-reliant
individualism within the framework of competitive contracting.
As noted, the empirical problem is formidable because of the
generally meagre and uneven statistical coverage of private
plans. There is normally reliable information on life insurance
plans, mandated occupational plans, and funded, or trusteed,
labor market pensions. The major shortcoming is lack of data on
unfunded "gratuity-type" pensions. For some countries, Japan
especially, we will thus underestimate the scope of private
provision. Historical data is, moreover, almost impossible to
assemble. This limits our inquiry to very recent data; for two
nations, Austria and Italy, there exist no up-to-date information
whatsoever on private occupational pensions. It is universally
agreed that their scope is tiny; we therefore decide to score
them equal to the lowest within the sample Can approximation for
Italy is facilitated by known private pension expenditures for
1 9 7 0) . © The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
The following empirical overview examines the public-private mix
within the two dominant social welfare program areas: pensions
and health care. For pensions, we present two different indicies;
one, total expenditure by program category; two, the sources of
income among aged households. With respect to the former,
expenditure data was prefered over coverage or finance data,
since it reflects the actual status at present; finance data,
given long-term funding, might only reflect a possible future
scenario. And the problem with coverage data is that broad
coverage does not necessarily imply that private pensions play
any significant role; in Sweden, for example, there is virtually
universal coverage of occupational private plans, but the
benefits paid out are exceedingly marginal.With respect to the
1atter, it is for a number of countries , p oss i b 1e to utilise
survey data to est i mate the r e 1 at i ve i mpor t an ce of public and
pri vate pensi ons, as well as of work income and individual
savi ngs i n the total income package of aged households. This
affords us the possibility of examining the continued importance
of work as well, and it permits us to compare expenditure-based
data with income-source data as a
means to ascertain data reliability. Issues of definitions and
data sources for pension expenditures are treated in detail in a
seperate appendix, attached to the tables.
For health care, we are able to present over-time data (1960 and
1980) on the public-private expenditure mix, thanks to the work
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
of Poullier on behalf of the OECD. A third, and final, indicator
of the public-private mix will be presented in the form of
employer contributions to legislated, as opposed to voluntary,
fringe benefits. This offers yet another means of gauging the
reliability of the expenditure data, and of estimating the
relative importance of private provision within the labor force.
3.1. STATE AND MARKET PROVISION OF PENSIONS
For a few countries, it is possible to trace expenditure
developments of occupational plans since the 1950's. Table 1
presents estimates of occupational pension expenditure as a
percent of GDP between 1950 and 1980 for 12 nations.
Table 1 illustrates a number of important phenomena that we shall return to subsequently. First, that private (funded) occupational
pensions played a relatively marginal role until very recently.
In 1950, private pensions (as well as public) absorbed a small
proportion of national resources. Secondly, the table reflects
the differential trends among nations. In some countries,
especially Australia,France, Switzerland, and the United States,
private plans have grown very large, indeed. A substantial
growth also occurred in Denmark, Canada and Holland. The opposite
is true for Finland, Sweden, and the U.K. where,of course,
government legislation or mandating marginalized the private
sector. The bi—modal distrubution of private pensions across
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
nations remains when we examine more closely the breakdown of pension expenditures around 1980. See Table 2.
Table 2 shows the relative importance of social security,
government employee, private occupational , and individual
pensions as a percent of GDP for 18 nations. For all four
categories,the cross-national variance is substantial. Social
security pensions vary from a low of just above 2 percent in
Japan to a high of almost 10 percent in Sweden. Government
employee pensions range from less than 1 percent in Australia and
Canada to a high of almost 4 percent in Austria. Private
occupational plans are insignificant in Austria and Italy, but
very large in France, Switzerland and the U.S.. Individual
annuities (which may include some group plans, too) play a
dominant role in Canada, Denmark and Germany, but are peripheral in Finland, Italy and the U.K.
A clearer picture of the relative pension mix can be established if we present the ratio of each of the four categories to total pension expenditure, as seen in Table 3. Note that we have here set Austria's and Italy's occupational pension expenditure equal to 0.1 percent of GDP.
Table 3 shows a distinct cross-national clustering. Low levels of
social security in the overall m i x seems clearly accomp an i ed by
large shares of private sector provision. This is almost
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
tautological- Countries are surprisingly bi—modal with regard to
their emphasis on public employee pensions. In one group, their
iinfluence is overwhelmingly strong: Austria, Belgium, Finland,
France, Ireland, Italy and Japan. This is important to note,
since this is the same group of c ountries in which soci a1
security generally evolved along strongly demarcated status-
corporative lines (see Esping-Andersen, 1986). In contrast, the
liberalistic and social democratic traditions display low levels
of civil service pension expenditure (Australia, Canada, Denmark,
New Zealand, Norway, Sweden, and Switzerland). What primarily
distinguishes the latter is the relative position of social
security versus private sector provision. Australia, Canada and
the U.S. are examples of modest social security and strong
private commitments, whereas Norway and Sweden (and possibly also Denmark) exemplify the opposite. A preliminary classification of "pension regimes" is thus possible:
1- Corporative systems, in which status is a key element in the
pension program structure. In this regime, the private market is
generally marginal, and social security tends to be highly
occupationally segregated with particularly pronounced civil
servants' privileges: Austria, Belgium, Germany, Italy, Japan,
with the possible inclusion of Finland and France.
2- Residualist systems, in which the market tends to prevail at the expense of either social security or civil service privilege,
or both: Australia, Canada, Switzerland, the United States, and
possibly also France.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
3. Universalistic systems, in which population-wide social
rights marginalize both privilege and markets: New Zealand,
Norway, Sweden, with the possible inclusion of Denmark and
Hoi1and.
This classification leaves us with only one really mixed case,
the U.K.. The same basic patterns would be expected to emerge
when we turn our attention to the structure of income among aged households- We would anticipate that the relative importance of
work,investment, and private pension incomes would be highest in
those cases with residualist regimes.
As can be seen in Table 4, there is some truth in this, albeit
with important exceptions. But, first of all, Table 4 by and
large confirms our previous picture, even if it is unable to
distinguish civil service pensions from general social security.
Work incomes are important in the countries where we would expect
it: Canada, and the U.S., but also in Denmark , Ireland, and the
U.K. For Denmark and the U.K., the reason is probably the
straightforward one that the social security system offers only modest flat—rate pensions, and that a second-tier either does not
exist, or is too recent to provide for substantial pension
income. For Ireland, the reason would primarily have to do with
the continued importance of rural self-employment. However, given
that the table refers to households, the work income will, in
large measure, refer to spouses.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
The distinction between investment income and private pensions in
Table 4 corresponds to the relation between individual sel f-
reliance and collective bargaining. In this sense, individualism
appears especially pronounced in Canada, New Zealand, and the
U.S., and least so in Norway, Sweden, and Finland. For a few of
these countries we have household income survey data for the
early 1960's, as well. This permits us to trace major structural
changes for Denmark, Canada, the U.K. and the U.S. The primary
trend is a fall in the significance of work, especially in
Denmark and Canada; the rise of social security; and also a rise
of investment income, especially in Denmark and Canada (OECD,
1977; Goodman, 1986') .
The statistical correspondence between our two types of public- private pension mix indicators is fairly strong. Based on the 10 nation sub-sample for which household income data are available,
the zero-order correlation between a) the private pensions share
of household income , and b ) occupational pension expenditure as
a percent of total, is +.602. Similarly, the correlation between
the two indicators of the social security share is +.683. Both
work incomes' share, and private pensions' share of household
incomes is strongly negatively correlated with the social
security variable (-.694, and -.636, respectively).
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
Table 5 presents the private-public mix in health care
expenditures for 1960 and 1980. At first glance, the table
illuminates two important factors. One, a pervasive shift from
private to public over the two decades. The 18-country mean of
private as a percent of total declines from 37.3 to 21.5. This
trend is accompanied by a certain international convergence:the
standard deviation declines from 17.9 to 13.5. Two, there appears
to be a somewhat similar public-private pattern to that which
existed for pensions; that is, the private-public mix is
consistent across different policy domains. Nations that are
biased in favor of private pensions seem also to be so in health
care. Thus, the "residualist" pension group includes by and large
the same nations with regard to health care: Australia, Canada,
Switzerland, the United States, with the possible inclusion of
France. The only real exception is Austria, where private
pensions are virtually non-existant, but where private health is
substantial. The zero-order correlation between the share of
private pension expenditures (including individual annuities),
and private health expenditure shares for 1980, is .630.
Finally, Table 6 presents data on the private-public mix in
employer payments towards fringe benefits. Expressed as a percent
of total labor costs, the table differentiates between legally
required employer payments (social security contributions, for
example) and voluntary contributions (which may be either
gratuituous or collectively negotiated). Although the table
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
covers only 16 of the total IS nations, the pattern that emerges
would in general seem to confirm our previous findings regarding
the public-private distribution- Based on the 16 nation sub
sample, the zero-order correlation between the share of voluntary fringe benefits and private occupational pension expenditure as a percent of total is .823.
In sum, the evidence suggests a generally good empirical
correspondence among our various indicators of the public-
private mix in social protection. This implies also that the
relevance of a "regime approach" to the cross-national comparison of welfare states may be fruitful. The clustering along our basic dimensions is both sufficiently clear on single indicators, and
seems also to be upheld when we cummulate different
indicators. We now turn from the quantitative overview to the
question of how differential private-public structures evolved
historically. The discussion will be limited to the field of
pensi ons. © The
Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
4. THE HISTORICAL ORIGINS OF PENSION STRUCTURES.
It would be a serious analytic fallacy to impose upon the 19th
Century the conceptual meaning of pensions and retirement that we
have today. The practice of retirement was a marginal phenomenon
until World War II (Graebner, 1980; Myles,1984). Neither public
policy,.nor private choice, assumed that a person would normally
withdraw from active working life at a certain specified age, and henceforth enjoy old age in leisure. There were of course persons
drawing a pension, but until recently only very few aged would
have been in command of sufficient income to circumvent
dependency, poverty, or the compulsion to work.
Social security pensions emerged at the close of the 19th
Century, spread more rapidly in the interwar decades, but hardly
provided an institutionalized means of retirement until after
World War II (Perrin, 1969). This does not imply, however, that
the world of pensions was once a predominantly private domain to
be successively crowded out by the state. Indeed, from their
origins to the present, pensions have continuously progressed
through an intricate mix of private and public; the state has
consistently been highly present.
Income security for the aged in the 19th century was a question
of protection against impaired working capacity or loss of
breadwinner. Besides working (the norm), the principal sources of
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
old age income protection lay in family care, thrift or charity,
within the private sector, and poor relief within the public
sector. As actual pension plans arose, the state was on center
stage. In its role as employer, the state was often a pioneer of
occupational pensions . The British government introduced civil
service pensions as early as 1834, the same year in which the new poor laws established the principle of less eligibility; New York City provided pensions for its employees by 1857. Taxation policy came to play a critical role in boosting private sector pension plans; government regulation of early friendly societies as well
as of rules pertaining to tax-exempt pension payments has
directly shaped the structure of private pension developments.
And, finally, the state's indirect stimulus has naturally been
formidable; the absence of legislated pensions, insufficient
coverage, meagre benefits, or restrictive eligibility conditions would almost automatically inspire private alternatives.
Any discussion of the history of pensions must take into account
the radically different structural conditions that prevailed in
early industrial capitalism. Due to class structure and
demographic conditions, objective need would be defined
di f fer ent1 y .
A large bulk of the 19th Century population was rural; self-
employment remained significant. Thus, in 1870 agriculture's
share of total employment usually surpassed 50 percent (65
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
percent in Austria, 52 percent in Denmark, 50 percent in Germany
and the United States, but only 23 percent in Britain). The
relatively marginal commodification of labor that this implies
means that dependence on wages or insurance incomes in old age
would be somewhat peripheral.
Employment structure notwithstanding, prevailing demographic
conditions would have subdued massive demands for old age
pensions. Around 1820,life exepectancy at birth was normally well below 40 years; by 1900, still below 50 years (in comparison to contemporary societies with a life expectancy of more than 70 years). To adjust for high rates of infant mortality, we may note
that life expectancy at age 20 around the turn of the Century
hardly surpassed 40 years: 40 years in Austria, 41 years in
France, 42 years in the U.S., and 46 years in Sweden (United
Nations, Demographic Yearbook, 1949). In other words, very few
survived beyond the age of 65. Hence, the population ratio of
those 65 years and older was, around 1870, between 3 and 5
percent (as compared to 11-15 percent in the mid-1970's)
(Maddison, 1982).
The need for old age pensions in 19th Century industrial
societies was perhaps modest, but certainly not absent. With the
consolidation of the commodity status of labor emerged the risk
that work incapacity would jeopardize survival. Widows, the
disabled, and the old were easily the victims of extreme poverty.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
Von Balluseck (1983:219) reports that the vast majority of those
receiving alms in Berlin in 1867 were over 60, or widows. But
abject or prospective need for protection was hardly matched by
any sustained capacity to ensure that it be met.
In the 19th Century, old age income protection was managed in a
variety of ways. One, most continued to work; a norm that was
pervasive well into the 20th Century. Ball (1978:80) reports that almost 70 percent of American males over age 65 worked in the
1890’s;Gui11emard (1980) reports similar figures for Trance. In
fact, the early social security schemes or employer plans were
not meant to replace work income, but rather to supplement or
compensate for diminished working capacity (Myles, 1984).
Employers would often provide sheltered jobs for their aged and
less productive personel.
The second major means of securing a livelihood in old age was
family. The importance of family was two-fold. One, the
traditional practice of transferring the means of production to
the younger generation and, subsequently, living off the
"dividend"; two, the necessity of relying on the family’s general
welfare function. A 1929 New York survey showed that more than
half of the aged persons were dependent on support from family
and friends (Weaver, 1982:42).
The third avenue was charity, in many nations predominantly
organized by the church. The same New York survey cited above,
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
showed that percent of the aged were solely dependent on charity, but this misrepresents its real significance. As late as
1927 in the United States, total private charity payments were 6
times greater than total public welfare expenditure (Weaver,
o p.c i t . p .20).
The fourth avenue was publicly provided poor relief which, as
discussed above, remained practically the only state income
maintenance program for its citizens, at least until the turn of
the century. As the German example showed, the relief rolls were often swelled with old workers with neither work nor property.
Even as late as 1954, there were one million aged Britons
dependent on social assistance (Brown and Small, 1985:136).
However, where political adherance to hard-core liberalism held
sway, poor relief was not a particularly dependable source. In
the United States, there were many states which refused to grant
cash assistance to the needy (Weaver, op.cit); in Britain, the
reign of the poorhouse would ensure that all but the most
desperate would turn elsewhere.
The fifth and sixth avenues, namely either state or private
pension insurance, are today the all—dominant but were in the
19th Century extremely marginal. Bismarck's pioneering pension
insurance first came in 1889, and most nations failed to
introduce public pensions for workers until well into the 20th
Century. This does not mean that the state was wholly inactive;
indeed, states pioneered the principle of occupational pensions
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
(for their own civil service), and frequently mandated them in selected, high-risk or high-priority occupations, such as seamen
or miners. But, clearly, these kinds of plans failed to cover
those citizens with the greatest potential need, namely the
property-less wage worker. One type of government provision was occasionally of some importance, and that was veterans' pensions,
especially in the United States. Skocpol and Ikenberry (1983)
argue that the surprising lack of popular pressure for pension
legislation in the U.S. was due to the rather frivolous
disbursement of Civil War pensions.
Private sector insurance was not capable of substituting for the
huge gaps left by family, charity and state. For our 19th Century
forebears, there were normally two private sector pension
options. First, and most important, were friendly societies and
their kindred. These were generally thrift organizations for
distinct social groups, be they occupationally or trade union
defined; they often evolved out of the ancient guilds. In some
nations their scope was considerable. Gilbert (1966) has
estimated that about 50 percent of working class males in Britain in 1880 were members of friendly societies;in the United States around 1890, there were 3.7 million members — equivalent to about
5 percent of the labor force (Weaver, 1982:46). Ashford
(1986:151) notes that the French mutual societies counted two
million members in 1902. Friendly societies were, on the other
hand, only marginally involved in old age pension disbursements.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
Their activities centered on sickness protection, unemployment,
burial costs, and survivors. Moreover, their membership largely
consisted of the better-off, skilled sections of the working
class: those capable of furnishing the required weekly
contributions. As a result, their capability of securing pension incomes among the aged was somewhat less than impressive. This is
clear in the United States during the 1920fs ; while membership
had grown to more than 5 million, the total number of pension
recipients in 1928 was only 11 thousand (Weaver, 1982).
The alternative source of private sector pension protection was
employer plans. There were a small number of industries that
established private pension plans at an early date, notably in
the railroads, mining, and among seamen. These early forerunners
of the industrial pension were, moreover, often sponsored by
governments. In addition, a handful1 private firms began in the
19th Century to establish company pensions. These were almost
universally corporate vanguard firms, such as American Express
(then a shipping company), ATT, Carnegie Steel, and Kodak in the
United States; Krupp, Siemens, and Hoechst in Germany; and
Cadbury, Lever, and Rowntree in Britain. Such plans, however,
were primarily addressed to salaried staff and they were, above
all, paternalistic and gratuituous. Benefits were discretionary,
usually contingent on life-long and loyal service, and were
financially precarious. Thus, rather than being premised on
contractual entitlement principles, they were paid out of current
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
ompany revenues- As such, a person's pension prospects were company fortunes.
ntimately connected to the vagaries of
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
5. THE HISTORICAL EVOLUTION OF THE PRIVATE-PUBLIC MIX
The foregoing portrait of 19th Century pensions indicates that
the market system was a fairly residual element in pension
provision.Therefore, when state pension insurance eventually
emerged, there was hardly any "crowding-out" effect possible. The 19th Century private market had, in fact, not discovered much of a niche for pensions.
Paradoxically, public and private pensions emerged and grew in
tandem. What was gradually crowded out were the pre-capitalist
legacies of social protection, such as the family, and charity,
together with the poor laws and friendly societies.
The causal structure of pension evo1ution rests on a combination
of sociological variables (demography and employment) and
political transformation. The demographic structure began to
change dramatically around the turn of the Century, particularly
with respect to family structure and life expectancy. In the
three decades after the turn of the century, male life expectancy at age one jumped by almost 10 years in most nations; the ratio of citizens 65- and older thus grew (United Nations, op.cit). It was also in this era that the shift from self-employment to wage
labor, and from agriculture to industry was particularly
powerful. The share of agricultural employment declined from 50
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
percent in 1370 in Germany and the U.S. to about 33 percent in
1910 . Hence, neither the family nor the farm would afford much
old age protection for the average worker of the new social
order. And, meanwhile, need grew objectively.
It was also an era in which the meaning of work and employment
was recast; from artisan-type shops and small manufactures to
modern mass production, from a stress on labor intensity to a
growing concern with maximizing productivity. The Progressive
Era, be it in the United States or Europe, was one that
inaugurated ideas of scientific management and optimal efficient use of labor power. Thus arrived also the desire of management to shed itself of ageing workers (Myles, 1984; Qraebner, 1980).
While need for pensions obviously grew, so did citizens'
collective capacity to demand them. Over the course of the 19th
and even 18th centuries, there had been countless proposals and
plans for both public and private pension plans. Daniel Defoe had
proposed a "pension office" as early as 1697;Thomas Paine wrote
it into his basic rights of man, and presented an actual plan to the Lower House; both revolutionary and Napoleonic France debated
old age protection CAlber, 1982:32-33;Ashford, 1986). They all
came to naught because there was no political will. The period
around the turn of the Century, however, changed the conditions
under which political will could find expression. In the labor
market emerged trade unions and, gradual1y , industry— and
nation-© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
wide labor associations. Their legal recognition was generally
established during the closing years of the 19th Century, and
their growth was almost everywhere explosive. Universal
suffrage spread around World War I, allowing the rising labor
parties representation and some leverage. In other words, the
"social question" could become political.
In this historical nexus emerged the modern blend of private and
public old age protection. In the private sector, the movement
was towards two basic systems; one, individual (life) insurance;
the other, collective types of occupational and industrial
pensions that, incidentally, should include government employee
occupational pensions. In the public sector emerged two models of
social security pensions; one, a basic, usually flat-rate,
minimum with its roots in the social assistance tradition
(Denmark, Australia, for example); the other, an insurance scheme
pr emi sed on actuarial 1ogi c and an emp1oyment r ecor d .
In a nutshell, the thrift element shifted from the friendly
society to the modern insurance company; the employer gratuity
pension was gradually transformed into a contractual fringe
benefit in collective bargaining; government poor relief became
social security.
The early initiatives within both public and private pensions
were frequently motivated by desires to weaken the labor
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
movement s . Private sector employers instituted pensions as a
means of weakening union loyalties, divide the employees, and
glorify management (Myles, 1984; Graebner, 1980; King, 1978;
Jackson, 1977). Unions were therefore antagonistic. Similarly, early state legislation of pensions was typically undertaken as a
means to arrest the growth of labor movements, and to redirect
workers' loyalties toward the state and the existing order
CRimlinger, 1971). This was the clear rationale behind the German , Danish and Austrian reforms; this was also what guided the 1891
Papal Encyclical, Rerum Novarum, which came to exercize
tremenduous influence on social policy within the Catholic
nations. The rise of modern pension schemes was therefore not catapulted by labor; its influence was little more than that of the passive kind.
Given their relative powerlessness to influence either state or market,it is not surprising that the labor movements concentrated
on cultivating their own plans. It was during the turn of the
Century that these came to enjoy a fast pace of growth. As noted,
the Arnerican trade union societies grew from 3.7 mi 11 ion to 5.3
million members between 1390 and 1900; they continued to grow
until the Great Depression, when they suffered severe financial
difficulties and gradually lost out to private insurance
companies, employer plans and public pensions (Weaver,
1982:46ff). It should be kept in mind, that probably only l/4th
among the large numbers covered in trade union plans were covered
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
for pensions (Weaver, op.cit. p.48), and very few actually became recipients.
The situation was fairly similar in Britain where trade union
fund membership grew to 5.5 million by 1338 (Brown and Small,
1385). That is equal to 24 percent of the British labor force of
that period.
From the point of view of the labor movements, the friendly
society strategy was discovered to be less than optimal. It did,
of course, offer an added attraction for workers to join the
unions, but its marginal dividend for mobilization was easily
offset by its segmentary consequences. It spoke to the better-off
workers, able to contribute and easier to organize, but largely
excluded the weakest, lowest paid, and most marginal workers—
precisely those that would be most important to organize and most
in need of protection. The workers' insurance model embodied in
the trade unions was, accordingly, potentially dualistic.
Besides, it was prone to financial difficulties due to recurrent
unemployment and business cycles. Thus, Weaver (1382) argues that the decay of trade union funds in the United States was primarily caused by their inability to weather the Depression.
The institutionalization of both private and public pensions
coincided during the early decades of the 20th Century. In
several respects, public policy actually nourished market
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
expansion. First, governments began in earnest to build up
occupational plans for their employees at both central and local
level. In Britain, for example, all local governments instituted
teachers pensions by 1898; by 1937, this was mandated for all
employees of local governments (Brown and
Small, 1985). In the United States, there was a rapid growth of
both federal employee pensions and state/local government
coverage. By 1928, total coverage of public employees reached
about 1 million; about 25 percent of state/local employees
(King, 1978:200; Weaver, 1982:48). In the same year, veterans
pensions were still the greatest source of pension income,
benefitting almost 500.000 individuals, or somewhat close to 85
percent of all pension recipients in the U.S. (Weaver, op.cit.
p.48). This was also the era in which many public utilities, such
as transportation, gas and electric, were nationalized; as a
consequence the membership of government employee plans grew
additionally. Even in the United States, the Federal government
came to the rescue of the railroads, leading to the
nationalization of their industrial pension plans in 1935.
The direct impact of government occupational plans on private
sector growth was two-fold: one, they were important as agenda
setters, stimulating demands for equal protection among private
sector employees; two, being often administered by insurance
companies, they helped bolster their role as key institutions in
the further growth of private sector pension protection in terms
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
of both individual, group, and industrial schemes- In brief, government employee plans helped construct the private market.
A second important influence of government were its fiscal and
regulatory interventions. Between the two world wars especially,
governments introduced the idea of tax expenditures to give the
market incentives for social protection (for example, the 1921
Finance Act in the U.K.; the 1922 and 1924 revenue act in
Denmark, and the 1926 revenue act in the U.S.), by allowing
deductions for contributions. In turn, such tax privileges
spurred governments to regulate private sector initiatives,
stipulating financial solidity, funding, contractual rights and
accountability. Thus, governments came to further bolster the
importance of insurance companies; concommitantly, they helped
reshape the nature of employer pensions, discouraging the
traditional discretionary gratuity principle and encouraging the
rise of regular,negotiated, and contractual fringe benefit plans
for employees.
Ironically, the legislation of social insurance schemes also
provided considerable room for private pensions. Where
legislation was slow to happen, as in the United States, the
incentive was obvious. Where it did happen, the private sector was typically antagonistic, fearing a crowding out effect. When
Germany, for example, legislated pensions for salaried employees
in 1911, the insurance industry was vehemently opposed CJantz,
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
1961:149!). But there seems to be no unequi vi cal 1 y negative effect
of legislated pensions on private pension growth. The social
security reforms that were introduced before World War II offered
very meager benefits, incomplete coverage, and where the
insurance model was adopted, the contribution requirements meant
that they really only applied to future generations.Thus, in
Britain the 1908 law provided only means-tested benefits to aged
citizens beyond the age of 70; the subsequent contributory
pension of 1925 assumed 40 years contribution and was, anyhow,
designed so as to provide only a minimum to be supplemented with
private occupational pensions or individual savings. The German
ArV workers' pension insurance was launched as a disability
pension for those unable to work, targeted for workers over 70,
premised on 35 years contribution. As Myles (1984!) shows, those
Germans that did receive a pension would probably have been
unab1e to 1i ve on i t .
Generally speaking, government pension legislation before World
War 11 was in the spirit o f o f f er ing mini malist pr ot ect ion, so as
not to discourage labor supply; it was assumed that necessary
pension supplements should be purchased in the private market. In
the United States, the spirit of minimalism was, perhaps,
unusually extreme. But the same basic principle applied also to
European practice. Welfare Capitalism was a slogan that depicted
well the definition of government responsibility for the era as a whole. © The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
The evolving pension mix of the first decades of the 20th Century
mirrors a welfare capitalist model. It embodies a matrix of
developments that bridge the largely pre-capitalist terrain of
the 19th Century, and the welfare statism of the post World War
II period. Objective need for pensions had clearly asserted
itself; wage earners' power to demand action was increasingly an
inescapable reality; industry's new productivism diminished the
value of aged workers; states had taken decisive steps to
encourage and even create a market for pensions; social security
had been set in motion, but was still no serious contender to
private pensions. In brief, this was the era of Titmuss'
residualist state.
Private pension schemes grew markedly during the first decades of the 20th Century. But more important than their growth was their transformation. The movement was from discretionary gratuities to
contractual arrangements; from unfunded plans to insured,
trusteed schemes; and from catering to narrow strata of upper-
echelon staff, they were gradually extended to manual workers.
Within this process occurred also a transformation of the age-old
tradition of thrift. The friendly societies Cor the piggy-bank)
gave way to the modern insurance company's life insurance plan.
More slowly, the savings embodied in the family farm gave way to
ur ban homeowner ship.
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research
The promotion of Welfare Capitalism, it was believed, would allow
the modern capitalist enterprise to substitute for the
unpleasantly communistic flavor of social insurance. It spoke to
the new world of the Progressive Era, the emerging modern
corporation, scientific management, and the heightened concern
with good labor relations. It was premised on hopes that social
protection was indeed a cause to be happily embraced by private enterpr i se.
Both companies and whole industries began, in the United States,
to establish funded and trusteed pension plans, increasingly to
be placed with insurance companies. By 1930, the insurance
industry held a total of 83 million policies (including
individual, group, and company plans), and was paying out
benefits amounting to $ 2 Billion (more than charity, public
employee plans and the scattered state pension plans paid out
together)(Weaver, op.cit.p42). This staggering amount, of course,
included mainly other risks than pensions, but it was in the
pension field that the life insurance industry's fastest growth
occurred. Group plans (almost exclusively contracted by
individual manufacturing firms) accounted for 1 percent of its
business in 1915; 15 percent by 1935 (Weaver, op.cit. p.47).
Industrial pensions that addressed manual workers grew at a fast pace. With quite modest contributions (that were tax-deductible)
, such pension plans spread quite impressively in the 1320's. Ely
© The Author(s). European University Institute. version produced by the EUI Library in 2020. Available Open Access on Cadmus, European University Institute Research