Full text of DOI:10.1111/j.1564-913X.2000.tb00408.x

EI nq t ue ra nl at tri eo an t aml eL na tb ,o su or c Ri ae lv pi er wo t, e Vc ot i lo. n1 3a 9n d( 2 0i n0 c0 o) ,m Ne o s. e2 curity for women
149
Equal treatment, social protection and
income security for women
Linda LUCKHAUS*
1
nsuring income security is a major function of social protection. Arguably,
E
that is its defining task. However, the ability of social protection to pro-
vide an adequate and reliable source of income for women is problematic —
partly because of the other functions that, historically, social protection has
been required to perform. These include the promotion of specific forms of
financial dependency which are rooted in and characteristic of gender and em-
2
ployment relations in the wider society. These relations of dependency pro-
vide a fragile source of income for women; this fragility is then imported into
the social protection systems giving them effect. The insistence on sex- and
3
gender-based equality in the field of social protection can minimize or even
eliminate this fragility, depending on how the notion of equality is defined and
its purpose construed, and there are many ways of doing this. Which of these
definitions dominates legal and political discourse at any one time depends on
prevailing views of the proper role of social protection and the appropriateness
of the pattern of income provision and dependency to which this role relates. It
also crucially depends on the changing nature of cultural and economic prac-
tices which ultimately set new limits in this sphere.
*
Senior Lecturer in Law, University of Warwick. The research on which this article is
based was funded partly by the British Academy, under its Research Leave Scheme, and by the
Robert Schuman Centre for Advanced Studies of the European University Institute, Florence.
The article is based on a background paper commissioned for the ILO’s World Labour Report
2
000: Income security and social protection in a changing world.
1
Here defined as social insurance, social assistance and private schemes, the premia for
which are not wholly market determined.
2
Social protection systems may also reinforce non-financial aspects of gender and em-
ployment relations by, for example, facilitating the recruitment, retention and disciplining of
labour. Additional functions of social protection include: encouraging social cohesion and
stability and the avoidance of public disorder; promoting growth of capital markets and finan-
cial services; facilitating economic growth (through expansion of demand or, alternatively, by
encouraging saving).
3
The term “gender” is seldom, if ever, used in national and international instruments
prescribing equality, equal treatment or non-discrimination between men and women. The term
normally used in this legal context is “sex” and this applies especially to instruments adopted by
the European Community (hereinafter the EU) which are the main legal focus of this article.
Claims to equality under the relevant EU legislation must be made in terms of sex rather than
gender and this, rather than any insistence on “biological differences” accounts for use of the
term alongside gender which, for all other purposes in this article, is the more appropriate term.
Copyright © International Labour Organization 2000

1
50
International Labour Review
This article first explores the links between income security, sources of
income support, and definitions of sex and gender equality in social protection.
The focus is social protection as provided in developed countries, particularly
those of western Europe; reference is made to developing and transition econo-
mies where possible. Next, the article identifies the different types of discrimi-
nation still to be found in social protection systems, applying the definitions of
equality developed in the preceding section. Discriminatory practices are se-
lected for examination from among recent developments in those systems for
what they show of shifts towards greater or lesser discrimination. Finally, the
article examines the relationship between equal treatment, social protection and
income security, concluding with some normative thoughts on which of the
various definitions of equal treatment is to be preferred, and with some ideas
on how sex and gender equality may be secured in social protection.
Income security and social protection
In everyday terms, the two important aspects to income security are: the
amount and adequacy of income; and the regular flow of income. In other
words, income security means that there should be a constant flow of income
adequate to live on. Probing further, it is clear that both adequacy and reliabil-
ity are contingent on the source of the income, that there is more than one
possible source, and that one source may be more effective than another. In
developed, western economies, there are three main sources: sexual relation-
ships between cohabiting married or non-married partners of the opposite sex;
4
social protection systems; and employment and self-employment. Income de-
rived from employment may take the form of pay or benefits. Benefits derived
from the employment relationship may be regarded both as deferred pay and as
a form of social protection.
However, these main income sources are not immutable; nor is the pattern
of use made of them by individuals and groups. As the economic activity of
married women in most west European states has increased in recent years, so
has their reliance on paid employment as an income source. There is, however,
one fairly constant feature in the pattern of use. People caring at home for
children, the elderly or persons with disabilities (i.e. not as part of a commer-
cial arrangement) are precluded from engaging in paid employment during that
period and thus from access to income from this source. Such people, there-
fore, must look either to their partner or to social protection for material sup-
port so long as they are engaged in this activity. However, they are not free to
choose in this. Social protection systems are generally structured in such a way
as to ensure that a partner is the first port of call in these situations. There is,
for example, no social insurance benefit to cover the risk involved in giving up
4
A developing country and transition economy perspective would require the list to
incorporate wider “family” relationships, national and international aid organizations, and em-
ployment relationships in the informal economy.

Equal treatment, social protection and income security for women
151
5
paid employment to care for children, the elderly, people with disabilities, etc.
Someone engaged in informal unpaid caring work of this kind is assumed to
have a partner who is a paid worker (a “breadwinner”) and, therefore, able to
provide for them. Only if the breadwinner partner is not engaged in paid work
(
because of unemployment, sickness, old age, etc.) can an unpaid worker have
recourse to social protection. However, being regarded as a dependant, the
unpaid worker tends not to receive the money directly, and payment is made to
their breadwinner partner on their behalf.
The gender-based nature of social protection systems has been much dis-
6
cussed but bears repeating here because of its relevance to the adequacy and
reliability of income sources and hence to people’s income security. (The switch
of focus to the income security of people rather than women is deliberate.) The
gender-based nature of social protection systems and the income insecurity to
which this gives rise are rooted in the designation of the unpaid care worker as
a dependant of the paid worker-provider. Unpaid caring work, not the sex of
the unpaid worker, is the clue to understanding why social protection systems
prove a fragile source of income for women.
The lack of income security of unpaid workers in social protection sys-
tems arises in two related sets of circumstance. First, because the partner con-
cerned may be engaged in low-paid, erratic employment, he/she may not pass
on income he/she may have, or may do so only on terms which are unaccept-
able for sexual, emotional or physical reasons. Moreover, income flow from a
partner may be terminated at any time by separation, divorce or death. Second,
the lack of income security for unpaid workers arises because their other source
of income (social protection) imports — through the gender-based structure
outlined above — most if not all of the insecurities associated with the financial
relationship between an unpaid worker and their “breadwinner” partner. The
United Kingdom’s cohabitation rule which provides for withdrawal of benefit
from widows and single parents immediately they form a new sexual relation-
7
ship, irrespective of their financial or other circumstances, is perhaps the clearest
example of the fragility of the support offered to unpaid workers by the social
protection system.
Another aspect to the relation between income security and social protec-
tion and how this produces insecurity for women concerns paid employment.
Even in the present state of fluctuating labour markets, this income source can
produce adequate and reliable incomes for some categories of workers. There
are many, however, for whom periods out of employment and periods in em-
ployment in part-time or low-paid work with little prospect of enhancement
are the norm. For these workers — men or women — paid employment offers
5
Although there are care allowances, short-term maternity and parental leave which
enable paid and unpaid leave to be taken for care purposes.
6
In relation to northern European and north American schemes, but not in relation to
southern European schemes. For a contribution filling the latter gap, see Addis (1999).
7
These circumstances include participation in the New Deal for Single Parents training
scheme, a place on which, in practice if not in accordance with regulations, is immediately lost
along with benefit) if a single parent is held to be cohabiting.
(

1
52
International Labour Review
neither an adequate nor a reliable income source. More to the point, for present
purposes, social protection systems tend to reproduce these insecurities rather
than remove them.
This is because entitlement is conditional, in respect of occupational ben-
efits, on employment with a specific employer and, in respect of many social
insurance benefits, on having been in paid work (in order to be able to pay
contributions). For both occupational benefits and social insurance, entitlement
may be further restricted because it is conditional on beneficiaries being in paid
work for a certain number of hours or earning a certain amount per week. Even
where the existence of a low-earnings threshold does not in effect exclude low-
paid workers from access to schemes, if the method of calculating benefits is
geared closely to earnings their benefit payments will be low. People who move
in and out of employment will have difficulty satisfying the lengthy contribution
periods required to qualify for benefits such as pensions. Intermittent employ-
ment involving periods of low pay or no pay can also affect the amount of retire-
ment pension awarded, depending on thecalculation formula used. Finally, move-
ment in and out of work for different employers makes it difficult for workers to
build up entitlement to occupational benefits, particularly pensions.
This income insecurity in social protection systems is rooted in unpaid
work and in either non-employment or employment in forms of work which
are not full time, relatively highly paid and continuous over a working life. It
follows that these systems tend to provide an unreliable and inadequate source
of income for persons in such situations. Moreover, and this is where women
are re-introduced into the analysis, groups with these unpaid work and precari-
ous labour market characteristics are largely — though not entirely — made up
of women. For women continue to undertake most unpaid caring work, whilst
entering the labour force in increasing numbers. They also continue — more
8
than do men — to move in and out of employment and to dominate part-time,
9
low-paid jobs in many economies, being obliged in many instances to seek
these precarious and peripheral forms of employment because of the demands
10
of their unpaid caring work. This link between unpaid work activity and
participation in precarious forms of employment is crucial to explaining why
social protection systems fail to provide women with an adequate and reliable
source of income.
The next part of the article analyses the relationship between equal treat-
ment and social protection by looking at the various ways in which equal
treatment for men and women in social protection has been defined. One im-
8
In the EU, 83 per cent of part-time workers are women (Commission of the European
Communities, 1998a, p. 15).
9
In 15 countries in the early 1990s, the proportion of women in low-paid employment
was consistently higher than the proportion of men and in some countries considerably so. For
example, Japan 37.2 per cent women, 5.9 per cent men; the United Kingdom 31.2 per cent
women, 12.8 per cent men; Switzerland 30.4 per cent women, 6.8 per cent men; and Germany
2
5.4 per cent women, 7.6 per cent men (OECD, 1996, table 3.2, p. 72).
10
On the economic inactivity and part-time work rates of women and men with children
under 17 years, see Commission of the European Communities (1998a), pp. 4-5.

Equal treatment, social protection and income security for women
153
portant question underpinning the analysis is whether women’s income insecu-
rity in social protection systems is considered relevant to the definition of equal
treatment — whether poverty is regarded as an equality issue.
Equal treatment and social protection
There are at least five variants of the meaning of equal treatment in social
protection. The legal version is that constructed through the practice of legis-
lators and the courts. The other four have been developed by interest groups,
policy-makers and academics, with a view to influencing government policy.
Thefirst of the political notions seeks to end thedisadvantages suffered by women
in social protection systems through non-coverage of their unpaid work. The
second specifically aims for the financial independence of women. The third
regards full social protection rights for women as a means of encouraging their
labour market participation, while the fourth seeks to remove those aspects of
social protection perceived as acting as a disincentive to women’s paid work.
The legal formulation of equal treatment
Many international and national legal instruments espousing the cause of
equality could provide the basis for a discussion of the legal notion of equal
treatment. However, relatively few of these human rights instruments make
specific reference to sex or gender on the one hand and to social protection on
1
1
the other. Equality instruments adopted by European Union (EU) institutions
12
are rare examples of the latter and form the basis of the legal notion of equal
treatment discussed here.13
EU equality law defines the principle of equal treatment in terms of a pro-
hibition on discrimination; more specifically, a prohibition on discrimination on
grounds of sex and marital status. This ban extends to both indirect and direct
forms of discrimination. An example of direct discrimination, once common in
11
The International Covenant on Economic, Social and Cultural Rights, and the Euro-
pean Social Charter are partial exceptions here. The International Convention on the Elimina-
tion of all Discrimination against Women is another partial exception since it specifically men-
tions discrimination against women. Other more general human rights instruments are the
Charter of the United Nations, the Universal Declaration of Human Rights and the International
Covenant on Civil and Political Rights. To complete the picture, regional-based instruments such
as the African Charter on Human and Peoples’ Rights and employment-based instruments such
as ILO Convention No. 156 on the reconciliation of employment and family responsibilities
should also be mentioned.
12
Council Directive 79/7/EEC on equal treatment for men and women in statutory social
security schemes (hereinafter referred to as Directive 79/7); Council Directive 86/378/EEC as
amended by Council Directive 96/97/EEC on equal treatment of men and women in occupa-
tional social security schemes. Despite the existence of the latter Directive, following the Barber
case (Barber v Guardian Royal Exchange Assurance Group (Case C-262/88) [1990] ECR
I-1889), Art. 119 of the EC Treaty (now renumbered Art. 141 by the Treaty of Amsterdam)
prescribing equal pay for men and women is the principal provision governing occupational
benefits, including occupational pensions.
13
It should be noted that this EU-derived legal formulation does not necessarily accord
with interpretations under other international and constitutional instruments.

1
54
International Labour Review
northern European systems but now largely abolished, is the granting of unem-
ployment benefit to married women at a rate lower than that applied to men and
single women (explicitly on grounds of sex and marital status). An example of
indirect discrimination is a rule limitingaccess to benefits to those earning above
a certain threshold or working more than a specified number of hours a week. In
this case, the same rule is applied to all workers, but it has a differential effect,
depending on people’s earnings and weekly hours: the effect is adverse on those
with earnings or hours below the threshold (i.e. low-paid or part-time workers)
since they will be barred from access to benefit. Where a rule of this kind ad-
versely affects more women than men (or men than women), it is said to consti-
tute indirect discrimination on grounds of sex, unless the rule or threshold can be
justified on “objective” grounds, i.e. on grounds other than sex.
The basic idea underlying this somewhat technical concept of discrimina-
14
tion is that people in the same situation should be treated in the same way, and
people in different situations should be treated differently. The ban on direct
discrimination meets the first requirement, to treat like as like. Married women
should accordingly be treated the same as men and single women, i.e. granted
the same, not a lower, rate of unemployment benefit. The ban on indirect dis-
crimination meets the second requirement, to treat unlike as unlike. Accord-
ingly, the same earnings rule should not be applied both to the low-paid and the
highly-paid, since the two groups of workers are not in the same situation.
The treat-like-as-like approach to equality is both elastic and flexible, one
reason being the absence from EU equality legislation of specific criteria for
determining what constitutes the same and what constitutes different situations.
Are persons engaged in paid work in the same situation as those in unpaid
work? Are people with high and low earnings in different situations?Although
15
the EU legislation gives some guidelines, the answers to these crucial ques-
tions have largely been left to the European Court of Justice to determine in the
cases brought before it and to EU Member States when implementing the EU
legislation at the national level.
Another reason for this flexibility is the narrow interpretation given by
the European Court of Justice and others of what is intended by equal treatment
16
in the social protection field. According to this narrow interpretation, equal
treatment has two aims. The first is to ensure that social protection rules treat
people as individuals and do not make assumptions about their characteristics
1
4
“
...women are entitled to be treated in the same manner, and to have the same rules
applied to them, as men who are in the same situation...” Borrie Clarke v Chief Adjudication
Officer (Case 384/85) [1987] ECR 2865.
15
Directive 79/7 applies to the “working population”, thereby seeming to exclude unpaid
workers and the non-employed from the scope of the equal treatment principle embodied in it.
It applies to statutory schemes covering specified risks — such as unemployment, old age,
sickness — traditionally associated with paid employment; caring is not included. These provi-
sions covering the personal and material scope of Directive 79/7 would seem to imply limitation
of the Directive and the equal treatment principle to paid work.
1
6
The EU equality instruments do not assist a great deal. Directive 79/7 merely states the
purpose of the Directive to be “the progressive implementation of equal treatment between men
and women in matters of social security”.

Equal treatment, social protection and income security for women
155
17
on the basis of their membership of a particular group. This provided the
justification for abolishing the different, and adverse, treatment of married
women in social protection schemes, since the adverse treatment involved the
assumption that as a group married women did not engage in paid work and
were financially supported by their husbands — an assumption that proved to
be incorrect in many cases, given married women’s increasing labour market
1
8
activity. A second aim is to ensure that the law relating to social protection
accurately reflects socio-economic reality. This argument was also used to jus-
tify abolition of the adverse treatment of married women on the basis that the
rationale for the treatment (married women’s non-participation in the labour
market) was an increasingly outmoded assumption no longer in accord with
economic reality (see Hoskyns and Luckhaus, 1989).
Seeking to promote the integrity of the individual and to safeguard the
integrity of the law are not unimportant aims, but they do not relate specifically
to social protection, nor do they seek to alter the pattern of existing financial
and social relationships (unlike some of the political notions of equal treat-
ment). Furthermore, they are largely procedural in nature. As a result, when
major substantive questions have arisen, e.g. whether equality must be imple-
mented so as to avoid increasing poverty or whether Member States may be
required to alter their social policy in order to achieve equality, they have been
determined by the European Court of Justice in a negative way.
A narrow interpretation of the purposes of equal treatment in the social
protection field was not inevitable. It was open to the European Court of Jus-
tice to adopt a more expansive interpretation, using some of the political no-
tions as a guide. And in a different political and economic climate, it might
well have done so.19
Political notions of equal treatment
In their current forms, the political formulations of equal treatment emerged
in political debates at European and national level from the late 1960s onwards.
The first of these notions, calling for an end to the financial disadvantages
suffered by women in social protection systems, arose out of the detailed and
rigorous analyses of these systems carried out by policy-makers, women’s groups
and academics. This analysis linked women’s disadvantaged position in social
protection systems to their unpaid work and to their (often connected) partici-
pation in the sort of part-time, low-paid, intermittent employment poorly cov-
ered by social protection systems. For this reason, the demand for equal treat-
17
See, for example, the Advocate General’s opinion in Coloroll Pension Trustees Limited
v Russell and Others (Case C-200/91) [1994] ECR I-4389.
18
See, for example, the Advocate General’s opinion in Caisse d’assurances sociales pour
travailleurs indépendants ‘Integrity’ v Nadine Rouvroy (Case C-373/89) [1990] ECR I-4243.
1
9
In its 1986 judgment in Drake v Chief Adjudication Officer ((Case 150/85) [1986] ECR
1
995), the European Court of Justice by some clever reasoning managed to bring a benefit for
caring within the material scope of Directive 79/7, so enabling the exclusion of married women
from that benefit to be held unlawful; see Luckhaus (1986).

1
56
International Labour Review
ment was specifically framed in terms of equality for women. The idea that
men would need to invoke the equality principle was not conceivable within
that analytical framework. The focus on women’s disadvantaged position within
social protection also implied a close link between this notion of equality and
the avoidance of poverty. This link in turn presupposed that the disadvantages
would be removed and equality secured by “levelling up” women’s position so
that they would enjoy the same degree of income security within social protec-
tion systems as men do. Of all the variants of equality discussed here, this
formulation gives the greatest priority to securing income security for women.
The second political formulation of equal treatment, the claim for
women’s financial independence, is again specific to women but the emphasis
is on ending the institutionalization of female unpaid workers’ financial de-
pendence on their male breadwinner partners. Calling as it does on women to
end their financial dependence on men by securing their financial independ-
ence through employment, thisclaim concerns more than social protection alone.
This broader claim also incorporates a narrower one (specific to social protec-
tion), namely, that the rules and concepts determining entitlement within these
systems should be restructured around an individualized model of entitlement
rather than the relationship between breadwinner/paid worker and dependant/
unpaid worker. In short, this notion of equal treatment seeks to remove the
concept of dependency between partners from social protection systems.
The third and fourth political formulations of equal treatment both give
priority to promoting women’s paid employment and both regard social pro-
tection systems as playing an instrumental role in this respect, but the mecha-
nism is different in each case. The third version sees the guarantee to women of
full rights to social protection as a way of encouraging them to participate in
the labour market in order to secure these rights. Here, social protection is seen
as a financial incentive to paid work. This notion has its origin in the late 1960s
and early 1970s when there was concern in the EU about labour shortages and
their effect on national economies. Interestingly, a variant of this argument
focusing on part-time work has recently surfaced in Spain where there is con-
cern in some quarters about both the relative absence of part-time forms of
employment and women’s low labour market participation. The argument is
being made that full social protection rights should be extended to part-time
work in order to encourage non-employed women to take up part-time work;
see European Industrial Relations Review (1999).
The fourth political formulation of equal treatment concerns the ways in
which social protection systems operate so as to discourage women from taking
or remaining in paid employment. Social protection is therefore seen as a dis-
incentive rather than an incentive to women’s labour market participation. Con-
cern about the disincentive effects of social protection emerged, in the United
Kingdom at least, in the late 1970s and 1980s and initially focused on men
rather than women. It coincided with a macroeconomic concern to lower wages
and social charges in order to improve competitiveness and reduce unemploy-
ment. In the 1990s, as preoccupation with work incentives grew in the United
Kingdom, attention turned to women whose partners were unemployed and

Equal treatment, social protection and income security for women
157
claiming social assistance; it was said that women in such situations left the
labour market or remained non-employed because it was not financially worth
holding on to what could only be — if the argument is to ring true — low-paid
employment. One report typifying this preoccupation with the perceived disin-
centive effect of social protection systems on women’s employment explicitly
linked it to the issue of women’s independent income (see Duncan, Giles and
Webb, 1994).
This seemingly innocuous link needs to be examined closely. Though such
a task is beyond the scope of this article, two points may be made. One is to
note statisticians’ growing preoccupation with increasing numbers of house-
2
0
holds composed of jobless persons of working age in the United Kingdom
21
and many industrialized countries. At one level, the preoccupation arises from
concern for the material welfare of people in these households (OECD, 1998a,
p. 1); at another, it reflects a desire to identify “labour market slack” (OECD,
1
998a, p. 1) and to have data offering some support for the otherwise specula-
tive proposition that social protection systems discourage the female partners
22
of men in receipt of benefits from getting (low-paid) jobs. The second point
is that “women’s independent income” is likely to be very different from that
intended in the claim for financial independence. Since the preoccupation with
disincentives coincides with a concern to reduce wages and social charges (and
hence with removing people from receipt of benefit), the reference to “inde-
pendence” in this context probably has more to do with non-reliance on social
protection as an income source than with non-reliance on a partner (as is the
case with the claim for financial independence). Moreover, given the relatively
low rates of benefit (at least in the United Kingdom), the logic of the disincen-
tives argument strongly suggests that this independent income will almost cer-
tainly be derived from forms of part-time, low-paid and intermittent employ-
ment which provide little, if any, income security for women.
Implications for equal treatment of recent
developments in social protection
The various formulations of equal treatment described above provide the
criteria for assessing whether, and to what extent, existing social protection
systems may be said to be discriminatory. The first part of the assessment deals
20
In spring 1999, some 17.2 per cent (3.2m) of all working-age households had nobody
in work (Cooper-Green, 2000, p. 28).
21
See OECD (1998a), p. 21. It should be noted that much of the increase is attributable to
a growth in households consisting of a single jobless adult, but statistics also show the share of
households consisting of two or more jobless adults to be either static or increasing slightly in
some countries. The report also makes specific mention of the particular situation of women
in the latter type of household.
22
The OECD report on employment illustrates how this can be achieved indirectly, by
explaining the preponderance of what it calls “work-poor” households in terms of “the disin-
centive effects arising from the interactions of the tax and benefit systems”; see OECD (1998a),
p. 21.

1
58
International Labour Review
with aspects of systems reflecting the concept of dependency and the breadwin-
ner/dependant model of entitlement. The second deals with provisions within
these systems that privilege full-time, relatively well-paid, continuous employ-
ment patterns and disadvantage more precarious forms of employment and
unpaid work. Reference is also made in this part to provisions within a social
protection system which may operate to mitigate the adverse effects of the bias
elsewhere in the system against precarious employment and unpaid work. The
third part deals with two sets of social protection rules — relating to pension
age and life expectancy — which raise sex and gender issues.
Social protection and the concept of dependency
This section examines three sets of social protection provisions, all linked
through the concept of financial dependency between adult partners (hereafter
referred to simply as “dependency”). They are widow’s (and survivor’s) ben-
efits, social assistance schemes and the rules providing for pension sharing on
divorce. Widow’s benefits are structured around dependency because they link
benefit entitlement to contributions by the deceased spouse, because they insure
against the loss of a breadwinner-provider, and because they may be withdrawn
(
as in the United Kingdom) if the recipient enters a new sexual relationship.
Survivor’s benefits are here taken to be benefits granted on the death of a spouse,
i.e. to widowers as well as widows. They are based on the same principles as
widow’s benefit but are not sex-specific. The second set of social protection
provisions, social assistance schemes, are structured around dependency in so
far as entitlement is determined according to the resources and needs of a unit
comprising a man and a woman in a sexual relationship. The third set of provi-
sions enabling pension sharing on divorce presuppose dependency between the
two members of a formerly married couple. They also presuppose that the
acquisition of pension rights is uneven between spouses and that the lesser or
non-existent rights of the wife will be lost if the marriage is terminated by
divorce prior to the retirement or death of the husband.
Widow’s and survivor’s benefits
Widow’s benefits are a long-standing feature of social insurance schemes
and of occupational schemes. Two recent developments have been the intro-
duction of survivor’s benefits in both statutory and occupational schemes and
the conversion of some statutory schemes from entitlement based only on con-
tributions to entitlement based on means testing or a mix of means testing and
contributions. The introduction of statutory survivor’s benefit is well advanced,
at least in the EU. Periodic payments of this kind now exist in all EU Member
States except Denmark. In 1992, Denmark abolished the statutory periodic
payment to widows and replaced it with the more limited provision of a lump-
sum payment to both widows and widowers on the death of their spouse. The
United Kingdom is the last EU Member State to remove the special treatment
accorded to widows in statutory schemes and is doing so by extending provi-
sion — including periodic payments — to male survivors on the same terms as

Equal treatment, social protection and income security for women
159
those applicable to female survivors. The 1999 legislation abolishing widow’s
benefit and inserting the new system of bereavement benefits is due to come
into force in April 2001. As regards occupational schemes in the EU, there is
some evidence pointing to extension of widow’s benefit to widowers, in the
process of reform rendered necessary by a European Court of Justice ruling on
the discriminatory nature of occupational widow’s benefit in 1993.23
The second development concerning statutory survivor’s benefit involves
its conversion to a means-tested benefit. Governments in France, Italy, Sweden
and the Netherlands have all introduced an element of means testing into their
schemes (United Kingdom, 1998a, chapter 3, para. 7). The United Kingdom’s
1
999 reforms adopted a similar approach. In brief, the widow’s pension pay-
able on a periodic basis to women aged 45 and over for as long as they remain
widows or do not cohabit has been replaced by a bereavement allowance avail-
able to both male and female surviving spouses aged 45 and over. However,
this allowance is restricted to 12 months’ duration. Thereafter, widows and
widowers will be able to claim means-tested income support but will only gain
entitlement to this benefit if, and so long as, they can demonstrate insufficient
income, availability for work and the other conditions applicable to this ben-
efit. In the long term, these new rules are expected to lead to a significant
reduction in the level of provision for older widows (United Kingdom, 1998a,
24
chapter 3, para. 5).
An assessment of the equal treatment implications of widow’s benefit and
the recent developments in their regard is appropriate here. In principle,
widow’s benefit offends the legal notion of equal treatment since it directly
discriminates against men. The restriction of this benefit to female survivors in
occupational schemes was held unlawful by the European Court of Justice in
1
993, as already noted. This benefit continues to be lawful in EU statutory
social protection schemes by virtue of the exemption of the broader category of
survivor’s benefit under EU equality law. Survivor’s benefits under occupa-
tional schemes are non-discriminatory since they treat both male and female
25
23
Ten Oever v Stichting Bedrijfspensioenfonds voor het Glassenwassers-en Schoonmaak-
bedrijf (Case C-109/91) [1993] ECR I-4879. The French AGIRC and ARRCO occupational
schemes introduced survivor’s benefits in 1994 and 1996 respectively. For trends in occupa-
tional schemes in the United Kingdom, see Luckhaus and Moffat (1996), pp. 78-79.
24
Men and women aged 55 or over when the legislation comes into force and who are
widowed in the subsequent five years will be able to claim income support when their bereave-
ment allowance expires without being available for work. They will also receive an additional
amount of income support bringing that payment up to the level of the bereavement allowance.
This transitional provision will ensure that widows do not lose out financially with the introduc-
tion of the bereavement allowance, but this protection is not afforded to those under 55 when
the legislation comes into force and who are subsequently widowed.
25
Art. 3(2) of Directive 79/7. The United Kingdom’s decision to reform widow’s benefit
was prompted in part by legal proceedings begun in 1997 in Willis v UK (application No.36042/
7 to the European Commission on Human Rights). The absence of widower’s benefit was
9
challenged as contrary to the ban on sex discrimination in the European Convention on Human
Rights, there being no exemption of survivor’s benefit in this instrument, as there is in Directive
7
9/7.

1
60
International Labour Review
survivors the same, as dependants. Statutory survivor’s benefits also satisfy this
non-discrimination principle (subject to the terms being identical for men and
women, which is not always the case), but their discriminatory status is not a
live issue since, as already noted, they are exempted from application of equal
26
treatment principle by EU equality law. The fact that the means testing of
survivor’s benefit can result in reduced social protection for some widows raises
in stark form the question of whether, according to EU law, equality between
men and women can be implemented by reducing provision for some or all
groups: in short, whether equality can be implemented in a way which may
result in poverty for some or all of those concerned.
The levelling down of provision in the process of implementing equality is
a common strategy in both statutory and occupational schemes. In the 1970s and
1
980s, in the first cases to come before it in which this point was raised directly
or indirectly, the European Court of Justice either proceeded on the assumption
that the result of a successful discrimination claim would lead to extension of
provision to the disadvantaged group or simply avoided the issue. It then
developed a neat formulation of the equal treatment principle, the consequence
of which was automatic extension of the privileged treatment to the disadvan-
taged group. However, when pressed to rule on whether Member State legisla-
tures could implement equality by reducing provision, it responded positively,
expanding on the idea, developed elsewhere in its case law, that the practical
implementation of equal treatment was a matter of social policy over which
27
28
2
9
30
Member States continued to retain considerable discretion and control.
To summarize the legal situation, widow’s benefits are discriminatory
whereas survivor’s benefits are not, although statutory widow’s benefit is ex-
empted under Directive 79/7 from application of the equal treatment principle.
In addition, EU Member State implementation of equal treatment involving a
reduction in provision for some or all groups previously entitled to benefit is
2
6
Art. 3(2) of Directive 79/7. All exemptions under Directive 79/7 are due to be abolished
as a result of the “progressive implementation of the principle of equal treatment”. A proposal
for a draft directive aimed at completing equal treatment in statutory schemes was introduced in
1
987 (COM(87) 494). This proposal suggested exemptions should be abolished by individual-
izing entitlement in preference to extending dependency-related provisions to men. The pro-
posal was amended heavily in negotiations, and was finally shelved.
27
In the 1970s’ cases on equal pay under Art. 119 of the EC Treaty (renumbered Art. 141
by the Treaty of Amsterdam). The idea that women would take an equal pay claim to secure
reduction of men’s pay to theirs did not appear to be within the contemplation of the European
Court of Justice nor, arguably, of anyone else.
28
Teuling v Bedrijfsvereniging voor de Chemische Industrie (Case 30/85) [1987] ECR
2
497.
2
9
Women were to have the same rules applied to them as men in the same situation since,
where the Directive has not been implemented, those rules remain the only valid point of
reference: see Borrie Clarke v Chief Adjudication Officer (Case 384/85) [1987] ECR 2865.
3
0
Queen v Secretary of State for Health ex parte Richardson (Case C-137/94) [1995] ECR
I-3407 in respect of statutory schemes and Smith and Others v Avdel Systems Limited (Case
C-408/92) [1994] ECR I-4435 in respect of occupational schemes.

Equal treatment, social protection and income security for women
161
permissible under EU law. As to the political notions of equal treatment, wid-
ow’s benefit goes some way to meeting the call for an end to the disadvantages
women suffer in social protection systems through non-coverage of their un-
paid work activity, although it does so by focusing on sex and marital status
rather than on engagement in unpaid caring work (or indeed on participation in
the peripheral sections of the labour market). The extension of widow’s benefit
to male survivors is not compatible with this notion of equal treatment, nor is
the levelling down of provision which has tended to accompany that process.
The existence of widow’s benefit runs directly counter to the ideal of ending
dependency in social protection systems, a situation exacerbated by the exten-
sion of this status to men in the form of survivor’s (widower’s) benefit. Finally,
the application of means testing to survivor’s benefit may raise the prospect of
increasing disincentives to work for women, contrary to the notion that equal
treatment requires their reduction.31
Social assistance
The application of means testing to survivor’s benefit is indicative of an-
other development in social protection with implications for equal treatment,
namely the expansion of social assistance. A study of the 24 OECD countries
showed that social assistance is becoming more important in nearly all the
countries, in terms both of expenditure and claimant numbers (Eardley et al.,
1
996, p. 178). The osmotic spread of means testing to nearly all forms of social
protection including in many non-OECD countries is illustrated by its applica-
tion to contribution-based family allowances in Italy, the introduction in the
United Kingdom in the early 1990s of a means-tested in-work benefit for per-
32
33
sons with disabilities, and the announcement in Bangladesh in 1998 that a
means-tested pension of 100 takas (about 2 US dollars) per month was to be
awarded to some 90 people in each of the country’s 4,479 districts. State social
assistance programmes for the elderly are also run in other parts of Asia, nota-
bly India; see ILO/EASMAT (1997), p. 24. A feature common to these pro-
grammes, to all the schemes operated in OECD countries and, most likely, to
social assistance everywhere, is the testing of means on the basis of a unit larger
than the individual seeking entitlement to the benefit. In some instances, the
unit may be very large and embrace distant as well as near relatives. In the
majority of OECD countries, the unit is restricted to a claimant and his/her
spouse (Eardley et al. 1996, p. 65) and, in some cases, to a claimant and the
(hetero)sexual partner with whom he/she is held to be cohabiting.
3
1
This point is discussed below in relation to social assistance schemes. It has also been
suggested that widow’s benefit itself provides a disincentive to work since it enables a woman
not to have to take paid work if her husband dies.
32
First introduced in 1934, the assegni familiari have had an uneven history, but since
994 have been favourably regarded and expanded in scope by successive Italian administra-
tions; see Addis (1999).
1
33
Called disability working allowance and now replaced by a very similarly structured
benefit called a disability tax credit.

1
62
International Labour Review
Potentially, the implications for equal treatment of this spread of social
assistance and the linking of entitlement to these benefits to the resources of
both partners are quite far-reaching. The root of the problem with means test-
ing based on a couple’s resources is that where women’s earnings are consider-
ably lower than men’s, as is customarily the case, more women than men fail to
establish entitlement to means-tested benefits because their partner’s earnings
bring the couple’s combined resources above the minimum income allowed.
Moreover, where a male partner succeeds in establishing entitlement (aided by
the low or non-existent earnings of his partner), the element of the benefit
intended to meet his partner’s needs is paid to him and not to his female partner.
As to the discriminatory status of the dependency notion in social assist-
ance, a number of cases have come before the European Court of Justice seek-
ing to establish that taking into account a partner’s income in means-tested
schemes constitutes indirect discrimination against women. Acknowledging its
adverse effect on women’s entitlement, the Court none the less ruled that the
practice was not discriminatory because it could be objectively justified on the
grounds that provision of a minimum income was an integral part of Member
States’ social policy and that Member States had a “reasonable margin of dis-
cretion” in determining the nature of social protection measures and the actual
34
means by which they were to be implemented. As with its application in
relation to levelling down by Member States, this formula nicely illustrates
how easily the abstract form of the “treat-like-as-like” approach to equality can
be moulded into the proposition now clearly established in this area of EU law
that poverty and material well-being are irrelevant to the concept of equality, as
interpreted by the Court.
As to compatibility with the political notions of equal treatment, social
assistance schemes themselves can be a welcome lifeline for a woman who is not
entitled to other forms of social protection because of her unpaid work activity.
These schemes do not actually conflict with the call for an end to the disadvan-
tages women encounter in these other spheres of social protection as a result of
their unpaid work. However, the treatment of awoman’s partner’s (usually higher)
income as hers tends to displace if not destroy the potential of these schemes to
alleviate her income insecurity. At the same time, the income assessment rules
bring the schemes into conflict with the equal treatment ideal of ending disadvan-
tages for women arising from their unpaid work activity. In addition, the pay-
ment of money to a man to meet his partner’s needs runs counter to the principle
of equality which regards the financial independence of a woman from her part-
ner as paramount. Finally, means-tested schemes are a prime target for advocates
of the notion that equal treatment requires an end to the disincentives to work
embedded in social protection schemes. This argument goes as follows: the ten-
34
Commission v Belgium (Case C-229/89) [1991] ECR I-2205; Moelenbroek v Bestuur
van de Sociale Verzekeringsbank (Case C-226/91) [1992] ECR I-5943. These rulings have the
effect of taking poverty-related benefits outside the scope of Directive 79/7 and the equal
treatment principle, a point specifically decided by the Court in relation to the United King-
dom’s income support in Jackson and Cresswell v Chief Adjudication Officer (Case C-63-64/
9
1)[1992] ECR I-4737.

Equal treatment, social protection and income security for women
163
dency for a woman’s earnings to be low (or non-existent) permitting her (unem-
ployed, disabled, older) partner to qualify for means-tested benefits results in a
disincentive to work for women in (low-)paid employment, because the couple is
35
generally not much better off if the woman does so.
Pension sharing on divorce
A third development in social protection serving to extend further and
consolidate the principle of dependency in social protection systems is the prac-
tice of dividing the pension rights of spouses on divorce. In principle, pension
sharing (or pension splitting, as it is sometimes known) is applicable to all
types of pension schemes and has a fairly long history in statutory schemes. It
was introduced in Germany in 1977 (Solcher, 1978, p. 308) and in Canada in
36
the following year. It has attracted attention recently in relation to occupa-
tional schemes. Pension sharing in occupational schemes was introduced in
Ireland in 1996. It was proposed in the United Kingdom in 1997 and in South
Africa in 1998. In the United Kingdom, the proposal became law in 1999 when
legislation was passed providing for the transfer of pension rights from one
spouse to another on divorce. The legislation supplements existing arrange-
ments which can be made by British courts, but which fall short of outright
transfer of rights; the main beneficiaries of the 1999 legislation are expected to
be women (United Kingdom, 1999). In South Africa, the Law Commission’s
1
998 consultation paper (South African Law Commission, 1998) suggested
pension sharing might be extended to all spouses, including those entering
marriage under customary law. Its recommendation following consultation,
however, is to restrict pension sharing to marriages recognized as such by exist-
ing law (SouthAfrican Law Commission, 1999). In the United States, a variant
of pension sharing (called earnings sharing) is being actively promoted by the
Cato Organization as part of a plan to abolish the United States social security
system and replace it with a fully funded, defined contribution system financed
solely by the individuals receiving the benefit (Shirley and Spiegler, 1998; see
also Olsen, 1999).
If pension sharing on divorce is available to men and women on the same
terms, it seems to be compatible with the legal notion of equal treatment.37
35
This is because, in their turn, the women’s earnings are deducted from the benefit that
would otherwise have been awarded. Whether or not people make employment decisions exclu-
sively on monetary terms is a moot point, not least because women might decide to take
employment in order to gain direct access to the money. If one accepts the logic of the disincen-
tives argument, it can only operate in relation to work which is low paid. Women (or men)
motivated solely by money are not going to give up the financial rewards of highly paid work
for the relatively low levels of income granted to their partners under social assistance schemes.
36
O’Neil and Ciffin, 1978, pp. 64-77. For details of pension sharing in some EU Member
State schemes, see Luckhaus (1994), pp. 147-161. Under the US social security system, a
divorced woman is entitled to benefits equal to 50 per cent of her former spouse’s benefits
provided the marriage lasted ten years; see Shirley and Spiegler (1998).
37
Although it will undoubtedly benefit more women than men and, by the same token,
adversely affect more men than women, it would presumably be held to be objectively justifi-
able.

1
64
International Labour Review
Since the aim of pension sharing is to secure a fair division of the wealth
acquired during a marriage on the ground that both spouses have contributed to
that acquisition through the sharing of paid and unpaid work, pension sharing
also seems to conform with the call for an end to disadvantages in social protec-
tion systems arising from the non-coverage of unpaid work. Pension sharing is
sometimes misguidedly regarded as a means of individualizing benefit rights
(
establishing the financial independence of one partner from the other). But
pension sharing proceeds on the basis of one spouse being financially depend-
ent on the other and so extends and reinforces the principle of financial depend-
ency throughout social protection systems.
In summary, then, financial dependency between partners provides the
analytical link between the social protection developments described so far in
this article. The striking feature about these developments is the extent to which
this form of financial dependency is spreading in both statutory and occupa-
tional schemes. Recent attempts at EU level to achieve recognition of homo-
sexual (cohabiting) relationships as equivalent to heterosexual (cohabiting) ones
in order thereby to secure for the dependent homosexual partner the financial
advantages awarded to the dependent partner in a heterosexual relationship
indicate a readiness on some people’s part to consider further expansion of this
38
form of dependency in the social protection schemes of the future.
Social protection regulations and different forms of
employment
The rules governing the transmission of income security from the labour
market to the social protection system are examined in this section, as are rules
within the social protection system capable of countering the adverse effects of
these labour market rules to some extent.
The labour market rules are: obligation to be in paid work, satisfaction of
an earnings or hours threshold, entitlement linked to payment of contributions
38
In the United Kingdom case of Grant v South-West Trains Ltd (Case C-249/96) [1998]
CMLR 993, the applicant employee challenged the failure by her employer to grant her female
1
partner travel concessions granted to married or cohabiting heterosexual partners of employees,
as a violation of the equal pay principle under Art. 119 of the EC Treaty (renumbered Art. 141
by the Treaty of Amsterdam). The European Court of Justice ruled against this being an instance
of sex discrimination but pointed favourably to the new Art. 13 of the EC Treaty (as renumbered
by the Treaty of Amsterdam) enabling the EU Council to adopt measures to eliminate discrimi-
nation on the grounds of sexual orientation. The implications of (ultimately) requiring the same
treatment of homosexual and heterosexual couples in relation to dependency in social protec-
tion systems could be far reaching. It seems Belgium already takes into account the income of a
homosexual partner for the purposes of assessing entitlement to means-tested supplements: see
Royal Decree of 4 August 1996 (Commission of the European Communities, 1997, p. 47). This
extension of dependency leads to savings in expenditure on social protection and a loss to the
individual who might be deprived of benefit because their partner’s earnings are taken into
account in assessing that individual’s entitlement and to the couple (the couple rate being less
than two individual rates). Extending survivor’s benefits to homosexual partners and permitting
pension sharing on “divorce” (to name just two of the possibilities) would, however, increase
expenditure in occupational as well as statutory schemes, a reason perhaps why there may be
some reluctance to extend dependency in this sphere.

Equal treatment, social protection and income security for women
165
over a lengthy period, benefit amounts linked to earnings in work and lengthy
contribution periods. These rules feature in some statutory and occupational
39
schemes but not in others. The earnings and hours thresholds and the lengthy
40
contribution conditions in statutory schemes are discussed here, partly be-
cause of recent or proposed changes and partly because their potentially dis-
criminatory status has been the subject of legal challenge.
An earnings or hours threshold is a fairly common feature of social pro-
tection systems. A comparison by the OECD of provision in 21 OECD coun-
tries published in 1998 showed that the systems in all but five of those countries
—
Greece, Hungary, New Zealand, Norway and Spain — operated either an
41
hours or an earnings threshold or a combination of both. Recent develop-
ments in relation to these thresholds were mixed. Some involved a reduction in
the hours or earnings threshold and thus raised the possibility for part-time or
low-paid workers of gaining access to the systems. Thus, in 1994, the threshold
applicable to the social protection system in Japan was reduced from 30 to 20
hours a week and in Canada earnings and hours thresholds previously applying
to unemployment benefit were abolished from 1997 (OECD, 1998a, p. 172).
In Germany, in 1997 too, the first step was taken to reduce the hours and
earnings thresholds governing access to the pension, health and unemployment
insurance schemes. This involved reducing the 18-hour threshold, below which
people could not gain access to the unemployment insurance scheme. Thence-
forth, people could participate in all three schemes — pension, health and
unemployment — if they worked 15 or more hours a week or earned more
than the minimum earnings threshold. This threshold is fixed at a low level
(
630 DM per month in 2000) so, in practice, it is this earnings threshold rather
than the hours threshold which determines who can participate in the insurance
schemes. InApril 1999 further changes to the German pension insurance scheme
were introduced related to this minimum earnings threshold. The changes are
complex but the most important element, for our purposes, is the partial re-
moval of the minimum earnings (630 DM) threshold in relation to pension
insurance. Henceforth, employers are obliged to pay contributions of 12 per
cent on earnings below this level, while employees (earning between 300 and
6
30 DM per month) may make voluntary contributions of 7.5 per cent. Em-
ployees who do not make the voluntary contributions acquire less than the full
3
9
See Luckhaus and Ward (1997), pp. 237-253, for a detailed analysis of rules in statu-
tory and occupational schemes and of how they interact with different forms of employment in
EU Member States.
40
Discussion of their incidence in occupational schemes is hampered by lack of data.
Some information concerning hours thresholds in the United Kingdom is to be found in Luckhaus
and Moffat (1996), pp. 81-85. Cases concerning hours thresholds in Germany and the Nether-
lands continue to be referred to the European Court of Justice, which suggests that such thresh-
olds still operate in occupational schemes in those countries.
41
OECD, 1998a, pp. 170-172. The report referred to the three benefits — public health,
old age and unemployment — said to be the most likely to operate thresholds.

1
66
International Labour Review
range of pension rights. The idea behind these changes was to help low-paid
42
female employees improve their social protection in old age.
Other developments concerning thresholds have been lessfavourable to low-
4
3
paid workers. Thus, in Finland in 1996 an income threshold was introduced
44
into the sickness insurance scheme for the first time. In the United Kingdom,
the lower earnings limit is gradually being raised to bring it into line with the tax
4
5
threshold, the possible consequence being an expansion of the 2.6 million peo-
46
ple presently excluded from statutory social protection by that threshold.
Rules linking pension entitlement to a specified number of years of con-
tribution have also undergone change, often as part of a general package of
government reforms designed to reduce social protection costs (and budget
deficits). In general, changes are likely to restrict access to pensions for people
with a pattern of intermittent employment. Thus, in Sweden the contribution
period required to achieve full pension entitlement was raised from 30 years to
total lifetime contributions; while, in Italy, the ability of private-sector workers
to obtain a full pension after 35 years of contribution and public-sector workers
after 20 years has been removed and replaced by a condition requiring lifetime
contributions (Pierson, 1999, p. 25). Changes to pension earnings formulas in
47
these two countries may also result in reduced pensions for this group. In
42
In addition the reform was aimed at avoiding the creation of jobs with pay set at just
below the minimum earnings threshold (Germany, 1999). This was thought to have been
encouraged by the absence of a requirement on employers and employees to pay contributions
on employment below the minimum earnings threshold, a 20-per-cent tax being levied on
employers in respect of this employment instead. The 20-per-cent tax has been abolished now
that employers are obliged to pay contributions on employment below the 630 DM threshold.
43
It should be noted that social protection was excluded from the scope of EU legislation
Directive 97/81) providing for an end to discrimination against part-time workers.
(
44
Kuhnle, 1999, p. 17. In 1994, 8.7 per cent of women, compared with 3.3 per cent of
men were in low-paid employment in Finland (OECD, 1996, table 3.2, p. 72). This difference
between men and women might not be sufficient to establish disproportionate adverse effect for
the purposes of an indirect discrimination claim under EU law.
45
The two-fold aim being (according to the new Labour Government’s 1998 Green Paper
on welfare): “to increase the incentives for employers to take on staff and to make work pay for
all employees” (United Kingdom, 1998b, para. 31, p. 30). The first step in this process was the
raising of the limit for payment of employers’ contributions to the tax threshold, with effect
from April 1999. The second step is the raising of employees’ contributions over two stages, so
that by April 2001 the level of earnings at which contributions must be paid will be fully aligned
with the tax threshold.
46
Of the 2.6 million, 2 million are women and 0.6 million are men; 70 per cent of the men
are under 25 (McKnight, Elias and Wilson, 1998, p. vi). It should be noted that the United
Kingdom Government has taken steps to try to prevent an increase in the numbers already
excluded by the lower earnings limit by providing that people earning above this limit but below
the new (tax) threshold will continue to qualify for contributory benefits by assuming “notional
payments”.
47
In Sweden, prior to reform the amount of pension was calculated on earnings in the best
5 years and after reform on average earnings over the entire working life. The “best” earnings
1
formula enables years of no or low-paid work to be excluded from the pension calculation and
in this way to benefit intermittent and low-paid workers. Italy has modified its pension formula
from last five and last one year (private and public sector workers, respectively) to an average
of the entire working life. Other European states using the final salary or best earnings formula
are also moving in this direction.

Equal treatment, social protection and income security for women
167
France, Portugal, Austria and Finland, the number of years of earnings on
which the pension is based has been increased (Commission of the European
Communities, 1998b, p. 24). In the United States, there is concern that a pro-
posal to raise the number of years of covered earnings from 35 to 38 in order to
gain full entitlement to a pension will have a disproportionate effect on women,
the estimate being that fewer than 30 per cent of women retiring in 2020 will
have the required 38 years (compared with nearly 60 per cent of men) (Shirley
and Spiegler, 1998). In Spain, however, there has been a slight improvement in
the number of years part-time workers are required to contribute in order to
gain full entitlement to a pension. An agreement between the Spanish Govern-
ment and Spanish trade unions signed in October 1998 resulted in a change in
the method of calculating the contribution years of part-time workers, giving
them entitlement to a pension after 21 years rather than 30. Full-time workers
continue to gain entitlement after 15 years. As already noted, one aim of this
reform is to encourage women to take up part-time work (European Industrial
Relations Review, 1999, pp. 30-31).
The question needing to be addressed is whether these labour market rules
are compatible with the legal and political notions of equal treatment outlined
earlier. As regards the legal construction of equality, it is possible that these
rules are incompatible with the requirement of equal treatment if a case can be
made that they are indirectly discriminatory and that they are not objectively
justifiable. Since the low-paid, part-time, intermittently employed workers
adversely affected by the rules tend to include more women than men (because
of their unpaid work activity), a discriminatory claim may have some chance
of success, particularly since the European Court of Justice generally seems
prepared to regard part-time (low-paid) workers as being in a different situa-
tion from full-time (higher-paid) workers and has not always demanded that
detailed statistical evidence needed to establish differential effect be produced
in support of the claim. However, the availability of the defence of objective
justification gives the Court at EU level and national levels considerable flex-
ibility if it wishes to reject the claim. The discriminatory status of these labour
market rules, therefore, remains unknown until a claim is made and the Court
gives its ruling. In the meantime, though the rules may be regarded as poten-
tially discriminatory they are none the less lawful.
There is some certainty with respect to the rules excluding part-time and
low-paid workers from access to social protection schemes. Here, as with
widow’s benefit, the legality of the rules differs as between statutory and occu-
pational schemes. In occupational schemes, the European Court of Justice took
the view that an hours threshold excluding part-time workers from access to a
48
pension is indirectly discriminatory and cannot be objectively justified. In
two cases concerning the mix of hours and earnings thresholds existing, prior
to 1997, in theGerman unemployment, invalidity and old-age insurance schemes,
the Court held that their discriminatory effect was objectively justified. The
48
Vroege v NCIV Instituut voor Volkshuisvesting BV and Pensioensfonds NCIV (Case
C-57/93) [1994] ECR I-4541.

1
68
International Labour Review
justification accepted by the Court was that employers should be relieved of the
liability of paying contributions on this form of employment (employment of
fewer than 18 and 15 hours a week, in which earnings fell below the minimum
specified), so that they would be encouraged to create the type of part-time,
49
low-paid jobs covered by these thresholds. It may be noted that the Court is
here accepting that equality is not only restricted to those engaged in paid (as
opposed to unpaid) work, but also to those engaged in the type of employment
—
schemes.
full-time, higher-paid — already privileged within social protection
5
0
So far as the four political notions of equal treatment are concerned, two
call for particular comment in the context of these social protection rules
privileging full-time, higher-paid continuous employment. Since these rules
are a major cause of the income insecurity suffered by women in social protec-
tion systems — women’s unpaid work being a reason for their non-employ-
ment or engagement outside this privileged realm — they are in direct conflict
with the notion of equal treatment which calls for an end to the structural
features of social protection systems which disadvantage unpaid work activity.
To a lesser extent, these structural features are incompatible with the idea that
full social protection rights should be accorded to women to encourage them to
take up paid employment, since in this way they will gain access to those rights.
It could be argued, for example, that knowledge of the lengthy period of con-
tribution and relatively high earnings needed to qualify for a pension, com-
bined with a realization that caring activity will prevent those requirements
being met, may prevent the promise of such a pension acting as an incentive to
women to enter the labour market.
Attention will now briefly be given to another set of rules some of which,
despite the austerity currently ruling European statutory social protection sys-
tems, are being preserved and, it seems, slightly enhanced. These rules operate
so as to enhance coverage for people engaged in unpaid work either by specific
reference to that activity or by breaking the link between social insurance and
participation in paid work. The first type targets unpaid caring work either by
transforming the work into a form of (non-market remunerated) paid work (as
the United Kingdom’s invalid care allowance and Finland’s home childcare
allowance effectively do), or by ensuring that periods of unpaid work do not
jeopardize entitlement to social insurance benefits by crediting the individuals’
accounts with contributions for the relevant periods of care activity. The latter,
termed caring credits, are effective in a number of European states in enhanc-
49
The assumption being that employers are less likely to create jobs of this kind if they
have to pay contributions in respect of them: Nolte v Landesversicherungsanstalt Hannover
Case C-317/93) [1995] ECR I-4625; Megner v Innungskrankenkasse Rheinhessen-Pfalz (Case
C-444/93) [1995] ECR I-4740.
(
5
0
The discriminatory status of the United Kingdom’s low earnings threshold was chal-
lenged indirectly and unsuccessfully in legal proceedings before the Employment Appeal Tri-
bunal in Banks v Tesco Stores Ltd and Secretary of State for Social Security, Judgment 15 Sept.
1
999.

Equal treatment, social protection and income security for women
169
51
ing pension entitlement for carers. Contribution credits were introduced into
the (AVS) pension scheme in Switzerland in 1995 along with pension split-
ting (as it was then called) for married and divorced couples (Bonoli, 1997,
pp. 119-122). The United Kingdom and Ireland have implemented a variant of
the caring credit in their home responsibilities protection, a measure whereby
years of low or no earnings may be disregarded in the computation of pension
amount. In 1996, Ireland increased the number of years for which such protec-
tion could be available by increasing the age of qualifying children from 6 to
1
2 (Commission of the European Communities, 1997, p. 50).
It should also be noted that a form of crediting for carers operates in
relation to occupational pension schemes where employers continue to make
contributions on behalf of employees on maternity (or parental) leave. The
extent of this practice by employers is unknown. However, it is a requirement
of EU law that contractual rights (including acquisition of pension rights) be
maintained during the 14 weeks (normally paid) maternity leave mandatory in
5
2
the EU. A recent decision of the European Court of Justice paves the way for
an expansion of caring credits in occupational systems. The Court ruled that it
was unlawful for employers not to continue contributing to their pension fund
on behalf of employees on unpaid maternity leave, where the pension funds are
53
financed by employers’ contributions alone.
Examples of the second type of rule, which breaks the link between social
protection entitlement and participation in employment, are non-contributory,
residence-based benefits and provisions permitting voluntary contributions. In
1
997, Italy combined both the voluntary contribution principle and the target-
ing of unpaid work by establishing a new fund for homemakers (casalinghe),
which provides insurance against occupational accidents and old age and which
is entirely voluntary.
As to the discriminatory status of these compensatory provisions, it is
clear that when they are available to both men and women they are not in
conflict with the legal notion of equal treatment — at least as far as the prohib-
ition on direct discrimination is concerned. In practice, their effect may be to
advantage more women than men. Even if an argument could be made out that
this adversely affects more men than women, rendering the provisions indi-
rectly discriminatory, the Court would almost certainly be encouraged to hold
the provisions to be objectively justified. The legal position is different where
the compensatory provisionssingle out women for special treatment by exclud-
ing men. The European Court of Justice, again, would almost certainly hold
sex-specific provisions of this kind to be in conflict with the ban on direct
5
1
Credits are also granted with the benefits for caring, such as the German long-term care
insurance and the United Kingdom’s invalid care allowance. For further details of these credits
in the EU, see Luckhaus and Ward (1997), pp. 249-250.
52
Directive 92/85 on the safety and health of pregnant or breastfeeding workers.
5
3
Boyle v Equal Opportunities Commission (Case C-411/96) [1998] All E.R. (EC) 879.
The particular point was decided by the European Court of Justice under Council Directive
9
2/85/EEC concerning the safety and health of pregnant workers, rather than under EU legis-
lation making specific reference to equality.

1
70
International Labour Review
5
4
discrimination, but the point is not likely to arise in relation to statutory
schemes because such provisions are exempted from application of the equal
treatment principle by EU law itself (Art. 7(1)(b) of Directive 79/7).
The two political notions of equal treatment relevant here, again, are the
calls for an end to the disadvantages in social protection schemes arising from
peripheral paid work and unpaid work activity and for measures to encourage
women to enter the labour market. Provisions targeting unpaid work as part of
social protection schemes based on employment (as opposed to residence) —
such as caring credits — are particularly apt to satisfy these two notions of equal
treatment and hence to produce gender equality. They also sit happily with no-
tions of financial independence of one partner from the other and do not appear
to operate as a disincentive to women to take employment. A qualification must
be entered here, however, for it is not clear that the level of benefit resulting from
provisions such as caring credits and residence-based entitlement will be such as
to produce an adequate income for beneficiaries, while voluntary contributions
are extremely problematic for those who by virtue of their unpaid work have no
source of income other than through their partner.
To summarize, then, social protection rules which penalize unpaid work
activity look set to stay, although some minor improvements have been made.
These rules — in statutory but not occupational schemes — appear to be com-
patible with the legal notion of equality but are in conflict with the political
ideals of removing disadvantages in social protection schemes arising from
non-coverage of unpaid work and of using the offer of full social protection
rights to encourage women into the labour market. Compensatory measures,
especially those targeting caring, appear to be holding their ground well. Unu-
sually, the compensatory measures appear to conform comfortably with both
legal and political notions of equality, except where the measures are available
to women only.
Social protection, pension age and life expectancy
The two aspects of social protection discussed here have a demographic
link. The first concerns the setting of different pension ages for men and women.
The second concerns assumptions about individual life expectancy based on
statistical averages and the use of such factors to justify different benefit rates
for men and women.
Where there is a difference between men’s and women’s pension ages,
that of men has traditionally been set, seemingly without exception, at the
higher level. Thus, in 1995, 15 OECD countries applied the same pension age
for men and women and eight applied different pension ages. The situation in
54
See, for example, the European Court’s ruling in Van Cant v Rijksdienst voor Pensionen
Case C-154/92) [1993] I-3811 (footnote 56 below). The construction of equality by Constitu-
(
tional Courts in this context is also of interest. The Austrian Constitutional Court has declared
illegal legislation restricting to women the entitlement to an allowance for part-time working up
to the time a child reaches 18 months (Equality Quarterly News, 1998, p. 28). But see the
German Federal Constitutional Court’s decision (footnote 59 below) which viewed provisions
exclusively for women as justified by past unpaid work activity and hence non-discriminatory.

Equal treatment, social protection and income security for women
171
1
961 was roughly the same. In 1995, all but two of the 15 countries (France
and New Zealand) operated equal pension ages of 65 or above. Of the EU states
operating differential pension ages, six have implemented or propose imple-
menting reform involving raising women’s pension age to that of men. These
six states are: Austria, Belgium, Germany, Greece, Portugal and the United
Kingdom (Commission of the European Communities, 1998b, p. 24). In the
case of the United Kingdom andAustria, the equalization process is to be phased
in over a long period. Outside the EU, Australia, Hungary and Japan are said to
be equalizing by raising women’s pension age to that of men (OECD, 1998b,
p. 53). Switzerland raised women’s pension age from 62 to 64 in 1995 (men’s
pension age remained 65). Raising the pension age for women therefore seems
to be becoming the norm in statutory schemes, a pattern which may be repro-
duced in occupational schemes, at least in the United Kingdom (Luckhaus and
Moffat, 1996, pp. 72-78). It should be noted, however, that the reforms in
individual countries can be complex in detail, as governments and employers
attempt to combine equalization with the introduction of various forms of “flex-
ible” retirement provision and with transitional measures intended to soften the
financial blow for those who would otherwise suffer a concomitant reduction
in entitlement. This is part of a wider scenario in which European governments
are moving towards raising pension age and extending working life as ways of
coping with rising pension costs and increased longevity.
As regards the compatibility of different pension ages with the legal no-
tion of equality, such a difference is precluded, in principle at least, by the ban
on direct forms of discrimination. For this reason, the different pension ages in
occupational pension schemes were held to be unlawful by the European Court
55
of Justice as from 17 May1990. However, differential pension ages in statu-
tory schemes continue to be lawful by virtue of the provision exempting them
under EU law (Art.7(1)(a) of Directive 79/7). In relation to these statutory
schemes, the Court has held further that once Member States take action to
equalize pension ages (even though unequal ages are exempted) they cannot
retain or introduce measures relevant to that scheme which treat men and women
differently. In particular, they cannot retain a sex-specific provision, which
involves calculating the amount of pension on a more generous basis for women
than for men, as the Belgian Government attempted to do when it equalized
56
pension ages at 60 in 1990. In this respect, the Court was unimpressed with
5
5
The date of the European Court of Justice’s judgment in Barber v Guardian Royal
Exchange Assurance Group (Case C-262/88) [1990] ECR I-1889, when it was finally and
conclusively settled that occupational pensions constituted pay under Art. 119 of the EC Treaty
(
renumbered Art. 141 by the Treaty of Amsterdam), and that different pension ages breached
the equality principle in that article.
56
Van Cant v Rijksdienst voor Pensionen (Case C-154/92) [1993] I-3811. The Belgian
Government has since reinstated the previous pension age of 65 for men and provided for the
increase in women’s pension age from 60 to 65 to be phased in over a 13-year period. The
flexible retirement provisions have been retained and women’s pensions are still calculated on a
more generous basis than men’s. The Court has held in De Vriendt and Others v Rijksdienst voor
Pensioenen and Others (C-377/96 to C-384/96) [1998] ECR I-2105 that the retention of differ-
ent methods of calculating the amount of pension is permissible under Directive 79/7 while
different pension ages for men and women are maintained.

1
72
International Labour Review
the argument that the more generous method of calculating pension amounts
for women was a justifiable means of compensating the women for the disad-
vantages they encountered in the social insurance system because of their un-
paid work activity.
In addition, the Court has held that it is permissible for employers to
implement equality in occupational schemes by raising women’s pension age
57
to men’s, even though this may worsen the women’s financial position, and
that it is impermissible for employers and pension funds to introduce transi-
tional measures to protect the women who suffer loss of rights because of the
decision to implement equality by raising their pension age rather than reduc-
5
8
ing men’s. This nicely illustrates, once again, the point that the legal version
of equality is not geared in any way to the social protection function of provid-
ing income security and avoiding poverty. So far as the political notions of
equal treatment are concerned, the most significant one in the pension age
context is the call for an end to the disadvantages for women arising from
social protection’s non-coverage of unpaid work. On one view, women’s lower
pension age is fully compatible with this notion of equal treatment, because it
can be regarded as a way of compensating women for their past years of unpaid
care work and of offsetting some of the disadvantages arising within the social
protection system as a result of that activity. Thus the original lower pension
59
age for women was justified in the German Constitutional Court as a means
of compensating women for their “double burden” and for the difficulties they
encountered in fulfilling the insurance contribution period needed to obtain a
pension. By the same token, equalizing pension ages by raising that applicable
to women deprives them of this “reward”.
It can be argued more pertinently that raising the pension age for women
can bring both advantages and disadvantages in financial terms. For women in
full-time, higher-paid, highly skilled jobs, the prospect of being able to work
longer is likely both to be welcome as an end in itself and to facilitate acquisi-
tion of pension rights and, ultimately, higher pension benefits. For women in
more precarious forms of employment, however, a higher pension age is likely
to mean an exacerbation of the income insecurity they already encounter in
employment — and indeed, in the social protection system, through its incor-
poration of the insecurities attached to precarious forms of employment.
The other demographic aspect concerns the different benefit rules applied
on the basis of an assessment of the differential risk to which men and women
are said to be exposed.The attribute currently receiving the greatest attention is
women’s propensity to live longer than men. The result of this differential risk
5
5
7
8
Smith and Others v Avdel Systems Limited (Case C-408/92) [1994] ECR I-4435.
Van den Akker and Others v Stichting Shell Pensioenfonds (Case C-28/93) [1994] ECR
I-4527.
5
9
Ruling of the Federal Constitutional Court of 28 Jan. 1987 concerning Article 3(2) GG
and s.25(3) Employees’ Insurance Act and s.1248(3) Reich Insurance Act.

Equal treatment, social protection and income security for women
173
assessment is that the same pension cover may cost more for women than for
men, or women may receive a lower benefit for the same level of contribu-
tions. For technical reasons, some intermediary pension benefits may be larger
for women than for men. At the end of the day, however, women’s level of
pension tends to be lower than men’s once actuarial factors linked to differ-
60
ences in longevity are taken into account. The alternative is for pension schemes
to ignore these statistical averages and to calculate pension amounts on a “uni-
sex” basis. Men and women in the same situation would then receive the same
pension.
A full picture of the extent to which sex-based actuarial factors are used in
various types of schemes is not yet available. However, some information is
available on pensions and it is clear that occupational rather than statutory
schemes are the ones affected. A survey of occupational schemes carried out in
the United Kingdom in 1991 showed that sex-specific factors were used in
various ways by many of the schemes, although a considerable number of them
6
1
adopted a unisex approach. In Denmark, in 1996 just under one half of
.3 million employees were in defined contribution schemes using actuarial
factors based on sex. The remainder of the employees were in unisex schemes
Commission of the European Communities, 1997, p. 53). The unisex approach
1
(
is believed to be widely adopted in insurance and pension schemes in the United
States; see, for example, Equal Opportunities Commission(1998), p. 29. When
proposals for the new Hungarian mandatory social security system and associ-
ated private pension scheme were being discussed in that country, the possibil-
ity of incorporating sex-based actuarial factors into the private pension scheme
was discussed and rejected. The scheme enacted in 1997 thusrequires that when
calculating benefits in the private scheme, the mortality rate of both men and
62
women shall be determined according to a unified mortality table.
The issue of sex-based actuarial factors gives rise to tricky conceptual and
practical problems in the light of the practices adopted by many insurance
companies and pension funds and the political weight these interests can bring
to bear on EU institutions. Asked to rule on the issue, the European Court of
Justice concluded that actuarial factors based on different assumptions about
men’s and women’s longevity do not generally fall within the scope of EU
60
In a defined contribution scheme, contributions produce a lump sum on retirement
which is then used to buy an annuity, the arrangement underpinning subsequent periodic
pension payments made until death. The same lump sum will buy a higher pension for a man
than for a woman because in her case the lump sum has to be spread over a longer period. The
intermediary benefit which has to be larger is the transfer value. This is the amount paid by one
scheme to another when a person changes employment and pension schemes. It is based on the
expected cost of buying a deferred annuity equal to the pension to be paid. So the transfer value
for a woman, taking the same benefits, will be larger at the same age than for a man. See Ward
(
1994), pp. 108-109.
61
Confederation of British Industry and William M. Mercer Fraser Ltd., 1991, p. 14. See
also Luckhaus and Moffat (1996), pp. 80-81.
62
Act LXXXII of 1997 on Private Pension and on Private Pension Funds, Article 31(2).
Unofficial translation of the legislation.

1
74
International Labour Review
6
3
equality law. The Advocate General, advising the European Court of Justice
prior to its decision, came to a contrary view. Discrimination exists for him
when men and women are treated not as individuals but as a group, and unequal
treatment for individual men and women arises as a result. Thus the treatment
of individual women on the basis of assumptions about their life expectancy
derived from statistics concerning women as a group was a clear case of direct
discrimination. TheAdvocate General’s reasoning here nicely evokes the integ-
rity of the individual and the undesirability of stereotypes which underpin the
notion of equality as usually (but not ultimately in this case) applied by the
Court.
In so far as sex-based actuarial factors do not, at present, contravene the
legal notion of equality as it relates to social protection, the same would seem
to apply as regards compatibility of these factors with the political notions of
equality. Clearly, however, if women have to pay more to acquire the same
pension benefits as men, this is going to compound the disadvantages they
already suffer within the social protection system on account of their peripheral
labour market activity and engagement in unpaid work.
Conclusion
Women’s dependency on men during periods of unpaid work activity
and their peripheral connection with the labour market give rise to income
insecurities which then tend to be incorporated into and reproduced by social
protection systems. The extension of this dependency, the increase in social
protection’s bias towards already privileged forms of employment signalled by
lengthening minimum contribution periods (an increase mitigated only slightly
by a relaxation in some earnings and hours thresholds and some extension of
caring credits), and the failure to resolve the pension age and longevity issues
in a way likely to increase rather than diminish women’s access to social pro-
tection in old age reinforce the tendency of social protection systems to repro-
duce wider income inequalities affecting women.
What then of equal treatment? Can its implementation eliminate or allevi-
ate the generally insecure nature of social protection as an income source for
women? Certainly, one can conclude that the legal notion of equality is exceed-
ingly ill equipped in this respect. It has had some financially beneficial effects:
enabling women to obtain the same rate of unemployment benefit as men is one
example; enabling them to gain access to benefits from which they were previ-
ously excluded on sex and marital grounds is another. But these effects are
63
More precisely Art. 119 of the EC Treaty (renumbered Art. 141 by the Treaty of
Amsterdam). The Court held, in relation to defined benefit schemes, that these matters were part
of the funding arrangements not the employer’s promise to pay a pension, the latter being a
matter of pay under Art. 119, the former not: Neath v Hugh Steeper Ltd (Case C-152/91) [1993]
ECR I-6935, Moroni v Firma Collo GmbH (Case C-110/91) [1993] ECR I-6591. The legality of
using sex-based factors in defined benefit schemes only is reaffirmed by Art. 6 of Council
Directive 86/378/EEC as amended by Council Directive 96/97/EEC.

Equal treatment, social protection and income security for women
175
fortuitous in that they depend on the vagaries of interpretation and on the will-
ingness of Member States to implement equality by extending provision rather
64
than by reducing or eliminating it. To be an effective mechanism to combat
the income insecurity encountered by women in social protection systems, the
legal notion of equality must address the problem of poverty and be concerned
with unpaid and peripheral workers as well as those employed in full-time,
higher-paid, continuous forms of work. Instead, as it has developed at EU and
Member State levels, equal treatment has been fairly effective in its limited
aims of preventing the stereotyping of individuals (especially married women),
and of helping ensure that social protection law bears some relation to social
and economic practices, in particular the increasing participation of married
women in the labour market. Its failure to address wider substantive issues,
such as poverty and unpaid work, however, ensures that it is also fairly effec-
tive at maintaining the status quo.
Four political notions of equal treatment were addressed in the context of
income security for women. Two concern financial incentives and disincen-
tives to women working. The equal treatment ideal of using social protection as
a financial incentive to women to work is conducive to improving women’s
position within social protection since the logic of the incentives argument
points to an extension and expansion of social protection for women. The re-
emergence in Spain of this notion of equal treatment dating from the 1960s-70s
—
adapted to the modern world of part-time work — is interesting. The more
recent interpretation of equal treatment as supporting the removal of financial
disincentives to women working appears principally concerned to delimit
women’s reliance on social protection systems, particularly the means-tested
ones, and to urge them into the labour market irrespective of the type of work
available to them there. This is the risk — one also carrying with it the risk of
perpetuating the pattern of income insecurity presently the lot of many women
inside and outside the social protection system.
The financial disincentives variant of equal treatment may well be the
dominant political form it assumes in the future. In the mean time, the two
notions struggling for hegemony outside the legal sphere are that of remedying
women’s disadvantage in social protection arising from their unpaid work and
that of financial independence of one partner from the other. The first, espe-
cially, has women’s financial insecurity as a central concern. The second has
the removal of dependency on a partner as its central concern partly because
this often proves a fragile source of income but also because of the risk, inher-
ent in all forms of dependency, of one partner in the relationship being sub-
65
jected to the arbitrary will of the other. If social protection provisions can be
64
Experience in the United Kingdom (in relation to invalid care allowance and the
European Court’s judgment in Drake v Chief Adjudication Officer ((Case 150/85) [1986] ECR
995) suggests that popular mobilization in support of an extension of benefit is essential in
order to prevent a successful equal treatment claim leading to a reduction in provision.
1
65
For a discussion of the controversial propositions made here, see Luckhaus (1994),
pp. 147-161.

1
76
International Labour Review
taken to be a measure of the strength of popular and official support for these
principles, the continuing spread of dependency-related provisions like sur-
vivor’s benefit and the comparatively healthy state of measures such as caring
credits, designed to end the disadvantages encountered by women because of
their unpaid work, suggest that there is slightly more support for ending those
disadvantages than for ensuring financial independence for women. However,
a realistic rather than an optimistic appraisal of the situation suggests that
neither principle has sufficient official support and popular acceptance to over-
turn the status quo.
The stalemate in the practical sphere is replicated in debates in the theor-
etical sphere. Proponents of one principle appear to have concluded that
advancement of their principle necessarily involves erosion of the other. For
example, it is argued that financial independence for women requires the indi-
vidualization of entitlement. This in turn means abolishing dependency-related
provisions, such as survivor’s benefit and pension sharing, which currently
operate so as to offset some of the financial disadvantages women suffer be-
cause of the poor or non-existent coverage of unpaid work in social protection
systems. Within this framework, women’s financial independence — their
autonomy from men — is pitted against their continuing income insecurity —
their poverty. The result is theoretical stalemate. There are, however, a number
of ways out of this impasse. One is to challenge the assumption that depend-
ency-related provisions are effective in offsetting the poverty which women
would otherwise suffer because of social protection’s poor coverage of unpaid
work. The references in this article to the ways in which and reasons for which
a sexual relationship is a fragile source of income for unpaid workers are in-
tended to do precisely this.
Another way is to acknowledge, first, that dependency-related provisions
are not the only way of offsetting the disadvantages encountered by women in
the social protection system arising from their unpaid work and, second, that
they may not be the most effective and efficient way of doing so. Arguably,
caring credits, caring allowances and other measures targeted on the unpaid
work activity of the non-employed and of those engaged in peripheral forms of
employment provide a more effective and efficient way of tackling these disad-
vantages. As demonstrated, though still in embryonic form, measures of this
kind are already a feature of some social protection systems. They may easily
be combined with individualization of entitlement. Moreover, so long as they
are available to both men and women, they can lay claim to being compatible
with all the notions of equal treatment discussed in this paper. There is there-
fore a way out of the theoretical stalemate between women’s financial inde-
pendence and their poverty — if the political will to jettison dependency is
there.
Individualization coupled with caring credits and other measures targeted
on unpaid work within social protection systems provide a starting point for
advancing the position of women currently caught in the web of income in-
securities caused, ultimately, by their unpaid work. But it is limited and does
not profess to provide a solution to the income insecurities experienced by

Equal treatment, social protection and income security for women
177
groups such as the non-employed, part-time workers and the low-paid, which
are not linked to unpaid work activity. Solutions to these problems are to be
found elsewhere and little is to be gained from rejecting individualization,
caring credits and other such measures on the grounds that they create inequi-
ties between different groups of paid and unpaid workers, all of whom risk
sharing the same impoverished state.
Even for unpaid workers, individualization and measures such as caring
credits (all of which can be implemented in an incremental, low-cost, way)
need to be part of a wider strategy developed in response to the question of how
best to provide for caring needs. A strategy focused on producing flexible work-
ing hours (without diminution of employment conditions and social protec-
tion), on developing and expanding the range of different kinds of caring serv-
ices of a formal and informal kind, and on providing a similarly wide-ranging
and innovative system of financial payments for men and women carers both
inside and outside the social protection system may go some way towards pro-
66
viding a satisfactory answer. Only then will it be possible to say that sex and
gender equality in the social protection field has been achieved.
References
Addis, E. 1999. “Gender in the reform of the Italian welfare state”, in South European Society
and Politics (London), Vol. 4, No. 2 (Autumn), pp. 122-149.
Bonoli, G. 1997. “Switzerland: Institutions, reforms and the politics of consensual retrenchment”,
in Clasen, pp. 107-129.
CBI (Confederation of British Industry) and William M. Mercer Fraser Ltd. 1991. Pensions post-
Barber: Equalising occupational pension schemes. London , Oct.
Clasen, J. (ed.). 1997. Social insurance in Europe. Bristol, The Polity Press.
Commission of the European Communities. 1998a. Reconciliation between work and family life
in Europe. CE-V/2-98-008-EN-C. Brussels.
—
—
. 1998b. Social protection in Europe 1997. DGV/E.2 (Apr. 1998). Luxembourg, Apr.
. 1997. Monitoring implementation and application of Community equality law 1996. CE-V/
2-98-007-EN-C. Brussels, July.
Cooper-Green, Emma-Jane. 2000. “Labour Force Survey household data: Spring 1999 analy-
ses”, in Labour Market Trends (London), Vol. 108, No. 1 (Jan.), pp. 25-34.
Duncan, Alan; Giles, C.; Webb, S. 1994. Social security reform and women’s independent
incomes. Equal Opportunities Research Discussion Series No. 6. Manchester, Equal
Opportunities Commission.
Eardley, Tony; Bradshaw, Jonathan; Ditch, John; Gough, Ian; Whiteford, Peter. 1996. Social
assistance in OECD countries. Vol. 1: Synthesis report. Research Report No. 46. Department
of Social Security. London, HMSO.
Equal Opportunities Commission. 1998. Equality in the 21st century: A new sex equality law.
Manchester.
Equality Quarterly News (Brussels). 1998. “Constitutional Court has withdrawn part-time
allowances for self-employed mothers”, in Spring issue, p. 28. Since the beginning of
1
999, available in electronic form only, on http://europe.eu.int /comm/dg05/equ_opp/
index_en.htm [visited 12 Apr. 2000].
6
6
The European Commission, away from the distraction of mainstreaming, is developing
an interesting multidimensional approach of this kind; see Commission of the European Com-
munities (1998a).

1
78
International Labour Review
European Industrial Relations Review (London). 1999. “Part-time work agreement signed”, in
Issue 300, Jan., pp. 30-31.
Germany, Government of. 1999. “Gesetzentwurf zur Neuregelung der geringfügigen Beschäf-
tigung eingebracht”. Press release on website http://www.bma.de/de/asp/aktuell
presse.asp?id+968 [visited 7 Apr. 2000]. Ministry of Labour and Social Affairs. Bonn,
22 Jan.
Hoskyns, C.; Luckhaus, Linda. 1989. “The European Community Directive on Equal Treatment
in Social Security”, in Policy and Politics (London), Vol. 17, No. 4, pp. 321-335.
ILO/EASMAT. 1997. Ageing in Asia: The growing need for social protection. Bangkok.
Kuhnle, S. 1999. The Scandinavian model: Trends and perspectives. European Forum Seminar
Paper WS/23. Florence, European University Institute.
Luckhaus, Linda. 1994. “Individualisation of social security benefits”, in McCrudden (1994),
pp. 147-161.
—
—
—
. 1986. “Payment for caring: A European solution?”, in Public Law (London), Winter,
pp. 526-537.
; Moffat, G. 1996. Serving the market and people’s needs? The impact of European Union
law on pensions in the UK. York, York Publishing.
; Ward, Sue. 1997. “Equal pension rights for men and women: A realistic perspective”, in
Journal of European Social Policy (Bath, UK), Vol. 7, No. 3 (Aug.), pp. 237-253.
McCrudden, C. (ed.). 1994. Equality of treatment between men and women in social security.
London, Butterworths.
McKnight, A.; Elias, P.; Wilson, R. 1998. Low pay and the National Insurance System: A
statistical picture. Equal Opportunities Research Discussion Series. Manchester, Equal
Opportunities Commission.
OECD. 1998a. Employment Outlook. Paris, June.
—
—
. 1998b. Maintaining prosperity in an ageing society. Paris.
. 1996. Employment Outlook. Paris, July.
Olsen, Darcy Ann. 1999. “Social security and women”, testimony of Darcy Ann Olsen, Entitlements
Policy Analyst, Cato Institute, before United States House of Representatives Committee
on the Budget. 4 May 1999. http://socialsecurity.org /dailys/05- 04-99.html [visited
31 Mar. 2000].
O’Neil, Maureen; Ciffin, Shirley. 1978. “Social security and divorce in Canada”, in Social
security provisions in case of divorce. Studies and Research No. 11. Geneva, International
Social Security Association, pp. 64-77.
Pierson, P. 1999. The comparative political economy of pension reform. European Forum Seminar
Paper. Florence, European University Institute, 24 Feb.
Shirley, Ekaterina; Spiegler, Peter. 1998. The benefits of social security privatization for
women. SSP No. 12. The Cato Project on Social Security Privatization. 20 July.
http://socialsecurity.org/pubs/ssps/ssp12es.html [visited 21 Dec. 1999].
Solcher, Hans. 1978. “Sharing of pension rights in the event of divorce under German law”, in
International Social Security Review (Geneva), Vol. 31, No. 3, pp. 308-317.
South African Law Commission. 1999. Sharing of pension benefits: Report. Project 112.
http://www.law.wits.ac.za/salc/salc.html [visited 7 Apr. 2000]. Pretoria, June.
—
. 1998. Sharing of pension benefits. Discussion Paper 77. Project 112. http://www.law.wits.ac.za/
salc/salc.html [visited 7 Apr. 2000]. Pretoria.
United Kingdom, Government of the. 1999. “Government announces consultation on draft
pension sharing regulations”. Press Release 99/316. Department of Social Security. Lon-
don, 15 Nov.
—
. 1998a. A new contract for welfare support in bereavement. Department of Social Security.
Command No. 4104. London, HMSO.
—. 1998b. A new contract for welfare. Green Paper. Command No. 3805. London, HMSO, Mar.
Ward, Sue. 1994. “Equality of treatment between women and men in occupational social security”,
in McCrudden (1994), pp. 99-113.