Values and Policy Attitudes: The Impact of Family Values and

Family Policy on Maternal Employment and Its Change from a

Comparative Quantitative Study*

*This paper is the preliminary draft for a presentation at the stream four on

‘Comparative social policy’, the 2010 annual SPA conference ‘Social Policy in Times

of Change’, 5-7 July 2010 at Lincoln University, Lincoln, the UK.

Nam K. Jo** **University of York

email: nj506@york.ac.uk

Abstract In spite of an increasing interest in cultural explanations of welfare, empirical evidence for

the effect of culture is still rare, especially from comparative quantitative perspectives. This

paper sees that one of reasons for this is, whilst there has been theoretical development

recently, the conceptualisation of the effect of culture was not refined yet to facilitate more

fruitful empirical examinations. The effect of culture can be seen as twofold, at the before

and after stage of decision-making, drawing on a recent suggestion that culture should be

seen at the same level with economic and political context for policy-making, which must be

embedded in society and stable over time. Here, it is attempted to develop this further with

focus on the ex-post cultural impact.

Conceptually, it is suggested that people’s values, which collectively form the cultural

context of society, should have influence on concrete policies’ fates through affecting policy

attitudes toward each policy. In particular, it is reasoned that a certain degree of cleavage

between people’s values and a concrete policy generates their negative attitudes toward the

policy, which should lead to ‘passive resistance’ to the policy such as not-taking-up of

entitled or eligible welfare support. Empirically, from all three waves of the EVS and WVS

data, the impact of traditional family values of mothers with young children on their working

decisions (status) was examined with controlling how strongly they are supported by family

policy across multiple OECD countries.

While it seems that family policy is clearly working – helping mothers to reconcile

working and parenting - it is shown that in the early 1980s the level of traditional family

values of mothers matter in their employment status where they are strongly supported by

family policy – that is, where their decisions to work much imply their take-up of support

offered by well-developed family policy. However, this impact appears to have faded away

over time, possibly reflecting the change of the nature of maternal employment. More

broadly and practically, this ex-post effect implies that a culturally better informed welfare

policy can be more effective.

1

Introduction The cultural analysis of welfare has recently observed significant development in theory. In

particular, there have been discussions about which perspective on culture can facilitate better

understanding of the relation between culture and social policy, with recognition that the

cultural accounts tend not to be meaningful and analytical when culture is seen to either

determine everything or be merely a reflection of social system (van Oorschot et al., 2008). In

this context, for example, the welfare culture approach (Pfau-Effinger, 2005) has offered a

systematic and profitable conceptual framework for the cultural explanation of welfare, which

sees that culture and the social system are interrelated but not decisively determined each

other. And it has been suggested that culture is another context for policy making, which is

different from the economic and political contexts (van Oorschot, 2006) - this is especially so

for ‘welfare’ policy-making which reflects the value-dimension more than ‘democratic’ and

‘capitalistic’ policy-making does (Marshall, 1972: 20).

This theoretical achievement, however, appears not to have led to an immediate increase

in cultural studies of welfare, especially from the comparative quantitative approach. I argue

that this is partly because our conceptualisation of how culture affects social policy was not

refined yet drawing on those discussions above. If it is briefly put without details, culture has

mostly been seen as either a general force for social structures or popular opinion/attitudes

within studies of the cultural dimension in relation to welfare. By the former type research,

the cultural foundations – ideological or religious traditions such as liberalism, conservatism,

socialism and Christianity – of welfare have been greatly illuminated (e.g. Evans, 1978;

Lockhart, 2001; Manow & van Kersbergen, 2009; O'Connor & Robinson, 2008; Opielka, 2008;

Stjernø, 2008; van Kersbergen, 1995; van Kersbergen & Kremer, 2008), but such a historical

process of confining policies to ways which are compatible with and enhance those cultural

traditions could hardly been examined analytically. By the latter type research, the variation of

welfare attitudes across countries has been known (e.g. Andreβ & Heien, 2001; Blekesaune &

Quadagno, 2003; Gundelach, 1994; Svallfors, 1997, 2007; van Oorschot, 2000, 2006, 2008;

van Oorschot & Halman, 2000), but either the ex-ante causal impact – i.e. setting and limiting

policy agendas and policy options (Alcock, 2001: 16; Freeman & Rustin, 1999: 18; van

Oorschot, 2006, p. 24) - or the ex-post legitimacy control effect – i.e. offering or withholding

support for policies (Cnaan et al., 1993: 124; van Oorschot, 2006: 24) - of welfare attitudes was

not so clear since welfare attitudes are vulnerable to issues and situations, and those can be

easily ‘moulded’ by policy-makers and the media (Page et al., 1987; van Oorschot, 2007: 134-

135; Whiteley, 1981: 461).

Instead, in a similar way we saw both the idealistic and materialistic approach cannot be

helpful for the cultural analysis of welfare in the above, it is assumed that conceptualising

culture at an in-between level which is neither as abstract as the cultural foundations nor as

concrete as welfare attitudes would be more fruitful. Culture as another context for policy

making should not be a determinant or dependant of the economic and political context, and at

the same time, it must be widely and deeply embedded in society and quite stable over time

(van Oorschot, 2006: 24). According to this reasoning popular and stable values can be more

desirable cultural dimensions as the cultural context for policy making, and with them how culture

affects social policy can be synthetically re-conceptualised: the cultural context (value-

characteristics of society) directs policy decisions to ways in accordance with and reinforcing it by

affecting public opinion on welfare issues which would set and limit social policy agendas and

options; and by influencing public attitudes toward welfare policies which would partly control

the legitimacy bases of those policies.

This paper is an attempt to contribute to the study of the relation between culture and

welfare by advancing this conceptualisation of a twofold effect of culture on social policy both

conceptually and empirically. In particular, I will focus on the effect of culture on social policy

2

at the ‘after’ stage of decision-making, which appears to be relatively less developed. In the first

part of this article, it is discussed further how such an effect of values at the policy

implementation stage can be conceptualised, especially in relation to take-up of welfare policy

support. The second part is an empirical analysis of this ex-post effect of culture on welfare

with focus on family policy and traditional family values across eight the most affluent

welfare states. While findings generally serve as evidence for the effect of culture, more

broadly and importantly, it is implied that the cultural analysis of welfare can inform the

change of welfare and contribute to more effective policy-making.

Values, policy attitudes and the ex-post effect of culture

The ex-post effect of culture on social policy is revealed through people’s evaluation of

concrete policies. When a policy instrument is introduced and implemented, people may

evaluate it individually or in groups - whether it is good or bad, desirable or undesirable,

would-be effective or not, and so on, either for themselves, or for targeted beneficiaries or for

the entire society. Having such judgments about a policy instruments is seen as policy

attitudes (Aalberg, 2003: 5-8). Meanwhile, there is almost always cleavage to a certain extent

between people’s values and a specific policy instrument. Even if people’s values are

reflected in policy-making as a causal factor, which was discussed as the ex-ante effect above,

the influence of values would be limited. Decision-making is not exclusively driven by the

cultural context but more by, for example, interests and power relations. Thus a gap has

inevitably arisen between concrete policy and values of people, especially those of policy

recipients, and this gap will influence their policy attitudes. People may find that a particular

policy is either more satisfactory or less in terms of their values and have either more positive

policy attitudes toward the policy or less. This does not mean that people’s policy attitudes are

dependent only upon their values. Similar to the process of cultural impact on the ‘before’

stage of policy-making, there would be other factors, such as interests and socio-demographic

conditions, all of which are not necessarily mutually exclusive. That is, people’s values, their

material interests and/or their particular situations, all of which would lead to, for example,

their negative evaluation of a specific policy.

It is reasonable, therefore, to conjecture that the evaluation of a concrete policy by people

(policy attitudes) affects the fate of that policy. When people favour and support a policy,

government will be encouraged to continue, enhance and expand it. If the policy is unpopular

there can be calls for it to be amended, repealed (abrogated) or replaced by another. In short,

whether people see a policy as being legitimate, which would partly determine the policy’s

future, is dependent upon the policy attitudes they have. The ex-post effect of culture on

policy is found here, where policy attitudes, which are partly shaped by values, have impact

on implemented policies’ survival.

Passive resistance and take-up decisions However, the policy attitudes of people may not be easily visualised publicly, especially when

they are positive toward policies – in this case their policy attitudes would remain unspoken

unless individuals are asked to reveal those (for example, by polls). Instead, we may be able

to expect that people would show their policy attitudes more clearly and actively (e.g. through

petitions or demonstrations), when these are negative towards a specific policy. Possibly, they

will manifest in more aggressive ways when their (material) interests are threatened by a

policy. If we can call such a voluntary and aggressive expression of negative policy attitudes

‘active resistance’, there must also be ‘passive resistance’ to a policy, that is, less aggressive

3

expression of their negative policy attitudes - when people find they just cannot approve a

policy as much.

It is assumed that the cleavage between values and policies is highly likely to lead not to

the former case - active resistance - but to the latter case - passive resistance. This does not

mean that the negative policy attitudes to certain policies can be differentiated by reasons (i.e.

either interests or values). Both would affect any policy attitudes simultaneously. However, if

we want to focus more on the cases which illuminate the effect of values better, examples of

passive resistance would be more desirable than those of active resistance which are probably

driven more strongly by interests than values.

I suggest that hesitation in taking-up a certain entitled service/benefit provided by a

policy instrument is a possible way of passive resistance to the policy. In fact, many ‘take-up’

studies (e.g. Hancock et al., 2004; Riphahn, 2001) imply the influence of the cultural

dimensions on decisions about ‘non-take-up’. Although they mostly attempt to explain non-

take-up with a rather economic concept of ‘hidden costs’ which outweigh the potential

benefits, it is also recognised that beneficiaries’ perceptions such as stigma, self-respect and

the unpleasantness of the claim process partly determine the amount of the hidden costs. Of

more relevance to social policy, van Oorschot (1991: 23-24) suggests that the recipients’

attitudes toward welfare policy in general are ‘another missing key factor’ influencing non-

take-up, possibly forming or raising an ‘instinctive barrier’ (Graham, 1984, and Ritchie &

Davies, 1988 in ibid).

This reasoning is supported empirically. For example, it is reported that the take-up of

parental leave scheme depends more on ‘motherhood culture’ than types of schemes – the

take-up rate in the ‘more strongly family-oriented culture’ of Germany is much higher than in

Finland in spite of the Finns’ more generous scheme (Pfau-Effinger, 2004, 2005: 12). The

public-funded professional caregiver policy was not successful (low take-up rate) in Germany

for the same reason – here they prefer relatives to professional care providers (Pfau-Effinger,

2005: 13-14). Taylor-Gooby (2008: 179-180) contends that service users’ decision-making is

often more dependent upon their values than their income, needs and the amount of

information they have, drawing on qualitative research across various welfare areas, which

have shown that unexpected outcomes are caused by inconsistencies between policy-makers’

expectations and service users’ values.

Traditional family values and ‘take-up’ of support by family policy Thus it is suggested that we are able to empirically examine the impact of values on social

policy at the implementation stage by looking at the relation between popular and stable

values and the take-up rate of a specific policy. In practice, however, ‘take-up’ data rarely

exist. Rather a thorough review of both the relevant literature and administrative sources

(Hernanz et al., 2004) reports that information on take-up rates is very limited even within

OECD countries – since governments do not produce it (except the UK) and it is almost

impossible to measure the precise number of ‘potential beneficiaries’. We can easily presume

that the ‘entitled-but-not’ claimants will not report their eligibility if there is a stigma attached

to the claim. For this, studies of take-up rates have commonly ‘estimated’ national rates from

data on individual socio-demographic and socio-economic status (Hernanz et al., 2004: 13-

17), although the comparability of these national level take-up rates has been questioned since

contents and characteristics of a policy always vary by society (Bruning & Plantenga, 1999:

199).

Given this, I attempt to estimate the take-up ‘decisions’ at the individual level instead of

estimating the take-up ‘rate’ at the national level. It is assumed that mothers’ decisions to

participate in the labour market show, at least partly, their decisions to take up supports

4

provided by family policy such as leave schemes and public child care, particularly within

societies strongly supporting maternal employment through family policy. However, this

assumption includes many logical leaps – above all, since their decisions to work are not

determined solely by supports they can get from family policy, their decisions not to work do

not necessarily mean that they have made ‘non-take-up’ decisions. Whilst this is a critical

limitation, some points were considered in this selection. Firstly, we can only ‘estimate’

individual take-up decisions - like the collective take-up rate, individual take-up decisions are

barely visible and measurable. Secondly, even if the take-up decision about family policy

supports is only partially reflected in working decisions, there are few alternatives for

assessing the actual effect of family policy other than to examine mothers’ working status.

That is, the examination of the impact of values on mothers’ employment with consideration

of family policy can be meaningful even without referring to the take-up and policy attitudes.

Thirdly and more practically, it is not easy to find any other take-up related dimensions

available within the same dataset from which we can draw stable value-dimensions. While

international attitudinal survey data are desirable for the latter, those surveys rarely cover

respondents’ welfare entitlements and service uses. Finally, we may be able to control some

suspected factors for mothers’ working decisions by including them into the analysis, and this

would help us to minimise exaggeration of the impact of values.

Meanwhile women’s decisions to work, especially of those who are more forced to make

reconciliation between career life and family responsibility, have been one of subjects where

the impact of values are suspected and examined more widely. Even if not all of us would

agree with Hakim (2000) that women’s choices between market work and family work are,

and will be, determined by their heterogeneous preferences (i.e. work-centred, family-centred

and in-between adaptive orientations), nobody may be able to deny her observation that

economic affluence, structural changes in the labour market and the development of welfare

policies have increased the importance of values and preferences in lifestyle decisions

(including working decisions) which were not quite the matter of choice but mostly

determined by social structural conditions in the past – this is particularly so for women’s

employment decisions due to ‘the contraceptive revolution and the equal opportunity

revolution’ (Hakim, 2000: Ch. 1 & 3). We already saw in the above that values matter in

mothers’ decisions to take the leave scheme (Pfau-Effinger, 2004, 2005: 12). It is also

reported that mothers tend to refer to moral and socially negotiated norms about ‘what is

acceptable’ when they make decisions about reconciliation between work and parenting

(Duncan & Irwin, 2004: 397 in Taylor-Gooby, 2008: 179-180).

According to Kangas and Rostgaard (2007) the studies of women’s employment have so

far emphasised either socio-structural factors such as education and income or welfare

institutional factors (welfare regimes and more specifically, family support policy) or

attitudinal factors (women’s preferences), but research combining all of those factors has been

rare, which is necessary to better understand the constraints on female employment. What

they found in their integrated analysis of those factors is, as they expected, that all of them

matter: more unemployed women with stronger home-centred preferences, less generous

leave schemes, poorer day-care services, more children under the school age and lower socio-

economic positions. However, they could not avoid questioning themselves about the causal

direction from opinion (preferences), which has often been presented in the studies of welfare

opinion/attitudes. As it was discussed earlier in this paper, the causal impact of the cultural

dimension would be better (if not should be) sought not from opinion which is issue- and

situation-dependent and unstable but from stable values forming the cultural context of

society and affecting opinion, on which the examination of the ex-post effect of culture on

welfare in this paper is based.

5

From a wider cross-cultural comparative perspective, it is argued that there can be a

distinction in Western Europe between ‘nations of families’ like Germany and Italy, where

the family plays an essential role in various aspects of life and it is more likely that a

breadwinner-father, homemaker-mother and dependent children are supported, and ‘nations of

individuals’ like Sweden and the UK, where strong individualism and social equality exists

and supporting working mothers and protecting children’s rights tend to be more concerned

about (Chesnais, 1996 in Brewster & Rindfuss, 2000: 284). This is echoed by a finding that in

such ‘nations of families’ breadwinner-fathers are supported by stronger employment

protection legislation and unemployment benefits (Boeri et al., 2001: 12).

Here we can assume a bit ironically that family policy cannot be strong in countries

where family is more highly valued and vice versa - the term ‘family policy’ is individualistic

oriented in that it is widely referred to policies designed to support women’s reconciliation

between employment and family responsibilities, mainly through childcare and parental leave

schemes. Family policy focused on supporting working mothers is possibly less popular

where stronger family values exist. If values affect policy attitudes as we conceptualised, it

can be hypothesised that family values should have a negative impact on mothers’ attitudes

(or an impact on their negative attitudes) toward family policy. For example, mothers with

strong family values can be less likely to rely on public childcare. They would be more likely

to choose parenting instead of leaving their children to be cared for by others while they are

working than those of weaker family values. In other words, mothers with strong family

values can have higher ‘instinctive barriers’ (van Oorschot, 1991: 23-24) to the take-up of

supports offered by family policy (more negative policy attitudes toward these policies).

Data and methods By our conceptualisation of cultural impact to get value-dimensions stable over time is the

key to this analysis. What should be noted here is that this stability is about not values

individuals hold but the cultural context – i.e. a collective level phenomenon, and not

immutability but being ‘permanent enough’ to enable us to predict consequences (Oyserman

& Uskul, 2008: 149-150). I attempt to draw value-priorities of individuals (individual-level

analysis) and aggregate these for value-characteristics of society (cultural context) following

most studies about values and culture (e.g. Hofstede, 2001; Schwartz, 1999) – since values of

individuals reflect not only unique individual experience but also shared culture of a

collectivity, the average values show ‘the central thrust of their shared

enculturation’(Schwartz, 1994: 92, 1999: 25-26). Thus stability will be examined by how well

the differences in the cultural contexts (aggregated value-priorities) between societies have been

maintained over time. It is expected that cultural differences are relatively enduring (Inglehart,

1990: 15) and, drawing on Hofstede’s method (2001), the correlation between ranks of countries

by collective value-characteristics (aggregated value-priorities) at differing time points will be

investigated. For this all three waves of the EVS (European Values Study) and corresponding WVS

(World Values Survey) data (at 1981/2, 1990 and 1999/2000) were analysed. This is one of

the few cross-cultural surveys focusing on values and attitudes for a relatively longer period

across a wider range of countries, and there have been few criticisms of the sample quality of

these data, especially for the highly industrialised countries (Larsen, 2006: 27). 22 OECD

countries was selected in order to minimise the impact of the differing socio-economic

context which may mask cultural difference (24 countries who joined before 1990 but

economically weak Turkey and no-data New Zealand). These countries share not only a

comparable level of wealth but also a common conception of social welfare historically as so-

called welfare states (Arts & Gelissen, 2002). The final data set consisted of 73,534 cases.

6

Only items which have been repeatedly asked at all three time points in most object

countries were selected and separated into 10 groups based on correlations. Seven items

formed the group of family:

One must always respect and love one’s parents (0~1: agree); Parents must always do best for their children (0~1: agree); Child needs a home with father and mother (0~1: agree); A woman has to have children to be fulfilled (0~1: agree); Marriage is an out-dated institution (0~1: agree); Future changes: More emphasis on family life (1: good thing ~3: bad thing); Woman as a single parent can be approved (1: disapprove ~ 3: approve).

These were factor analysed (by Principal Components Analysis) with the pooled dataset,

which extracted one underlying dimension. To obtain the comparability across countries of

this measure, which has been pointed out as an ‘immanent danger’ in comparative attitudes

research (Larsen, 2006: 26; Svallfors, 1997: 287), it was compared with factor solutions from

data by each wave, by each country and by each country at each wave (cf. van de Vijver et al.,

2008). Three items were discarded as they looked having different meanings across time

and/or place. Then the seemingly comparable dimension was compared with the result of

factor analysis of the same four items at the aggregated level with country means of items (cf.

van de Vijver et al., 2008) - the individual-level values and the culture-level values can be

different by differing internal logics (Hofstede, 2001: 17; Schwartz, 1994: 92-93). Only after

deselecting one more item, it appeared that the underlying dimension is cross-time and cross-

place comparable and cross-level equivalent, which is based on three items: child needs a

home of both parents; a woman has to have children to be fulfilled; and one must always

respect and love one’s parents. Differences in the level of this underlying dimension between

societies have been well maintained (with a significant rank correlation over 0.7 between

wave 1 and wave 3). This dimension was interpreted as ‘Traditional Family Values’ in that it

emphasises conventional ideas of family format, the nature of women and ‘the traditional

authoritarian familism’ (Gundelach, 1994: 43-44). This comparable measure echoes

suggestions in the literature – in a figure of countries’ profile of traditional family values

(with only some countries for a clearer view), we can find ‘nations of families’ (Chesnais,

1996 in Brewster & Rindfuss, 2000: 284) in the upper part and the ‘individualism-oriented’

Anglo-Nordic countries (Gundelach, 1994) in the lower side (Figure 1).

The employment status of mothers was drawn within the same EVS-WVS dataset. In

particular, the analysis focused on mothers living with children under the school-age (the age

of 5), who are at a more critical moment of decision-making about harmonising working and

parenting (so more targeted by family policy). This information was available only for wave 1

(1981/2) and wave 3 (1999/2000) data. 8 categories of the employment status were

dichotomised: working (full-time, part-time, self-employed and students) and not-working

(retired, housewife, unemployed and others). After selection of cases of mothers with young

children, it was not practically allowed (in terms of the number of cases) to differentiate

between full-time and part-time work although it can be more inspirational (e.g. see Kangas &

Rostgaard, 2007). Initially the age bracket 20-59 years was selected, when women are more

active in the labour market, but there were few cases over the age of 50 (living with young

children) and the age interval was decreased to 20-49.

7

Figure 1 Traditional family values in 13 OECD countries over two decades

wave1 (1981/2) wave2 (1990) wave3 (1999/2000)

Tra

diti

onal f

am

ily v

alu

es (

weak ~

str

ong)

Canada

Denmark

Finland

France

Germany

Italy

Netherlands

Norway

Portugal

Spain

Sweden

United Kingdom

United States

Information on socio-structural factors and especially about various conditions around

cases of mothers is very limited in this dataset. For example, there are no items on whether a

respondent is living with a partner, whether she lives in a dual-earner household, whether she

can get help with child care from grandparents and so on. Only three available compositional

factors were included: the level of education (10-point scale)1; whether the respondent is the

chief wage earner in the household; and their income level (10-point scale). It is assumed that

more educated mothers would, if other conditions are controlled, be more likely to work

because they have had, or can have, better jobs in terms of career building and salaries (the

larger opportunity cost). A mother would be more strongly forced to combine her work and

parenting if she is the chief wage earner in her household. Income level would also reflect the

opportunity cost – mothers of higher income level may have jobs with greater salaries and

will therefore be more likely to want to keep their jobs.

To what degree mothers’ employment is supported by family policy is gauged by

Gornick et al’s family policy index (Gornick et al., 1997). They examine the variation of

policy efforts which affect maternal employment patterns across 14 industrialised nations,

based on 16 policy indicators covering various policy areas such as benefits for parents, child

care, and public school policies drawn from data collected in the 1980s. Later, they developed

their index further and formulated the family policy index which covers wider policy areas

with more recent data - mostly of 2000 with a few data from the late 1990s (Gornick &

Meyers, 2003). Since they extended policy areas reflected in their index and changed the

formula for the index, these two indices are not directly comparable. However, variation

between the most developed countries in the policy effort to support mothers in combining

working and parenting appears clear in their index scores at each time points. According to

their work (Gornick & Meyers, 2003; Gornick et al., 1997), countries can be clearly classified

into three groups of strong, moderate and weak family policy at the mid-1980s (Table 1).

Although this distinctiveness becomes somewhat blurred in 2000, we can still divide those

into three groups with some borderline cases. These country groups by the level of family

8

policy will be used for our examination of mothers’ working decisions at 1981/2 and at

1999/2000, respectively.

Findings

Working mothers, their traditional family values and family policy in the early 1980s

The results from binary logistic regression analysis (cf. see Tabachnick & Fidell, 2007: Ch.

10) shows that the degree of traditional family values mothers have is a predictor for their

working status (Table 2). The stronger the traditional family values mothers have the less

likely to work appear they to be (with negative coefficients and the odds ratio below 1). In

addition, it is suggested that mothers who are the chief earners of their households, who have

higher levels of household income and who stayed longer in the formal education system are

more likely to work. As we assumed above, more educated and larger income earning

mothers seem to face a greater opportunity cost if they do not work. It is not surprising that

mothers are more likely to work when they are chief earners of their household.

Inclusion of the level of family policy effort into our prediction made the impact of

traditional family values clearer (with better significance level and larger contribution). This

supports Kangas and Rostgaard’s suggestion (2007) that the structural, institutional and

attitudinal dimensions should be considered together to better understand women’s and

mothers’ employment decisions. Meanwhile, it is shown that family policy is actually

‘working’ - mothers with young children in countries where their employment is strongly

backed up are more likely to work than those in countries of moderate and weak family policy.

It is probable that more mothers would find it hard to reconcile their jobs with child-rearing in

Table 1 Groups of countries by the degree of policy effort for supporting employment of mothers with young children

The mid-1980s The late 1990s - 2000

Country Scorea Country Score

b

High Effort Group

France 64.9 Denmark 0.94

Denmark 63.7 Sweden 0.89

Sweden 61.9

Finland 60.8

Belgium 55.6 - - - (Norway 0.80) - - -

Italy 50.6

Moderate Effort Group

Luxembourg 35.2 Finland 0.74

Germany 34.1 Belgium 0.73

Canada 32.4 France 0.66

Netherlands 32.0 Netherlands 0.65

Norway 31.2 Luxembourg 0.65

- - - (Germany 0.55) - - -

Low Effort Group

United Kingdom 21.6 United Kingdom 0.45

Australia 19.2 Canada 0.36

United States 17.1 United States 0.24 a Source: Index of policies supporting employment for mothers with children under 6 (Gornick, Meyers, & Ross, 1997).

b Source: Index of policies that affect families with children aged 0-5 (Gornick & Meyers, 2003).

9

countries only weakly and moderately supporting their employment and gave up their careers

than in societies strongly supporting maternal employment.

The separate look by family policy effort groups (the lower part of Table 2) provides us

more interesting scenes. That traditional family values matter (only) in the strong family

policy effort group is first and foremost – this shows that traditional family values become

more relevant to decision-making about working when such decisions become more a matter

of choice. In nations strongly supporting maternal employment, mothers may be more able to

keep their jobs (by taking up supports provided by family policy). According to our data,

mothers tend to refer to their family values to an even greater extent than to the fact that they

are chief bread-winners of their households (with a larger contribution by family values).

Where mothers’ employment is not strongly supported and their decisions are thus more

heavily dependent upon conditions – for example, the lack of child care support may force

them to quit their jobs to care for their babies or to keep their jobs to pay for the cost of

Table 2 Prediction of working status of mothers with children under the age of 5 at 1981/2

B Odds ratio Contribution

a B Odds ratio Contribution

a

Education completed age .16** 1.18 14.68 .17** 1.18 14.70

Chief Earner of household 1.37** 3.92 14.35 1.47** 4.33 15.67

Household Income .23** 1.26 18.34 .25** 1.28 20.34

Traditional Family Values -.21* .81 4.61 -.33** .72 9.74

Family Policy (referred to strong family policy group)

-

11.53

Weak family policy -

-.81** .44

Moderate family policy -

-.72** .49

N 453 453

Nagelkerke R square 0.149 0.179

↙

Weak family policy Moderate family policy Strong family policy

B Odds ratio

Contri-bution

a

B Odds ratio

Contri-bution

a

B Odds ratio

Contri-bution

a

Education Level .15 1.16 2.78 .13 1.14 2.34 .18** 1.19 7.71

Chief Earner 1.46* 4.29 7.69 1.06 2.90 2.52 2.23** 9.28 10.08

Income Level .29** 1.34 12.13 .11 1.12 1.63 .38** 1.47 16.17

Traditional Family Values -.16 .85 .80 -.21 .81 1.21 -.63** .53 12.29

N 132 144 177

Nagelkerke R square 0.170 0.069 0.298

Data: EVS-WVS 1981/2. ** Significant at the 0.01 level; * significant at the 0.05 level. Within 9 countries: Belgium, Denmark, France, Italy (strong family policy effort countries), Canada, Netherlands, Norway (moderate family policy effort countries), UK and the US (weak family policy effort countries) – see table 1. a By the change in -2log-likelihood if the variable is removed from the model.

10

private child care, the impact of family values is found to be weak. Given that a mother’s

decision to work partly but inevitably means their taking-up of support provided by family

policy within societies of strong family policy, it can be argued that the individual take-up

decision on family policy support is influenced by traditional family values. In other words, if

the take-up decision is a way of realising and expressing individual policy attitudes to a

concrete policy, it is suggested that policy attitudes are partly dependent upon values, as we

hypothesised.

At 2000

Data at this time point was more limited and it was not allowed to examine each policy effort

group separately. Unfortunately, we could not find any significant impact of traditional family

values on the working status of mothers (Table 3). There can be some possible explanations

for this. Although we saw above that traditional family values is quite stable over time at the

collective level, the level of family values of mothers at 2000 is clearly lower than that of

mothers in the 1980s at the individual level.2 As a result, the instinctive barrier to decisions to

work generated by traditional family values may not be so high for mothers at 2000. More

important than this, social-structural change would have made mothers’ participation in the

labour market more irrelevant to their family roles and responsibilities which have

traditionally been emphasised. In fact, it is now rather customary for women to be part of a

dual-earner couple, to keep their jobs and develop their careers over the course of a marriage

as well as having babies, taking their children to nurseries and pre-school units and so on.

Mothers who are working are not that different from women who are working nowadays. In

fact, within our data, the percentage of ‘working’ mothers who live with children under the

age of 5 has increased from 49.0% to 62.3% (at 2000, the percentage of working women is

68.2%). Development of family policy would have also contributed to this change. Even

within the weak family policy countries, according to Gornick and Meyers’ family policy

index (2003), paid maternity leave is given for 15 weeks at 2000 (except the US) while the

Table 3 Prediction of working status of mothers with children under the age of 5 at 1999/2000

B Odds ratio Contribution

a B Odds ratio Contribution

a

Education completed age .05 1.05 0.56 .06 1.07 0.78

Chief Earner of household .89* 2.43 6.09 .80 2.23 4.55

Household Income .37** 1.44 33.24 .34** 1.40 23.71

Traditional Family Values .08 1.08 0.42 -.01 0.99 0.00

Family Policy (referred to strong family policy group)

-

13.10

Weak family policy -

-1.51** .22

Moderate family policy -

-1.03** .36

N 286 286

Nagelkerke R square 0.170 0.224

Data: EVS-WVS 1999/2000. ** Significant at the 0.01 level; * significant at the 0.05 level. Within 5 countries: Denmark (strong family policy country), Belgium, France, the Netherlands (moderate family policy countries), and the UK (weak family policy effort countries) – see table 1. a By the change in -2log-likelihood if the variable is removed from the model.

11

strong family policy offers 18 weeks. Mothers at 2000 would not have agonised about their

working decisions as much as those in the 1980s.

At 2000 again, the impact of family policy is clear – where there is stronger support,

mothers with young children are more likely to work. As for the other factors, household

income level is an exclusively powerful predictor at 2000, whilst whether a respondent is the

bread-winner of the household has become a less influential determinant – mothers now do

not choose working for bread but for more income. Having a job has become something for

more than just ‘bread’ (survival) at least within the most affluent societies including those in

this analysis. In some respect, this possibly is the win of the development of family policy –

mothers and children are now supported enough (but maybe only enough) not to choose

working only for bread. Meanwhile, the vanished impact of the education level at 2000 can

also be partly accounted for by the increase in women’s and mothers’ employment in general.

‘Having a job’ and ‘developing a career’ is not a privileged thing for (more) educated women

(mothers) any more. In other words, ‘being more educated’ does not generate a significantly

larger opportunity cost for not-working than ‘being less educated’.

If we simply construct three models on mothers’ working/not-working status at different

time points (Table 4) – because of the lack of information within the dataset, cases are female

respondents who have one or more children, some patterns over time can be more visible

including this ‘fade-away’ trend of the effect of traditional family values, although the three

models in this table are not directly comparable (this is not longitudinal data and cases are

different). The level of traditional family values contributed nearly as much as the education

level at 1981 yet lost most of its effect at 1990 and hardly contributed to the prediction at

2000. Whether the respondent is the chief earner of her household took the place of the most

important predictor in 1981/2, but was a little challenged by the income level at 1990 and

finally gave up its position to household income level at 2000. The education level appears to

have begun losing its impact since 1990.

Discussion Many studies on public opinion/attitudes have been questioned in proving the causal direction

from public opinion because of its vulnerability to surrounded conditions, which leads

Table 4 Change in impacts of factors on mothers’ employment status over time within 12 OECD countries

1981/2 1990 1999/2000

B Odds ratio

Contri-bution

a

B Odds ratio

Contri-bution

a

B Odds ratio

Contri-bution

a

Education completed age .10** 1.11 25.8 .14** 1.15 45.0 .10** 1.10 16.5

Chief Earner of household 1.49** 4.43 78.5 1.53** 4.60 102.0 1.34** 3.83 95.4

Household Income Scale .18** 1.20 49.6 .21** 1.23 76.2 .31** 1.37 146.7

Traditional Family Values -.22** .81 19.6 -.10* .90 3.8 -.08 .92 2.5

N 1,725 1,708 1,702

Nagelkerke R square 0.143 0.183 0.199

Data: EVS-WVS. ** Significant at the 0.01 level; * significant at the 0.05 level. 12 countries: Belgium, Canada, Denmark, France, Iceland, Ireland, Italy, Japan, Netherlands, Spain, UK and the US. a By the likelihood ratio test, which shows change in -2 log-likelihood if the variable is removed from the model.

12

researchers themselves to ask whether public opinion is the ex-ante causal factor for policy,

the ex-post legitimacy control factor for policy, or merely a dependant of institutions and key

social actors (van Oorschot, 2007: 134-135). As recently suggested (Pfau-Effinger, 2005; van

Oorschot, 2006; van Oorschot et al., 2008), however, if culture should be seen at the same

level with economic and political context for policy-making in the pursuit of cultural

explanation of welfare, we can conceptualise that culture has both impacts (or twofold effect)

which is also empirically examinable. Focusing on the cultural impact at the policy

implementation stage, I suggested that popular and stable values which collectively form the

cultural context of society should affect people’s attitudes to concrete policies. More

concretely, I argue that negative policy attitudes can lead to passive resistance to a specific

policy instrument such as policy recipients’ non-take-up. This was empirically supported

although findings in the above were limited. We saw that in the early 1980s working status of

mothers with young children was clearly dependent upon how strong traditional family values

they have even if their employment is strongly supported by family policy. Whilst data at

2000 showed that this impact faded away (mainly by change of the nature of mothers’

employment) - and admittedly analysis with the earlier data was also limited with the lack of

information within the dataset, it may not be critically limited to show a way of hypothesising

and empirically examining the ex-post effect of culture.

If a broader viewpoint of the policy process is taken, in fact, this cultural impact on

policy attitudes at the after stage of decision-making is another causal effect on social policy

at the before stage. When there is less public support for a policy by negative policy attitudes,

or if the non-take-up rate is high, policy reform, policy innovation or the introduction of a

new policy would become necessary, which is more answering to the causes of such passive

resistances.

When people’s policy attitudes are linked to their take-up decisions, the ex-post effect of

culture has practical implications for social policy making. Whilst ‘controlling legitimacy

support’ for a policy is rather abstract, affecting take-up decisions is more critical and closer

to the ultimate aim of ‘effective’ policy making. Policy-makers may be sensitive to how

successful a policy instrument is since the low take-up rate of a policy may cause a conflict

over the efficient use of welfare resources. Therefore, if mothers’ values matter in family

policy, for example, we may be asked to develop more sophisticated family policy – where

mothers are more strongly oriented to traditional family values, policies more reflecting their

preference for home childcare over public care (particularly when their children are at pre-

school age) can help bring about more efficient policy instruments. For another example,

where the public funded residential care services of professional caregivers are eschewed for

traditional care by relatives (Pfau-Effinger, 2005: 13-14) we may need to adopt a different

approach like more support for informal caregivers. That is, taking policy recipients’ values

into consideration could be a way of effective policy making. Taylor-Gooby and Wallace

(2009) also contend that the lack of understanding of service users’ values in the recent social

policy reform of the UK has damaged the legitimacy of as well as public trust in welfare

provisions. Possibly, this cultural impact has been and will be another pressure on social

policy to change and to be diversified.

Notes 1 Since the standardised measure across societies for the education level was not offered, an

imperfect measure – the age the respondent completed her education – was used instead for

the education level.

13

2 This raises a question which cannot be dealt with within the limit of this research – why

mothers’ traditional family values have declined for two decades whilst those of others,

including women as a whole, have not.

References Aalberg, T. (2003), Achieving Justice: Comparative Public Opinion on Income Distribution,

Leiden: Brill.

Alcock, P. (2001), ‘The Comparative Context’, in P. Alcock and G. Craig (Eds.),

International Social Policy: Welfare Regimes in the Developed World, Hampshire:

Palgrave.

Andreβ, H.-J., and Heien, T. (2001), ‘Four Worlds of Welfare State Attitudes? A Comparison

of Germany, Norway, and the United States’, European Sociological Review, 17(4):

337-356.

Arts, W. and Gelissen, J. (2002), ‘Three Worlds of Welfare Capitalism or More? A state-of-

the-art report’, Journal of European Social Policy, 12(2): 137-158.

Blekesaune, M. and Quadagno, J. (2003), ‘Public Attitudes toward Welfare State Policies: a

Comparative Analysis of 24 Nations’, European Sociological Review, 19(5): 415-427.

Boeri, T., Börsch-Supan, A. and Tabellini, G. (2001), ‘Would you like to shrink the welfare

state? A survey of European citizens’ Economic Policy, 16(32): 7-50.

Brewster, K. L. and Rindfuss, R. R. (2000), ‘Fertility and Women's Employment in

Industrialized Nations’, Annual Review of Sociology, 26: 271-296.

Bruning, G. and Plantenga, J. (1999), ‘Parental leave and Equal opportunities: experiences in

eight European countries’, Journal of European Social Policy, 9(3): 195-209.

Cnaan, R. A., Hasenfeld, Y., Cnaan, A. and Rafferty, J. (1993), ‘Cross-Cultural Comparison

of Attitudes Toward Welfare-State Programs: Path Analysis with Log-Linear Models’,

Social Indicators Research 29: 123-152.

Evans, E. J. (ed.) (1978), Social Policy 1830-1914: Individualism, Collectivism and the

Origins of the Welfare State, London: Routledge & Kegan Paul.

Freeman, R. and Rustin, M. (1999), ‘Introduction: Welfare, culture and Europe’, in P.

Chamberlayne, A. Cooper, R. Freeman and M. Rustin (eds.), Welfare and Culture in

Europe: Towards a New Paradigm in Social Policy (pp. 9-20), London: Jessica

Kingsley.

Gornick, J. C., and Meyers, M. K. (2003), Families That Work: Policies for Reconciling

Parenthood and Employment, New York: Russell Sage Foundation.

Gornick, J. C., Meyers, M. K. and Ross, K. E. (1997), ‘Supporting the Employment of

Mothers: Policy Variation Across Fourteen Welfare States’, Journal of European

Social Policy, 7(1), 45-70.

Gundelach, P. (1994), ‘National Value Differences: Modernization or Institutionalization?’,

International Journal of Comparative Sociology, 35(1-2), 37-58.

Hakim, C. (2000), Work-Lifestyle Choices in the 21st Century: Preference Theory, Oxford:

Oxford University Press.

Hancock, R., Pudney, S., Barker, G., Hernandez, M. and Sutherland, H. (2004), ‘The Take-

Up of Multiple Means-Tested Benefits by British Pensioners: Evidence from the

Family Resources Survey’, Fiscal Studies, 25(3), 279-303.

Hernanz, V., Malherbet, F. and Pellizzari, M. (2004), Take-Up of Welfare Benefits in OECD

Countries: A Review of the Evidence: OECD.

Hofstede, G. (2001), Culture's Consequences: Comparing Values, behaviors, Institutions, and

Organizations Across Nations (Second ed.), California: Sage

14

Inglehart, R. (1990), Culture Shift in Advanced Industrial Society, New Jersey: Princeton

University Press.

Kangas, O. and Rostgaard, T. (2007), ‘Preferences or institutions? Work-family life

opportunities in seven European countries’, Journal of European Social Policy, 17(3):

240-256.

Larsen, C. A. (2006), The Institutional Logic of Welfare Attitudes: How Welfare Regimes

Influence Public Support, Aldershot: Ashgate

Lockhart, C. (2001), Protecting the Elderly: How Culture Shapes Social Policy, Pennsylvania:

Pennsylvania State University Press.

Manow, P., and van Kersbergen, K. (2009), ‘Religion and the Western Welfare State - The

Theoretical Context’, in K. van Kersbergen and P. Manow (eds.), Religion, Class

Coalitions, and Welfare States (pp. 1-38), Cambridge: Cambridge University Press.

Marshall, T. H. (1972), ‘Value Problems of Welfare-Capitalism’, Journal of Social Policy,

1(1), 15-32.

O'Connor, J. S., and Robinson, G. (2008), ‘Liberalism, Citizenship and the Welfare State’, in

W. van Oorschot, M. Opielka and B. Pfau-Effinger (eds.), Culture and Welfare State:

Values and Social Policy in Comparative Perspective (pp. 29-49), Cheltenham, UK:

Edward Elgar.

Opielka, M. (2008), ‘Christian Foundations of the Welfare State: Strong Cultural Values in

Comparative Perspective’, in W. van Oorschot, M. Opielka and B. Pfau-Effinger

(eds.), Culture and Welfare State: Values and Social Policy in Comparative

Perspective (pp. 89-114), Cheltenham, UK: Edward Elgar.

Oyserman, D., and Uskul, A. K. (2008), ‘Individualism and Collectivism: Societal-Level

Processes with Implications for Individual-Level and Society-Level Outcomes’, in F. J.

R. van de Vijver, D. A. van Hemert and Y. H. Poortinga (eds.), Multilevel Analysis of

Individuals and Cultures (pp. 145-173), London: Lawrence Erlbaum Associates.

Page, B. I., Shapiro, R. Y. and Dempsey, G. R. (1987), ‘What Moves Public Opinion?’ The

American Political Science Review, 81(1): 23-44.

Pfau-Effinger, B. (2004), Development of Culture, Welfare States and Women's Employment

in Europe, Aldershot: Ashgate.

Pfau-Effinger, B. (2005), ‘Culture and Welfare State Policies: Reflections on a Complex

Interrelation’, Journal of Social Policy, 34(1): 3-20.

Riphahn, R. T. (2001), ‘Rational Poverty or Poor Rationality? The Take-Up of Social

Assistance Benefits’, Review of Income and Wealth, 47(3): 379-398.

Schwartz, S. H. (1994), ‘Beyond Individualism/Collectivism: New Cultural Dimensions of

Values’, in U. Kim, H. C. Triandis, C. Kagitcibasi, S. Choi and G. Yoon (eds.),

Individualsim and Collectivism: Theory, Method and Applications (pp. 85-119),

London: Sage.

Schwartz, S. H. (1999), ‘A Theory of Cultural Values and Some Implications for Work’,

Applied Psychology: An International Review, 48(1), 23-47.

Stjernø, S. (2008), ‘Social Democratic Values in the European Welfare States’, in W. van

Oorschot, M. Opielka and B. Pfau-Effinger (eds.), Culture and Welfare State: Values

and Social Policy in Comparative Perspective (pp. 50-70), Cheltenham, UK: Edward

Elgar.

Svallfors, S. (1997), ‘Worlds of Welfare and Attitudes to Redistribution: A Comparison of

Eight Western Nations’, European Sociological Review, 13(3): 283-304.

Svallfors, S. (2007), ‘Class and Attitudes to Market Inequality: A Comparison of Sweden,

Britain, Germany, and the United States’, in S. Svallfors (ed.), The Political Sociology

of the Welfare States: Institutions, Social Cleavages, and Orientations (pp. 189-222),

Stanford, California: Stanford University Press.

15

Tabachnick, B. G., and Fidell, L. S. (2007), Using Multivariate Statistics (5th ed.), London:

Pearson Education.

Taylor-Gooby, P. (2008), ‘Choice and Values: Individualised Rational Action and Social

Goals’, Journal of Social Policy, 37(2): 167-185.

Taylor-Gooby, P. and Wallace, A. (2009), ‘Public Values and Public Trust: Responses to

Welfare State Reform in the UK’, Journal of Social Policy, 38(3): 401-419.

van de Vijver, F. J. R., van Hemert, D. A. and Poortinga, Y. H. (2008), ‘Conceptual Issues in

Multilevel Models’, in F. J. R. van de Vijver, D. A. van Hemert and Y. H. Poortinga

(eds.), Multilevel Analysis of Individuals and Cultures (pp. 3-26), London: Lawrence

Erlbaum Associates.

van Kersbergen, K. (1995), Social Capitalism: A Study of Christian Democracy and the

Welfare State, London: Routledge.

van Kersbergen, K., and Kremer, M. (2008), ‘Conservatism and the Welfare State:

Intervening to Preserve’, in W. van Oorschot, M. Opielka and B. Pfau-Effinger (eds.),

Culture and Welfare State: Values and Social Policy in Comparative Perspective (pp.

71-88), Cheltenham, UK: Edward Elgar.

van Oorschot, W. (1991), ‘Non-Take-Up of Social Security Benefits in Europe’, Journal of

European Social Policy, 1(1): 15-30.

van Oorschot, W. (2000), ‘Who should get What, and Why? On deservingness criteria and the

conditionality of solidarity among the public’, Policy & Politics, 28(1): 33-48.

van Oorschot, W. (2006), ‘Making the Difference in Social Europe: Deservingness

Perceptions among Citizens of European Welfare States’, Journal of European Social

Policy, 16(1): 23-42.

van Oorschot, W. (2007), ‘Culture and Social Policy: a Developing Field of Study’,

International Journal of Social Welfare, 16(2): 129-139.

van Oorschot, W. (2008), ‘Popular Deservingness Perceptions and Conditionality of

Solidarity in Europe’, in W. van Oorschot, M. Opielka and B. Pfau-Effinger (eds.),

Culture and Welfare State: Values and Social Policy in Comparative Perspective (pp.

268-288), Cheltenham, UK: Edward Elgar.

van Oorschot, W., and Halman, L. (2000), ‘Blame or Fate, Individual or Social? An

international comparison of popular explanations of poverty’, European Societies,

2(1): 1-28.

van Oorschot, W., Opielka, M. and Pfau-Effinger, B. (2008), ‘The culture of the welfare state:

historical and theoretical arguments’, in W. van Oorschot, M. Opielka and B. Pfau-

Effinger (eds.), Culture and Welfare State: Values and Social Policy in Comparative

Perspective (pp. 1-26), Cheltenham, UK: Edward Elgar.

Whiteley, P. (1981), ‘Public Opinion and the Demand for Social Welfare in Britain’, Journal

of Social Policy, 10(4): 453-476.

[DATA]

European and World Values Surveys four-wave integrated data, 1981-2004 (2006)

v.20060423.