Journal of Family Issues filenational perspective. Recent investigations for the general population strongly suggest that macrolevel factors, particularly gender inequalities,

http://jfi.sagepub.com

Journal of Family Issues

DOI: 10.1177/0192513X06296427 2007; 28; 399 Journal of Family Issues

Karsten Hank and Hendrik Jürges European Perspective

Gender and the Division of Household Labor in Older Couples: A

http://jfi.sagepub.com/cgi/content/abstract/28/3/399 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

can be found at:Journal of Family Issues Additional services and information for

http://jfi.sagepub.com/cgi/alerts Email Alerts:

http://jfi.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://jfi.sagepub.com/cgi/content/refs/28/3/399 Citations

at Universitaet Mannheim on February 3, 2009 http://jfi.sagepub.comDownloaded from

http://jfi.sagepub.com/cgi/alerts

http://jfi.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://jfi.sagepub.com/cgi/content/refs/28/3/399


Gender and the Division ofHousehold Labor inOlder Couples

A European Perspective

Karsten HankHendrik JürgesMannheim Research Institute for the Economics of Aging,University of Mannheim, Germany

Using microdata from the 2004 Survey of Health, Ageing and Retirement inEurope (SHARE), this study takes a cross-national perspective to investigatethe division of household labor among older couples (aged 50 years or more).Across nine continental European countries, the authors find considerablevariation in the overall distribution of housework between partners, withmore egalitarian countries in northern Europe and more traditional countriesin southern Europe. A multilevel analysis shows that about half of thebetween-country variance in the division of housework is due to differencesin older couples’ characteristics, but that there are no country-specific effectsof the main microlevel explanatory variables. Finally, the authors find a sig-nificant effect of macrolevel gender inequalities on couples’ division ofhousework, suggesting that older couples living in more gender-egalitariancountries are more likely to exhibit an equal sharing of household labor.

Keywords: gender; division of household labor; older couples; Europe;SHARE

Research on productive aging is growing rapidly (cf. Morrow-Howell,Hinterlong, & Sherraden, 2001; O’Reilly & Caro, 1994). Most of the

related literature deals with involvement of the elderly in volunteering (e.g.,Caro & Bass, 1995; Erlinghagen & Hank, 2006), grandparenting (e.g.,Hayslip & Kaminski, 2005; Pebley & Rudkin, 1999), or other work usuallyperformed for parties outside the individual’s household. However, workperformed within one’s own household also has a substantial economicvalue. Without home production, one would have to buy substantially

Journal of Family IssuesVolume 28 Number 3March 2007 399-421

© 2007 Sage Publications10.1177/0192513X06296427

http://jfi.sagepub.comhosted at

http://online.sagepub.com

399



more household services on the market (e.g., de Ruijter, Treas, & Cohen,2005; Stoller & Cutler, 1993). Although gender differences in other typesof unpaid or informal work are also recognized (e.g., Herzog & Morgan,1992; Hook, 2004), they are particularly pronounced in the division ofhousework between spouses. Despite some changes across cohorts, withmore recently born women doing less and their male partners doingsomewhat more (both relative and total), today men generally contributeat most one third of “core” housework tasks (e.g., Bianchi, Milkie, Sayer,Folbre, & Matheson, 2000; Breen & Cooke, 2005; Shelton & John, 1996).

Several studies dealing with postretirement changes in couples’ divisionof household labor show that the total amount of housework done mayincrease after retirement for both men and women, but that the traditionalpreretirement pattern largely persists (e.g., Dorfman, 1992; Solomon,Acock, & Walker, 2004; Szinovacz, 2000). So far, however, the division ofhousehold labor among older couples has not been analyzed in a cross-national perspective. Recent investigations for the general populationstrongly suggest that macrolevel factors, particularly gender inequalities,play a significant role in the distribution of housework between spouses(e.g., Breen & Cooke, 2005; Davis & Greenstein, 2004; Fuwa, 2004).Using data on couples aged 50 and older derived from the Survey ofHealth, Ageing and Retirement in Europe (SHARE), this article estimateshierarchical linear models to investigate the relationship between societalcontext and the division of routine household labor in nine continentalEuropean countries. Before presenting the empirical analysis, we willbriefly review the linkage between micro- and macrolevel aspects ofgender and housework.

400 Journal of Family Issues

Authors’ Note: Both authors contributed equally to this article. We are grateful for commentsby two anonymous reviewers. This research was supported by the European Community FifthFramework Program’s Quality of Life project under EC Contract QLK6-2002-002426(AMANDA). This article is based on data from Early Release 1 of SHARE 2004, which is pre-liminary and may contain errors that will be corrected in later releases. The SHARE data col-lection has been funded primarily by the European Commission through the Fifth FrameworkProgram (Quality of Life Project QLK6-CT-2001-00360). Additional funding came from theU.S. National Institute on Aging (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30AG12815, Y1-AG-4553-01, and OGHA 04-064). Data collection in Austria (through theAustrian Science Fund, FWF) and Switzerland (through BBW/OFES/UFES) was nationallyfunded. Please address correspondence to Karsten Hank, Mannheim Research Institute for theEconomics of Aging, University of Mannheim, Building L 13, 17, D-68131 Mannheim,Germany; e-mail: [email protected].



Gender and the Division of Housework

Various economic models play a prominent role in much of the literatureon household labor. Although the new home economics’ approach, put forwardby Becker (1981), proposes that men and women specialize to maximizehousehold utility or efficiency, the resource–bargaining power perspectivefocuses on power relations in the family (based on, for example, educationalor income differentials between the partners), and the economic dependencymodel is centered on the assumption that women are “forced” to exchangehousehold labor in return for economic support from a male breadwinner (cf.Brines, 1993; Greenstein, 2000, for example). These theories are compatiblewith general formulations of the relative resource hypothesis (a person withhigher income will do less housework) and the time availability hypothesis (aperson who spends more time in paid work will spend less time in house-work), “which are putatively gender neutral, emphasize choice, and assumethat housework allocation is governed by the rules and principles of exchangerelations” (Coltrane, 2000, p. 1214; see also Shelton & John, 1996, pp. 304ff.).

However, it has been suggested that “women’s employment, time avail-ability, resources, conscious ideology, and power do not account for whywives still do the bulk of family work” (Thompson & Walker, 1989, p. 857)regardless of demographic or life-course characteristics. The partner’sgender appears to be so influential that it is often considered to be the sin-gle most important determinant of the division of household labor. Theorieson socialization–gender role attitudes, for example, contend that peoplesocialized to believe in gender-segregated work will conform to thosebeliefs (e.g., Coverman, 1985; Cunningham, 2005). Thus, men and womenwith traditional attitudes are expected to share less housework, whereasmen and women with nontraditional attitudes are expected to share house-work more equally. It is assumed that people are automatically socializedinto rigid gender roles from childhood onward, going along with the devel-opment of relatively fixed attitudes and/or deeply gendered personalities.These strict assumptions are rejected by more recent gender constructiontheories that incorporate the symbolic and performance dimensions of gender(cf. Shelton & John, 1996). As Coltrane (2000) posits, “Doing specific house-hold tasks provides opportunities to demonstrate to oneself and to others thatone is a competent member of a sex category with the capacity and desire toperform appropriately gendered behaviors” (p. 1213). Thus, housework pro-duces not only household commodities but also gendered identities through-out the life course.

Hank, Jürges / Division of Household Labor 401



Gender ideology determines what a proper gender role is. Just as genderideologies vary across individuals (e.g., Greenstein, 1996), the social con-struction of gender is highly context dependent and varies across nations orcultures. Mason (1997) defines the societal gender system as

the socially constructed expectations for male and female behavior that arefound (in variable form) in every human society. A gender system’s expecta-tions prescribe a division of labor and responsibilities between women andmen and grant different rights and obligations to them. (p. 158)

Resulting macrolevel gender inequalities—that may be promoted or ame-liorated by the welfare state (Orloff, 1996; see also Geist, 2005)—materi-alize in various spheres (cf. Huber, 1990) such as the educational system(e.g., Jacobs, 1996), the labor market (e.g., Chang, 2000), or the politicalarena (e.g., Elder, 2004). Naturally, they are also reflected in spouses’ divi-sion of work in the family (e.g., Sarkisian & Gerstel, 2004; Sundström &Duvander, 2002; Thompson & Walker, 1989) and in the household. Thus,Coltrane (2000) suggests that the almost universally observed pattern ofhousehold labor

can only be understood by attending to the symbolic significance of house-hold labor in the social construction of gender and by analyzing the social,cultural, economic, and political contexts in which men and women formfamilies, raise children, and sustain households. (p. 1208)

Starting from Baxter’s (1997) five-country study—covering the UnitedStates, Sweden, Norway, Canada, and Australia—a number of studies haveexplicitly investigated the division of housework in advanced industrializedsocieties from a cross-national perspective (for an analysis of less devel-oped countries, see, for example, Sanchez, 1993, 1994). A universal find-ing is that wives’ contribution to household chores is still greater than theirhusbands’, even in the most egalitarian countries (e.g., Davis & Greenstein,2004). More differentiated insights can be derived from recent work usingmultilevel modeling. For example, Batalova and Cohen (2002), who focuson the role of premarital cohabitation, show that national cohabitation ratesin countries with higher levels of overall gender equality have equalizingeffects on couples’ division of housework regardless of their own cohabita-tion experience. Fuwa (2004) elaborates on the role of macrolevel genderinequalities, arguing that “male control over the political economy andmale dominated ideologies at the macrolevel may act as ‘discount factors’




against the power of individual women’s resources” (p. 752; see alsoBlumberg, 1984). Thus, she expects that individual-level factors will haveweaker effects on the division of household labor for women who live incountries with less pronounced gender equality—and vice versa. Using thesame data source (the 1994 International Social Survey Programme; ISSP)and selection of 22 countries on which the study by Batalova and Cohen(2002) is based, Fuwa (2004) indeed finds that women living in less egalitar-ian countries benefit less from their individual assets in the negotiation ofhousework. This is supported by an analysis of a subset of countries from the1994 ISSP, which shows that equal sharing of household tasks is particularlyrare in countries with a conservative welfare state regime, independent ofthe partners’ relative resources, time availability, or gender ideology (Geist,2005). Finally, utilizing the 2002 ISSP on Gender Roles, Stier and Lewin-Epstein (2005) examine the effects of employment-supportive policies,gender inequalities in the labor market, and general attitudes toward genderroles on households’ division of unpaid work in 25 countries. Although it isshown that employment policies have no direct effect on couples’ division oflabor, the authors’ findings suggest that gender inequalities in the labormarket and a country’s gender ideologies do affect the level of gender equal-ity in the family.

To our knowledge, though, no cross-national research has been carriedout yet that pays particular attention to the gendered division of householdlabor among older couples. Filling in this gap for continental Europe, ourstudy complements recent time-use research that points to significant inter-gender and intercountry differences in time-use patterns at older ages (cf.Croda & Gonzalez-Chapela, 2005; Gauthier & Smeeding, 2003). It alsoadds a European perspective to the so far almost exclusively U.S.-centeredliterature on the division of housework in later life, particularly after retire-ment. And finally, it investigates possible interactions between couple char-acteristics and the household’s country of residence in determining patternsof household labor.

Method

Data

The data for our study are drawn from the 2004 SHARE (see http://www.share-project.org for more information). The SHARE is modeled closelyafter the U.S. Health and Retirement Study, and it is the first European data




set to combine extensive cross-national information about the socioeco-nomic status, health, and family relationships of the elderly population (seeBörsch-Supan et al., 2005). Release 1 of the data contains informationabout some 22,000 individuals aged 50 or older from 15,000 households in10 countries, representing Europe’s economic, social, institutional, and cul-tural diversity from Scandinavia to the Mediterranean, including Sweden,Denmark, Germany, the Netherlands, France (excluded from our studybecause of missing macrolevel information), Switzerland, Austria, Italy,Spain, and Greece. Probability samples have been drawn in each participat-ing country; the weighted average household response rate in the face-to-facepart of the survey is 62 percent (a thorough description of methodologicalissues is contained in Börsch-Supan & Jürges, 2005).1 Our analytic samplewas restricted (a) to respondents living in a marital or nonmarital union at thetime of the interview and (b) to couples for whom at least one partner filledout the survey’s self-completion questionnaire (which includes the questionon which our dependent variable is based). This results in a total of 4,135observations (see Table 1 for details).

Variables

This article deals with “routine housework” (Coltrane, 2000, p. 1210). Theconstruction of the dependent variable modifies Davis and Greenstein’s(2004) measure, taking advantage of the fact that the SHARE provides bothpartners’ assessments of who takes the primary responsibility for routinehousehold chores (see Lee & Waite, 2005, for a discussion of alternative mea-sures). This is a major improvement over existing data sets such as the ISSP.The (generic) English version of the SHARE questionnaire asks, “Who in thecouple takes or took the main responsibility for cooking, cleaning the house,laundry and ironing?” with five answer categories: “myself only,” “myselfmainly,” “myself and my partner equally,” “my partner mainly,” and “mypartner only” (coded 1 through 5). Because this question was posed to bothpartners, responses were relabeled to distinguish husbands from wives. Toaccount for possible discrepancies in spouses’ responses (cf. Kamo, 2000;Lee & Waite, 2005), we use the mean of their respective answers. This infor-mation has been recoded, resulting in a variable ranging from 0 (both part-ners agree that the wife does all housework) to 1 (both partners agree that thehusband does all housework).2

On the right-hand side of the regression, we use a set of standard micro-level explanatory variables including the partners’ age, education, (gross)income,3 employment status, and health, as well as information about the




405

Tabl

e 1

Des

crip

tive

Sta

tist

ics

Aus

tria

Ger

man

ySw

eden

Net

herl

ands

Spai

nIt

aly

Den

mar

kG

reec

eSw

itzer

land

Gen

der

divi

sion

of

hous

ehol

d la

bor

.203

.242

.262

.252

.156

.179

.295

.130

.245

Prop

ortio

n of

cou

ples

with

.1

05.1

72.1

67.1

66.0

88.1

12.2

36.0

65.1

43eq

ual d

ivis

ion

of la

bora

Mal

e in

com

e <

fem

ale

inco

me

.105

.101

.182

.095

.050

.056

.157

.053

.111

Mal

e in

com

e >

fem

ale

inco

me

.630

.681

.550

.738

.739

.678

.519

.692

.635

Mal

e in

labo

r fo

rce

.241

.315

.405

.348

.255

.227

.467

.408

.429

Fem

ale

in la

bor

forc

e.2

00.2

93.4

50.2

70.1

26.1

54.4

62.2

14.3

81U

nmar

ried

cou

ple

.038

.048

.083

.039

.076

.017

.077

.012

.048

Cou

ple’

s m

ean

age

62.2

6862

.900

63.6

6462

.004

65.3

0962

.922

61.2

2860

.511

63.2

35M

ale–

fem

ale

age

diff

eren

ce3.

005

2.51

32.

898

2.73

62.

730

3.66

33.

014

5.75

02.

862

Mal

e,hi

gh e

duca

tiona

l deg

reeb

.272

.366

.298

.261

.070

.085

.336

.220

.312

Fem

ale,

high

edu

catio

nal d

egre

eb.1

43.2

35.3

12.1

39.0

35.0

98.3

79.1

44.1

90H

ouse

hold

inco

me

(nat

ural

log)

9.45

610

.178

10.6

8910

.527

9.13

39.

202

10.8

579.

169

10.4

27M

ale

retir

ed f

rom

labo

r fo

rce

.697

.589

.554

.495

.651

.746

.484

.557

.519

Fem

ale

retir

ed f

rom

labo

r fo

rce

.470

.431

.461

.140

.135

.368

.427

.241

.392

Num

ber

of c

hild

ren

2.10

71.

939

2.54

02.

434

2.72

72.

137

2.35

32.

079

2.22

8M

ale

care

s fo

r gr

andc

hild

ren

.277

.284

.339

.392

.267

.222

.362

.199

.201

Fem

ale

care

s fo

r gr

andc

hild

ren

.339

.332

.415

.446

.349

.305

.396

.230

.217

Mal

e lim

ited

by h

ealth

pro

blem

s.4

37.4

79.3

90.3

67.3

93.3

27.3

87.2

46.2

75Fe

mal

e lim

ited

by h

ealth

pro

blem

s.4

25.5

03.4

23.4

76.4

87.3

76.4

05.2

50.3

33O

nly

mal

e re

spon

se a

vaila

ble

.036

.032

.099

.068

.023

.066

.048

.030

.053

Onl

y fe

mal

e re

spon

se a

vaila

ble

.050

.031

.106

.080

.023

.051

.066

.046

.016

GE

M.7

46.8

491.

000

.888

.583

.181

.979

.000

.749

N41

958

760

566

434

141

035

156

918

9

Not

e:G

EM

=U

nite

d N

atio

ns’g

ende

r em

pow

erm

ent m

easu

re.

a. S

hare

of

coup

les

with

val

ues

equa

l to

or la

rger

than

.5 f

or th

e di

visi

on o

f ho

usew

ork

vari

able

.b.

Int

erna

tiona

l Sta

ndar

d C

lass

ific

atio

n of

Edu

catio

n ca

tego

ries

4 o

r hi

gher

.So

urce

:SH

AR

E 2

004

(Rel

ease

1),

auth

ors’

calc

ulat

ions

.



partners’ marital status and family responsibilities (see Table 1 for descriptivestatistics). One advantage of our data is that we have true couple information,that is, individual data for each question asked of each partner. This advan-tage also complicates matters because each individual variable type enters theanalyses twice. Overall, we have thus opted for rather parsimonious butrobust specifications.

We measure household income by the relative income position in eachcountry. More precisely, we compute country-specific income quintiles,which enter the regression as a set of dummy variables. This has the addi-tional advantage of combining flexibility in functional specification with easeof interpretation. We treat differences in the partners’ income as an indicatorof relative resources. Relative income of partners is coded as a set of dummyvariables for which the baseline category represents couples whose individ-ual income is in the same country-specific (individual) income quintile. Thus,the two dummy variables indicate whether the female partner is in a higheror lower income quintile than the male partner.

Employment status is used as an indicator of time availability. In theSHARE, respondents are asked to self-report employment status by select-ing one of the following categories: (a) retired, (b) employed or self-employed, (c) unemployed, (d) permanently sick or disabled, and (e)homemaker. We code as working all respondents who say they are currentlyemployed or self-employed. The employment status variable is also used tocreate an indicator for past labor force status (retired), which may have last-ing effects for retired couples (see below for a discussion).

Because less traditional gender ideologies have been shown to be closelyassociated with cohabitation (Batalova & Cohen, 2002; Cunningham, 2005;South & Spitze, 1994), union type is used to account for the gender ideol-ogy dimension at the level of couples.

We measure the age (or cohort; this is equivalent in cross-sectional data)of a couple by the average age of both partners. Age differences betweenpartners, which might indicate intracouple power relationships, are alsoincluded as control variables. They are measured as the male age minus thefemale age. Education enters the analysis as one dummy variable for “high”education for each partner, where high education is equivalent to anInternational Standard Classification of Education (ISCED) code of 4 orhigher (beyond upper secondary education).

Past and current family responsibilities are measured by the numberof children the couple has and by a dummy variable that equals 1 if therespondent said that during the past 12 months, he or she has regularly or




occasionally looked after grandchildren without the presence of theparents.

The SHARE contains a large amount of information about the respon-dents’ health status (including specific diagnoses, symptoms, etc.). In thecontext of our study, we are particularly interested in the effect of health-related limitations on daily activities (such as housework). Our health indi-cator is derived from a global activity limitation question: “For the past 6months at least, to what extent have you been limited because of a healthproblem in activities people usually do? (1) severely limited, (2) limited, butnot severely, (3) not limited.” In the empirical part, we do not distinguishbetween degrees of limitation.

Our main concern, however, is the role of societal factors in older cou-ples’ division of household labor. Like Batalova and Cohen (2002) andFuwa (2004), we use the United Nations’ gender empowerment measure(GEM; see United Nations Development Program, 2004) as a core measureof macrolevel gender inequalities. The GEM is an index based on the per-centage of parliamentary seats held by women, the percentage of femaleadministrators and managers, the percentage of professional and technicalworkers who are women, and women’s share of earnings income. It rangesfrom 0 to 1, where higher scores represent greater levels of empowermentfor women. Because the GEM is not available for France, this country wasexcluded from our analysis.

Analytical Strategy

To investigate the interaction between micro- and macrolevel determinantsof older couples’ division of household labor, we estimate hierarchical linearmodels (Bryk & Raudenbush, 1992). This and similar kinds of multilevelmodeling have become a popular alternative to estimating separate modelsfor each country (or other levels of context) in the analysis (see Teachman &Crowder, 2002, for a review). We follow a stepwise procedure, starting withthe “empty” Model 0, which includes no independent variables at all andexamines the overall couple- and country-level variances. In Model 1, all couple-level characteristics are entered into the regression with fixed effectsacross countries. In Model 2, we allow the coefficients of the three maincouple-level variables—the male–female income gap, the partners’ employ-ment status, and union type—to vary across countries. Finally, Model 3includes cross-level interactions between the GEM and the intercept as wellas the slopes of the three main couple-level variables.




Empirical Findings

Before presenting our multivariate results, we will briefly discuss twomain descriptive findings, displayed in Figure 1. First, there is an obviousnorth–south divide in the gender division of labor in the SHARE countries.This is reflected in the distribution of the proportion of couples in which thepartners share household tasks equally (including the small number of caseswhere the husband does more). Although, for example, in Greece and Spainless than 10 percent of couples aged 50 and older exhibit an equal division oflabor in the household, this is the case in about 17 percent of Dutch, German,and Swedish couples. The top rank is held by the Danes, where one out offour couples shares core household tasks equally.4 Second, there is strongindication for a close relationship between the division of household laboramong older couples and macrolevel gender inequalities.5 Plotting the pro-portion of couples with an equal division of housework against the GEMreveals a clear positive association. The four countries with above averageproportions of older couples sharing housework equally—the Netherlands,Sweden, Germany, and Denmark—are also the ones with the highest GEMscores (.8 or more). Similar associations are also found with single items ofGEM, such as the proportion of female members of parliament or thefemale–male wage ratio (details not shown here).

We now turn to the multivariate analysis (see Table 2; note that positivecoefficients indicate a larger share of the male partner in total housework).In Model 0, the between-country variance of the intercept is statisticallysignificant. The variance components in this empty model suggest thatabout 10 percent of the overall variance in the division of housework is dueto between-country variance (cf. Bryk & Raudensbush, 1992). A similarorder of magnitude is reported in the studies by Fuwa (2004) and Stier andLewin-Epstein (2005), for example. Including couple-level variables inModel 1 reduces the variance of country-level intercepts by almost half(from .0030 to .0017) and the couple-level variance by 7 percent (from.0279 to .0259). Turning to our main microlevel explanatory variables, wefind an asymmetric effect of relative income: Men with a lower income thantheir female partners tend to do somewhat more housework than those hav-ing about the same income as their partners (where equal income is definedas being in the same income quintile), but the difference is not statisticallysignificant (cf. Bittman, England, Sayer, Folbre, & Matheson, 2003).However, if the wife earns less than her husband, the husband’s share inhousehold duties is significantly lower than in couples with about equal




income. With regard to time availability, we find that men’s participationin the labor force also decreases males’ participation in household chores,whereas if the female partner engages in paid work, her husband’s share ofhousehold labor increases. Finally, living in a nonmarital union also increasesmen’s participation in housework strongly, supporting the view that cohabi-tation goes hand in hand with less traditional gender ideologies.

Both male and female higher education, which is also likely to be posi-tively correlated with less traditional gender ideologies, contributes to amore gender-equal division of housework (e.g., Coltrane, 2000). Older


Figure 1Gender Empowerment Measure and Shares of Couples With

Egalitarian Division of Household Labor Across Europe

Note: at = Austria; ch = Switzerland; de = Germany; dk = Denmark; es = Spain; gr = Greece;it = Italy; nl = Netherlands; se = Sweden.Source: SHARE 2004 (Release 1), authors’ representation.



410

Tabl

e 2

Det

erm

inan

ts o

f C

oupl

es’D

ivis

ion

of H

ouse

wor

k:R

esul

ts o

f H

iera

rchi

cal L

inea

r M

odel

s

Mod

el 0

Mod

el 1

Mod

el 2

Mod

el 3

βSE

βSE

βSE

βSE

Inte

rcep

t.2

176**

.018

4.2

177**

.014

7.2

175**

.014

1.2

172**

0.00

64G

EM

.116

8**0.

0194

Mal

e in

com

e <

fem

ale

inco

mea

.003

5.0

097

−.00

31.0

132

−.00

320.

0111

GE

M.0

760†

0.03

41M

ale

inco

me

> f

emal

e in

com

ea−.

0234

**.0

065

−.02

46*

.007

7−.

0246

*0.

0088

GE

M.0

089

0.02

50M

ale

in la

bor

forc

eb−.

0454

**.0

111

−.04

61**

.011

7−.

0452

**0.

0119

GE

M.0

227

0.02

12Fe

mal

e in

labo

r fo

rceb

.061

8**.0

079

.061

6**.0

086

.062

0**.0

096

GE

M.0

115

.025

2U

nmar

ried

cou

ple

.049

3**.0

121

.060

0**.0

160

.052

7*.0

177

GE

M−.

0364

.059

2C

oupl

e’s

mea

n ag

e−.

0021

**.0

005

−.00

21**

.000

5−.

0022

**.0

005

Mal

e–fe

mal

e ag

e di

ffer

ence

−.00

02.0

006

−.00

02.0

006

−.00

01.0

006

Mal

e,hi

gh e

duca

tiona

l deg

ree

.014

7*.0

067

.015

2*.0

066

.015

0*.0

066

Fem

ale,

high

edu

catio

nal d

egre

e.0

161*

.007

5.0

157*

.007

5.0

170*

.007

5H

ouse

hold

inco

me,

1st q

uint

ile−.

0118

.008

3−.

0113

.008

2−.

0123

.008

3H

ouse

hold

inco

me,

2nd

quin

tile

.010

9.0

082

.010

4.0

081

.010

2.0

081

Hou

seho

ld in

com

e,4t

h qu

intil

e.0

099

.008

2.0

099

.008

1.0

093

.008

2H

ouse

hold

inco

me,

5th

quin

tile

−.00

80.0

084

−.00

68.0

084

−.00

71.0

084



411

Mal

e re

tired

fro

m la

bor

forc

eb−.

0149

.011

6−.

0156

.011

5−.

0139

.011

6Fe

mal

e re

tired

fro

m la

bor

forc

eb.0

226**

.007

4.0

234**

.007

3.0

244**

.007

4N

umbe

r of

chi

ldre

n−.

0106

**.0

020

−.01

05**

.002

0−.

0108

**.0

020

Mal

e ca

res

for

gran

dchi

ldre

n.0

145†

.008

4.0

147

.008

4.0

148†

.008

4Fe

mal

e ca

res

for

gran

dchi

ldre

n−.

0178

.008

1−.

0174

†.0

081

−.01

77*

.008

1M

ale

limite

d by

hea

lth p

robl

ems

.000

8.0

056

.000

7.0

055

.000

4.0

055

Fem

ale

limite

d by

hea

lth p

robl

ems

.015

2**.0

054

.014

7**.0

054

.014

2**.0

054

Onl

y m

ale

resp

onse

ava

ilabl

e.0

422**

.011

4.0

413**

.011

4.0

409**

.011

4O

nly

fem

ale

resp

onse

ava

ilabl

e−.

0310

**.0

110

−.03

05**

.010

9−.

0298

**.0

110

Var

ianc

e co

mpo

nent

sIn

terc

ept

.003

0**.0

017**

.001

7**.0

002**

Mal

e in

com

e <

fem

ale

inco

me

.000

7.0

001

Mal

e in

com

e >

fem

ale

inco

me

.000

2.0

002

Mal

e in

labo

r fo

rce

.000

1.0

001

Fem

ale

in la

bor

forc

e.0

001

.000

2U

nmar

ried

cou

ple

.001

0.0

008

Lev

el 1

.027

9.0

259

.025

7.0

257

Not

e:G

EM

=U

nite

d N

atio

ns’g

ende

r em

pow

erm

ent m

easu

re.

a. R

efer

ence

cat

egor

y:m

ale

inco

me

=fe

mal

e in

com

e.b.

Ref

eren

ce c

ateg

ory:

mal

es (

fem

ales

,res

pect

ivel

y) w

ho a

re n

eith

er e

mpl

oyed

nor

ret

ired

.† si

gnif

ican

t at 1

0%. * si

gnif

ican

t at 5

%. **

sign

ific

ant a

t 1%

.So

urce

:SH

AR

E 2

004

(Rel

ease

1),

auth

ors’

calc

ulat

ions

.




couples (i.e., those with a higher mean age) exhibit the expected more tra-ditional pattern of housework, whereas the age difference between the part-ners has no significant effect. Similarly, the coefficients of the dummyvariables for different levels of household income turn out to be statisticallyinsignificant. However, there is some indication for a nonlinear relationshipbetween men’s participation in household tasks and household income: Thenegative coefficients for the lowest and highest income quintiles suggestthat the contribution of the male partner might be lowest at the upper andlower bounds of the income distribution.

If the woman has retired, her husband’s share of household laborincreases. At first glance, this might seem to be a surprising result, contra-dicting findings of previous studies (e.g., Szinovacz, 2000). In our model,however, the reference category consists of women who are neither gainfullyemployed nor retired. Because these are mostly housewives, retirement in ourmodel is an indicator of women’s previous labor force participation, whichis likely to facilitate a somewhat more equal division of housework betweenthe partners even after retirement (assuming continuity in household roles;e.g., Dorfman, 1992).6 Male retirement tends to be negatively correlatedwith the dependent variable, but the respective coefficient is not statisticallysignificant.

Family obligations have an effect on older couples’ division of house-hold labor in the senses that an increase in the number of children is asso-ciated with a decrease in males’ participation in household chores and thatmen caring for grandchildren also take larger responsibilities within theirown households. Limitations by health problems are not significant if theyaffect men, but they do result in a stronger engagement of the husband inhousehold duties if the female partner’s health is limited. As a final couple-level control variable, we use information about whether only the male oronly the female partner answered the question about the division of house-work. Consistent with previous studies (e.g., Kamo, 2000), men’s involve-ment in household chores is reported to be stronger if the informant is aman himself, whereas it appears to be significantly weaker if only thewife’s response is available (see below for further discussion).

Allowing the relative resources, time availability, and gender ideologyindicators to vary across countries in Model 2 does not change the coeffi-cients derived from the previous model. The variance components for theslopes of the respective variables are not statistically significant, whichmeans that their effects do not differ by country. Model 3 finally includesthe GEM’s effect on the intercept and on the slopes of the male–female




income gap, the partners’ employment status, and union type. The coeffi-cient for the effect of the GEM on the intercept is .1168 and is statisticallysignificant. This means that holding constant microlevel characteristics,men’s contribution to housework in more gender-egalitarian countries issubstantially larger than elsewhere. Moreover, including the GEM reducesthe remaining between-country variance of the intercept by almost 90percent, from .0017 to .0002! However, as already indicated in Model 2,the coefficients for the main microlevel variables generally do not varybetween countries with higher or lower GEM scores. The only exception isthe marginally significant effect of the GEM on the slope of the “maleincome < female income” dummy, suggesting that men in countries withhigher values of GEM contribute somewhat more housework if they earnless than their female partner, whereas this is not the case in countries withlower values of GEM.

Following a reviewer’s recommendation, we also estimated Model 3separately for men and women (see appendix). As expected from our ownprevious findings, the predicted average division of household labor islarger in the male model (.2317) than in the female model (.2031). Inparticular, the positive association between men’s contribution to house-hold chores and cohabitation or health limitations of the female partnerappears to be stronger if the interviewee is male. Moreover, the varianceof country-level intercepts is larger in the male model (.0018 vs. .0005),pointing to a greater cross-national heterogeneity in men’s than inwomen’s responses to the housework question in the SHARE. In contrastto the initial (couple-level) Model 3 and the male model, the positiveeffect of the GEM on the cohabitation slope is marginally significant if therespondent is female. This finding, however, cannot be considered as suf-ficiently robust to argue that gender ideology would be more effective inattaining a more egalitarian division of housework in more gender-equalcountries.

Discussion

Using microdata from the SHARE, this study is the first to investigatethe division of household labor among older couples in a cross-nationalperspective. Across continental Europe, we find considerable variation inthe overall distribution of household labor. One may roughly distinguishbetween more egalitarian countries in northern Europe, such as Sweden and




particularly Denmark, on one hand, and more traditional countries in thesouthern parts of Europe, above all Spain and Greece, on the other hand.Because we are dealing with cohorts born in 1954 or earlier, it is not sur-prising to find a generally lower level of men’s participation in houseworkthan might have been expected from studies that are representative of thewhole population (e.g., Davis & Greenstein, 2004, p. 1265).

The outcome of the multivariate analysis suggests an asymmetric, that is,gendered, effect of the partners’ relative income. If the wife earns less thanher husband, for example, the husband’s share in household duties is signifi-cantly lower than in couples with about equal income, but there is no statis-tically significant effect of relative income if the reverse case is considered.Men’s participation in the labor force decreases males’ participation inhousehold chores, whereas the female partner’s engagement in paid workincreases her husband’s share of household labor. In addition to this supportfor the time-availability hypothesis, we find evidence that less traditionalgender ideologies—indicated by cohabitation and higher education—contribute to a more gender-equal division of housework. However, ourmultilevel analysis reveals no country-specific effects of these couple-levelcharacteristics. The latter finding is different from Fuwa (2004), who pro-vides evidence that relevant microlevel factors have weaker effects on thedivision of household labor for women who live in countries with less pro-nounced gender equality. The lack of support for the macrolevel discount fac-tor argument in our study may result from a common baseline level of genderequality in our sample of nine countries, which might be too high to allow theidentification of effects such as those revealed in Fuwa’s analysis of 22 morediverse nations.

Although our results point to a greater cross-national heterogeneity inmen’s than in women’s responses to the housework question in the SHARE,a general finding is that about half of the between-country variance in thedivision of housework is due to cross-country differences in couples’characteristics. Still, we find a significant effect of macrolevel genderinequalities on couples’ division of housework. Even when controlling forindividual characteristics of the household, couples living in countrieswith higher scores of GEM are more likely to exhibit an equal sharing ofhousehold labor (see Figure 1). Discussing the mechanisms through whichgender empowerment may work, Batalova and Cohen (2002) suggest that“norms about the division of labor may . . . be affected by women’s visi-bility in positions of public authority and prestige” (p. 753). This points to




the role of broader cultural mechanisms in shaping cross-national variationsin the division of household labor—and although the GEM accounts for amajor share of the between-country variance in the distribution of house-work, our analysis still points to the presence of (statistically) significant,unobserved macrolevel heterogeneity. Bianchi et al. (2000) conclude thatmuch of the increase in men’s share of housework observed in younger U.S.cohorts is because of their increased willingness to perform this labor,which is likely to have resulted from

changed attitudes about what is expected, reasonable, and fair for men tocontribute to the maintenance of their home [. . . as well as from . . .] cul-tural change in ideas about ‘women’s work.’ It is likely more acceptable formen to cook and clean, indeed, welcomed, for men to show competence atmaking a home-cooked meal, for example. (p. 219)

Such intertemporal cultural changes are visible as cross-cultural differencesin our investigation.7

This study has some limitations that call for further research. First andforemost, the current SHARE data allow only a cross-sectional view. Thatis, we cannot observe actual changes in housework after retirement. Ourrough cross-sectional evidence as well as previous U.S. research suggeststhat such changes tend to be small. However, the magnitude of thesechanges is likely not only to increase in the future (when new generationsof more highly educated women will enter retirement), but also to varyacross national contexts. Exploiting such intertemporal and intercountryvariations should be a promising field for future research. Second, com-pared to the ISSP, for example, the sample of countries currently repre-sented in the SHARE is relatively small. In particular, former Socialistsocieties are yet missing. Future studies of the division of houseworkamong older couples should not only aim at an extension of the spatial andtime dimensions of their analyses, though. They should, third, also try at themicrolevel to account for complementary productive activities of eldersinside and outside their own homes (see Hook, 2004) and at the macrolevelto include indicators that allow development of a better grasp of the culturalfactors contributing to the persistence of the gendered division of (house-hold) labor.8 Although some suggestions in this latter regard have alreadybeen put forward (such as national cohabitation rates, used by Batalova &Cohen, 2002), much more systematic work needs still to be done.



416

App

endi

xD

eter

min

ants

of

Cou

ples

’Div

isio

n of

Hou

sew

ork:

Res

ults

of

Hie

rarc

hica

l Lin

ear

Mod

els,

Wit

h Se

para

teE

stim

atio

ns f

or M

ale

and

Fem

ale

Res

pond

ents

Mod

el 3

:Mal

eM

odel

3:F

emal

e

βSE

βSE

Inte

rcep

t.2

317**

.005

4.2

031**

.008

3G

EM

.123

6**.0

165

.111

9**.0

251

Mal

e in

com

e <

fem

ale

inco

mea

−.00

82.0

143

−.00

14.0

127

GE

M.0

777

.043

6.0

683

.039

0M

ale

inco

me

> f

emal

e in

com

ea−.

0265

*.0

109

−.02

41**

.010

2G

EM

.021

5.0

311

.005

3.0

288

Mal

e in

labo

r fo

rceb

−.04

57**

.013

8−.

0501

*.0

142

GE

M.0

250

.023

8.0

231

.024

1Fe

mal

e in

labo

r fo

rceb

.059

1**.0

113

.064

4**.0

120

GE

M.0

282

.030

0−.

0021

.031

7U

nmar

ried

cou

ple

.067

8**.0

203

.041

1.0

237

GE

M−.

0155

.069

0−.

0635

.078

6C

oupl

e’s

mea

n ag

e−.

0027

**.0

005

−.00

19**

.000

6M

ale–

fem

ale

age

diff

eren

ce−.

0007

.000

7.0

004

.000

7M

ale,

high

edu

catio

nal d

egre

e.0

138†

.007

7.0

163*

.008

1Fe

mal

e,hi

gh e

duca

tiona

l deg

ree

.013

1.0

088

.019

2*.0

092

Hou

seho

ld in

com

e,1s

t qui

ntile

−.00

94.0

098

−.01

63.0

102

Hou

seho

ld in

com

e,2n

d qu

intil

e.0

065

.009

6.0

121

.010

0H

ouse

hold

inco

me,

4th

quin

tile

.001

5.0

096

.013

7.0

100



417

Hou

seho

ld in

com

e,5t

h qu

intil

e−.

0100

.009

9−.

0091

.010

4M

ale

retir

ed f

rom

labo

r fo

rceb

−.01

17.0

137

−.01

75.0

142

Fem

ale

retir

ed f

rom

labo

r fo

rceb

.026

6**.0

085

.021

7*.0

091

Num

ber

of c

hild

ren

−.00

91**

.002

4−.

0129

**.0

025

Mal

e ca

res

for

gran

dchi

ldre

n.0

104

.010

0.0

145

.010

4Fe

mal

e ca

res

for

gran

dchi

ldre

n−.

0180

†.0

096

−.01

39.0

100

Mal

e lim

ited

by h

ealth

pro

blem

s−.

0057

.006

5.0

054

.006

8Fe

mal

e lim

ited

by h

ealth

pro

blem

s.0

179**

.006

4.0

094

.006

6V

aria

nce

com

pone

nts

Inte

rcep

t.0

018**

.000

5**

Mal

e in

com

e <

fem

ale

inco

me

.000

5.0

001

Mal

e in

com

e >

fem

ale

inco

me

.000

5.0

004

Mal

e in

labo

r fo

rce

.000

2.0

002

Fem

ale

in la

bor

forc

e.0

004

.000

4U

nmar

ried

cou

ple

.001

3.0

024†

Lev

el 1

.033

7.0

369

Not

e:G

EM

= U

nite

d N

atio

ns’g

ende

r em

pow

erm

ent m

easu

re.

a. R

efer

ence

cat

egor

y:m

ale

inco

me

= f

emal

e in

com

e.b.

Ref

eren

ce c

ateg

ory:

mal

es (

fem

ales

,res

pect

ivel

y) w

ho a

re n

eith

er e

mpl

oyed

nor

ret

ired

.† si

gnif

ican

t at 1

0%. * si

gnif

ican

t at 5

%. **

sign

ific

ant a

t 1%

.So

urce

:SH

AR

E 2

004

(Rel

ease

1),

auth

ors’

calc

ulat

ions

.




Notes

1. For details on survey participation, item nonresponse, and imputations in the Survey ofHealth, Ageing and Retirement in Europe (SHARE), see de Luca and Peracchi (2005) andKalwij and van Soest (2005).

2. In 95 percent of our observations, both partners chose the same or a neighboring answercategory. Values indicating that the male partner does the main share or even all of the house-work are observed for less than 2 percent of our sample.

3. Missing values are a particular concern in the case of income variables. In our study weuse imputed information about income provided from the data by the SHARE group. SeeBrugiavini, Croda, Paccagnella, Rainato, and Weber (2005) for a detailed description of theapplied imputation procedures.

4. The proportion of older French couples with an equal division of household labor (asestimated from the SHARE data) is 16 percent, that is, slightly above the average of the ninecountries included in our full analysis.

5. To ease the interpretation of the size effect of the United Nations’ gender empowermentmeasure, we rescaled the values published by the UN so that the country in our sample withthe lowest score has value zero and the country with the highest score has value one.

6. The coefficients of the variables for participation in the labor force and retirement aresignificantly different from each other (both for males and for females). Although this providessome indication for postretirement changes in the division of household labor, these changesare obviously too small to affect the basic distribution of work between men and women.

7. A closely related issue concerns cross-national variations in “equity points.” An unequal(i.e., not 50–50) distribution of household labor need not necessarily be perceived as unfair.However, only since the 1990s has research begun to isolate conditions associated with label-ing divisions of housework as fair or unfair (Coltrane, 2000). Recent work by Davis (2004) notonly reveals cross-national differences in women’s average perceptions of fairness of the divi-sion of household labor, but also shows that these differences are affected by a country’s polit-ical and economic history as well as by women’s overall empowerment.

8. This is not to say that economic factors contributing to greater gender material equalitywould be irrelevant (cf. Breen & Cooke, 2005).

References

Batalova, J. A., & Cohen, P. N. (2002). Premarital cohabitation and housework: Couples incross-national perspective. Journal of Marriage and Family, 64, 743-755.

Baxter, J. (1997). Gender equality and participation in housework: A cross-national perspec-tive. Comparative Family Studies, 28, 220-248.

Becker, G. S. (1981). A treatise on the family. Cambridge, MA: Harvard University Press.Bianchi, S. M., Milkie, M. A., Sayer, L. C., & Robinson, J. P. (2000). Is anyone doing the

housework? Trends in the gender division of household labor. Social Forces, 79, 191-228.Bittman, M., England, P., Sayer, L., Folbre, N., & Matheson, G. (2003). When does gender

trump money? Bargaining and time in household work. American Journal of Sociology,109, 186-214.

Blumberg, R. L. (1984). A general theory of gender stratification. Sociological Theory, 2, 23-101.Börsch-Supan, A., Brugiavini, A., Jürges, H., Mackenbach, J., Siegrist, J., & Weber, G. (Eds.).

(2005). Health, ageing and retirement in Europe—First results from the Survey of Health,




Ageing and Retirement in Europe. Mannheim, Germany: Mannheim Research Institute forthe Economics of Aging. Available at http://www.share-project.org.

Börsch-Supan, A., & Jürges, J. (Eds.). (2005). The Survey of Health, Ageing and Retirementin Europe—Methodology. Mannheim, Germany: Mannheim Research Institute for theEconomics of Ageing. Available at http://www.share-project.org.

Breen, R., & Cooke, L. P. (2005). The persistence of the gendered division of domestic labour.European Sociological Review, 21, 43-57.

Brines, J. (1993). The exchange value of housework. Rationality and Society, 5, 302-340.Brugiavini, A., Croda, E., Paccagnella, O., Rainato, R., & Weber, G. (2005). Generated income

variables in SHARE Release 1. In A. Börsch-Supan & H. Jürges (Eds.), The Survey ofHealth, Ageing and Retirement in Europe—Methodology (pp. 105-113). Mannheim,Germany: Mannheim Research Institute for the Economics of Aging.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and dataanalysis methods. Newbury Park: Sage.

Caro, F. G., & Bass, S. A. (1995). Increasing volunteering among older people. In S. A. Bass(Ed.), Older and active: How Americans over 55 are contributing to society. New Haven:Yale University Press, 71-96.

Chang, M. L. (2000). The evolution of sex segregation regimes. American Journal ofSociology, 105, 1658-1701.

Coltrane, S. (2000). Research on household labor: Modeling and measuring the social embed-dedness of routine family work. Journal of Marriage and Family, 62, 1208-1233.

Coverman, S. (1985). Explaining husbands’ participation in domestic labor. SociologicalQuarterly, 26, 81-97.

Croda, E., & Gonzalez-Chapela, J. (2005). How do European older adults use their time? InA. Börsch-Supan, A. Brugiavini, H. Jürges, J. Mackenbach, J. Siegrist, & G. Weber (Eds.),Health, ageing and retirement in Europe—First results from the Survey of Health, Ageingand Retirement in Europe (pp. 265-271). Mannheim, Germany: Mannheim ResearchInstitute for the Economics of Aging.

Cunningham, M. (2005). Gender in cohabitation and marriage: The influence of gender ideol-ogy on housework allocation over the life course. Journal of Family Issues, 26, 1037-1061.

Davis, S. N. (2004). Is justice contextual? A cross-national analysis of women’s perceptionsof fairness of the division of household labor. Unpublished manuscript, CarolinaPopulation Center, University of North Carolina at Chapel Hill.

Davis, S. N., & Greenstein, T. N. (2004). Cross-national variations in the division of house-hold labor. Journal of Marriage and Family, 66, 1260-1271.

de Luca, G., & Peracchi, F. (2005). Survey participation in the first wave of SHARE. InA. Börsch-Supan & H. Jürges (Eds.), The Survey of Health, Ageing and Retirement inEurope—Methodology (pp. 88-104). Mannheim, Germany: Mannheim Research Institutefor the Economics of Aging.

de Ruijter, E., Treas, J. K., & Cohen, P. N. (2005). Outsourcing the gender factory: Livingarrangements and service expenditures on female and male tasks. Social Forces, 84, 305-322.

Dorfman, L. T. (1992). Couples in retirement. Division of household work. In M. Szinovacz,D. J. Ekerdt, & B. H. Vinick (Eds.), Families and retirement: Conceptual and method-ological issues. Newbury Park, CA: Sage, 159-173.

Elder, L. (2004). Why women don’t run: Explaining women’s underrepresentation inAmerica’s political institutions. Women & Politics, 26, 27-56.

Erlinghagen, M., & Hank, K. (2006). Participation of older Europeans in volunteer work.Ageing & Society, 26, 567-584.




Fuwa, M. (2004). Macro-level gender inequality and the division of household labor in 22countries. American Sociological Review, 69, 751-767.

Gauthier, A., & Smeeding, T. (2003). Time use at older ages: Cross-national differences.Research on Aging, 25, 247-274.

Geist, C. (2005). The welfare state and the home: Regime differences in the domestic divisionof labour. European Sociological Review, 21, 23-41.

Greenstein, T. N. (1996). Husbands’ participation in domestic labor: Interactive effects ofwives’ and husbands’ gender ideologies. Journal of Marriage and Family, 58, 585-595.

Greenstein, T. N. (2000). Economic dependence, gender, and the division of labor in the home:A replication and extension. Journal of Marriage and Family, 62, 322-335.

Hayslip, B., & Kaminski, P. L. (2005). Grandparents raising their grandchildren: A review ofthe literature and suggestions for practice. The Gerontologist, 45, 262-269.

Herzog, A. R., & Morgan, J. N. (1992). Age and gender differences in the value of productiveactivities. Research on Aging, 14, 169-198.

Hook, J. L. (2004). Reconsidering the division of household labor: Incorporating volunteerwork and informal support. Journal of Marriage and Family, 66, 101-117.

Huber, J. (1990). Macro-micro links in gender stratification. American Sociological Review,55, 1-10.

Jacobs, J. A. (1996). Gender inequality in higher education. Annual Review of Sociology, 22,153-185.

Kalwij, A., & van Soest, A. (2005). Item non-response and alternative imputation procedures.In A. Börsch-Supan & H. Jürges (Eds.), The Survey of Health, Ageing and Retirement inEurope—Methodology (pp. 128-150). Mannheim, Germany: Mannheim Research Institutefor the Economics of Aging.

Kamo, Y. (2000). “He said, she said”: Assessing discrepancies in husbands’ and wives’ reportson the division of household labor. Social Science Research, 29, 459-476.

Lee, Y.-S., & Waite, L. J. (2005). Husbands’ and wives’ time spent on housework: A compar-ison of measures. Journal of Marriage and Family, 67, 328-336.

Mason, K. O. (1997). Gender and demographic change: What do we know? In G. W. Jones,R. M. Douglas, J. C Caldwell, & R. M. D’Souza (Eds.), The continuing demographic tran-sition (pp. 158-182). Oxford, UK: Clarendon Press.

Morrow-Howell, N., Hinterlong, J., & Sherraden, M. (Eds.). (2001). Productive aging: A con-ceptual framework. Baltimore: Johns Hopkins University Press.

O’Reilly, P., & Caro, F. G. (1994). Productive aging: An overview of the literature. Journal ofAging and Social Policy, 6, 39-71.

Orloff, A. (1996). Gender in the welfare state. Annual Review of Sociology, 22, 51-78.Pebley, A. R., & Rudkin, L. L. (1999). Grandparents caring for grandchildren: What do we

know? Journal of Family Issues, 20, 218-242.Sanchez, L. (1993). Women’s power and the gendered division of domestic labor in the third

world. Gender & Society, 7, 434-459.Sanchez, L. (1994). Material resources, family structure resources, and husbands’ housework

participation: A cross-national comparison. Journal of Family Issues, 15, 379-402.Sarkisian, N., & Gerstel, N. (2004). Explaining the gender gap in help to parents: The impor-

tance of employment. Journal of Marriage and Family, 66, 431-451.Shelton, B. A., & John, D. (1996). The division of household labor. Annual Review of

Sociology, 22, 299-322.Solomon, C. R., Acock, A. C., & Walker, A. J. (2004). Gender ideology and investment in

housework: Postretirement change. Journal of Family Issues, 25, 1050-1071.




South, S. J., & Spitze, G. (1994). Housework in marital and nonmarital households. AmericanSociological Review, 59, 327-347.

Stier, H., & Lewin-Epstein, N. (2005, May). Policy effects on the division of housework. Paperpresented at the spring meeting of the ISA Research Committee 28, Oslo, Norway.

Stoller, E. P., & Cutler, S. J. (1993). Predictors of use of paid help among older people livingin the community. The Gerontologist, 33, 31-40.

Sundström, M., & Duvander, A.-Z. (2002). Gender division of child care and the sharing ofparental leave among new parents in Sweden. European Sociological Review, 18, 433-447.

Szinovacz, M. E. (2000). Changes in housework after retirement: A panel analysis. Journal ofMarriage and Family, 62, 78-92.

Teachman, J. D., & Crowder, K. D. (2002). Multilevel models in family research: Some con-ceptual and methodological issues. Journal of Marriage and Family, 64, 280-294.

Thompson, L., & Walker, A. J. (1989). Gender in families: Women and men in marriage, work,and parenthood. Journal of Marriage and Family, 51, 845-871.

United Nations Development Program (UNDP). (2004). Human development report. NewYork: UNDP.



Documents

Journal of Family Issues filenational perspective. Recent investigations for the general population strongly suggest that macrolevel factors, particularly gender inequalities,