Living Arrangements of the Elderly in China and Consequences for their Emotional Well-being Qiang Ren, Peking University Donald J. Treiman, UCLA PAA, New.

Living Arrangements of the Elderly in China and Consequences for their

Emotional Well-being

Qiang Ren, Peking UniversityDonald J. Treiman, UCLA

PAA, New Orleans, 11-13 April 2013

2

Introduction• Chinese housing stock has changed rapidly in

recent years, especially in urban areas, with hutong (courtyard) housing replaced by high-rise apartments.

• There also has been massive migration, with more than 200 million people living other than where they are registered, mostly peasants migrating to cities.

• More than 20 million children have been left behind with single parents or grandparents.

• In consequence, the living circumstances of the elderly have changed substantially.

3House in village in Hebei Province (near Beijing)

4Beijing neighborhood

5Upscale high-rise apartments in Beijing

6

AgendaIn this paper we• Describe the current living arrangements of

the elderly.• Try to account for variations in living

arrangements—not presented here because results are very complex.

• Assess the consequences of living arrangements (and covariates) on indicators of emotional well-being.

7

Data• 2010 wave of the Chinese Family Panel

Survey, a multistage (nearly) national probability sample of Chinese families.

• 14,960 households were included.• All family members in each household were

interviewed (those under age 10 by proxy), resulting in 42,590 responses.

• Of these, 7,040 were age 60 or older. This is the sample we study.

8

Living arrangements

9

Determinants of living arrangements

• Not presented here.

10

Consequences of living arrangements

• Much popular speculation and some research on the consequences of living arrangements among the elderly in China.– Claim that living with children, and especially

grandchildren, enhances emotional health of the elderly—the dominant position.

– Counterclaim that living with children creates conflict and living with grandchildren imposes a burden, depressing emotional health of the elderly.

11

Empirical tests (1)

• Three outcome variables:– Happiness (5-pt scale ranging from “very

unhappy” to “very happy”). [mean: 3.8; s.d.:1.0]– Life-satisfaction (5-pt scale ranging from “very

unsatisfied” to “very satisfied”). [mean: 3.6; s.d.: 1.0]

– Depression (average score on 6-variable version of CES-D, with scores ranging from 1 “never” to 5 “almost every day”). [mean: 1.6; s.d.: 0.7]

12

Empirical tests (2)• Main predictor variables: 10-category living

arrangements typology shown earlier (living with spouse only is the reference category).

• Covariates: age, gender, years of schooling, vocabulary knowledge, one or more children still alive, health status, ln(per capita family income), and whether housing is inadequate.

• Analytic strategy: seemingly unrelated regression, (a) on typology; (b) adding covariates. Here we present only (b).

• First entire sample, and then separately for current rural and urban residents.

13

Consequences: results (1)• Is living with adult children good for the

emotional health of the elderly?– No. Compared to elderly couples living independ-

ently, couples living with an adult parent but no grandchildren tend to be less happy, less satisfied with life, and more depressed. (All coefficients are significant at <.01.). Effect sizes range from |.13| to |.23|.

– Living without a spouse but with grown children is even worse than living with both a spouse and grown children. Effect sizes range from |.27| to |.39|.

14

Consequences: results (2)• Do grandchildren reduce the negative effects

of living with adult children?– Yes. The negative effects on happiness and life

satisfaction become non-significant (p≥.10) and the effects on depression become substantially weaker (from .15 to .06 when the spouse is a household member and from .28 to .10 when the spouse is not a household member).

15

Consequences: results (3)• Does the presence of grandchildren but not

children (“generation-skipping households”) enhance emotional well-being compared to spouse-only elderly couples?– No. The elderly are significantly less happy, less

satisfied, and more depressed when grand-children are present. (All coefficients are significant at <.10.) Effect sizes range from |.07| to |.13|.

– Things are even worse when no spouse is present (p always <.07). Effect sizes increase to |.19| to |.30|.

16

Consequences: results (4)• Is living alone bad for emotional health?– Yes. The coefficients are all significant at < .001

and the effect sizes range from |.18| to |.32|• Which covariates matter? Consistent effects,

all in the expected direction, include:– Vocabulary knowledge, net of education, which

probably indicates continued cognitive capacity.– Having a least one child still alive.– Good health.– Income.– Urban residence and adequate housing matter for

happiness and depression but not for life satisfaction.

17

Consequences: results (5)• Do urban and rural residents differ? – Yes, but in somewhat complex ways. Here are

highlights of rural-urban differences:– 2-gen. family with spouse: no effect on happiness

(relative to spouse-only households) for rural elderly but negative effect for urban elderly.

– 3-gen. family with spouse: positive effect on happiness and also greater depression for rural elderly (inconsistent); no effect for urban elderly.

– 3-gen. family without spouse: same as above for rural elderly; greater depression for urban elderly.

18

Consequences: results (6)– Generation-skipping family with spouse: negative

effect on life satisfaction and greater depression for rural elderly; no effects for urban elderly.

– Generation-skipping family with no spouse: negative effect on life satisfaction for rural elderly; no effects for urban elderly.

– Urban males less happy, satisfied than females; no gender effect for rural elderly. (Both urban and rural males less depressed than females.)

– Children still alive: reduces depression for rural elderly; no effect in urban areas.

19

Consequences: results (7)• Seemingly unrelated regression results.

Residuals are correlated in the expected way:

20

Summary• From the point of view of emotional well-being,

living independently with one’s spouse is the optimal living arrangement for the elderly in contemporary China.

• Living with adult children, in 3-generation households, and in generation-skipping households, generally degrades emotional health for the elderly, particularly when household not shared with a spouse.

• There are some rural-urban differences, but they do not permit a coherent story.

21

Concluding conjectures• Despite claims about traditional values

emphasizing care for the elderly, it is evident that extended family arrangements do not enhance the emotional well-being of the elderly and on the whole degrade it. It is likely that this is due to the stress that multiple-generation relationships bring, in China as elsewhere. Stories of intergenerational conflict are the stuff of literary accounts extending over many centuries. The increasing urbanization and extensive internal migration in contemporary China may exacerbate such stresses.

22

Thank you