Unveiling the direct and indirect effects of sibling size on happiness: evidence from adults in early and mid-adulthood in China

China has shifted from stringent birth control to pronatalist policies amidst declining fertility and population aging, with growing emphasis on subjective well-being. Despite falling birth rates, recent reports show resilient or rising happiness, prompting investigation into how sibling size relates to happiness. Prior Chinese studies report mixed direct associations (positive, negative, and nonlinear), and often overlook indirect channels via income and education. Moreover, sibling number is endogenous to parental preferences and environments, complicating causal inference. This study asks whether sibling size directly affects happiness, or instead operates indirectly through income and education. It tests three hypotheses: (H1) no direct association between sibling size and happiness after controlling for income and education; (H2) sibling size reduces income, thereby lowering happiness; and (H3) sibling size reduces educational attainment, thereby lowering happiness. Using CGSS 2008 data and an instrumental-variable strategy leveraging the staggered establishment of Family Planning Leading Groups (FPLs) in the 1970s, the paper separates direct from indirect effects.

The literature suggests limited direct links between sibling count and happiness once socioeconomic covariates are controlled; studies of children often find minimal direct effects of sibship size, though birth order can matter. In China, empirical results are mixed: some report positive associations (e.g., sibling support mitigating burdens on only children), others negative (resource concentration benefits only children), and some nonlinear relationships. Income is a well-established determinant of happiness in China, with nonlinearities and sensitivity to inequality. Several studies find that more siblings are associated with lower adult income (resource dilution), though some Chinese evidence suggests nonlinear or context-dependent effects. Education is likewise positively related to happiness and may matter even more than income in some contexts. Competing mechanisms link sibling size to education: the quantity–quality trade-off (more siblings reduce per-child investment, lowering attainment) versus economies of scale and sibling learning. Recent Chinese evidence tends to support the trade-off, especially for women. Based on this, the study posits: H1, no direct effect of sibling size on happiness net of income/education; H2, sibling size lowers income and thus happiness; H3, sibling size lowers education and thus happiness.

Data: 2008 Chinese General Social Survey (CGSS), nationally representative of urban and rural China, with rich individual, sibling (including ages/birth order), parental background, and hukou information. The analytic sample includes individuals born 1969–1987 (restricted to births after 1968 and before 1988) to exploit exogenous variation from early family planning policies and to avoid incomplete schooling among the youngest; final sample sizes vary by model (up to 6102). Key variables: happiness (5-point self-report), income (log), education (years), number of siblings, birth order, gender, health status (1–5), marital status, Communist Party membership, hukou (rural/urban), parental education, parental occupations (at respondent age 14), province fixed effects. Models: (1) Ordered probit for happiness: latent happiness is modeled as a function of income (log), education (years), number of siblings, and controls, with province fixed effects. Estimations with and without income/education identify total versus direct effects of sibling size. (2) Linear models for outcomes Y in {income (log), education (years)} on sibling size and controls using OLS and 2SLS. Instrumental variables: To address endogeneity of sibling size, the study constructs a policy exposure instrument using the difference between an individual's birth year and the year the local Family Planning Leading Group (FPL) was established. Establishment years are compiled from provincial chronicles and the Encyclopedia of Chinese Family Planning, with within-province variation distinguishing provincial capitals from later dissemination to counties and villages. The instrument equals zero if establishment occurs after birth. Exclusion rests on FPLs targeting fertility rather than directly affecting later-life income/education; province fixed effects control for time-invariant provincial factors; additional checks show the instrument has no direct effect in second-stage when included directly. First-stage regressions show the instrument significantly and negatively predicts sibship size at the 1% level. To avoid collinearity between individual birth year and the instrument, some specifications use mother’s birth year/age at birth. Standard errors are clustered, and Hausman tests assess endogeneity where applicable.

- Apparent total association: Without controls, sibling size is negatively correlated with happiness (ordered probit coefficient −0.057, p<0.01). Once income and/or education are included, the sibling coefficient becomes insignificant, indicating no direct effect on happiness. In IV specifications, the sibling coefficient remains insignificant; exogeneity of sibling size is not rejected in that happiness model. - Income channel (Table 3): OLS shows sibling size reduces income: coefficient −0.155 (p<0.01) in a sparse model; −0.058 (p<0.01) with full controls; −0.028 (p<0.05) when controlling for education, implying both direct and education-mediated pathways. The indirect effect via education is about −0.030 (difference between models). 2SLS estimates are imprecise and not significant, but Hausman tests do not reject exogeneity of sibling size in income models, suggesting OLS is efficient and consistent with a negative effect. Education has a strong positive association with income (~0.087–0.090 per schooling year). - Education channel (Table 4): Sibling size robustly reduces educational attainment. OLS: −0.906 (p<0.01) in a sparse model; −0.325 (p<0.01) with extensive controls. 2SLS: −1.981 (p<0.01); exogeneity is strongly rejected (p=0.000), indicating endogeneity bias in OLS and validating the IV strategy. Parental education and stable parental occupations are positively associated with education; rural hukou is associated with substantially fewer schooling years (−2.403 to −3.042 years). - Determinants of happiness: Income (log) and education (years) both positively predict happiness at the 1% level when included; gender (female), better health, and (in some models) party membership are also positively associated with happiness. Overall: Sibling size does not directly reduce happiness; instead, it indirectly lowers happiness by reducing education and income, with strong evidence for the education pathway and supportive evidence for the income pathway.

Findings indicate that changes in sibship size per se do not directly alter subjective well-being among Chinese adults in the studied cohorts. The key mechanisms are indirect: more siblings dilute family resources, lowering educational attainment and income, which in turn depress happiness. The study further shows that sibling size depresses income both through reduced education and via more immediate dilution of non-schooling resources (inheritances, inter vivos transfers, social connections). Policy implications are twofold: (1) pronatalist policies aimed at raising fertility need not conflict with happiness goals if complemented by measures that protect or enhance human capital accumulation and earnings capacity; (2) targeted interventions to equalize opportunities—especially for rural residents and women—such as improving rural schools, enhancing access to higher education, and expanding credit/asset-building supports, could mitigate the negative indirect effects of larger sibships on future well-being.

The study provides causal evidence from China that the number of siblings has no direct effect on adult happiness once income and education are accounted for. Instead, larger sibships reduce educational attainment and income, which indirectly lower happiness. Thus, the happiness impact of fertility operates primarily through socioeconomic channels. For policy, efforts to increase fertility should be paired with investments that sustain or raise educational and income opportunities so that higher fertility does not impair future subjective well-being.

- Generational scope: The analysis focuses on cohorts born 1969–1987; effects may differ for younger generations exposed to later policy changes. Larger and more recent datasets are needed to assess cohort heterogeneity and contemporary applicability. - Policy evaluation: The study does not directly evaluate post-2015 two-child and 2021 three-child policy impacts; quasi-experimental or experimental designs using newer data are warranted. - Omitted pathways: Beyond education and income, other mechanisms influencing happiness (e.g., broader social capital beyond party membership, housing markets, labor conditions, mental health supports) were not fully modeled. Including richer measures of social networks and context could refine pathway estimates.