Marriage entry, divorce and reconciliation: The unintended consequence of the home purchase restriction policy in China

The paper examines how China's home purchase restriction (HPR) policy—introduced to curb speculative housing demand and cool overheated property markets—affects family formation and dissolution. Grounded in Becker’s economic theory of marriage, which frames marriage and divorce as utility-maximizing decisions influenced by socioeconomic factors, the study asks whether a non-family-oriented policy can unintentionally shape marriage entry, divorce, remarriage, and reconciliation. The context is China’s second demographic transition characterized by delayed marriage, higher divorce, and lower fertility. Given homeownership’s central role in Chinese marriage markets, the authors hypothesize two countervailing channels: (1) price-cooling effects of HPR may improve affordability and promote marriages; (2) eligibility rules and purchase quotas may incentivize strategic divorces to expand purchase eligibility and mortgage advantages, potentially delaying or destabilizing marriages. The study aims to quantify these effects overall and by city tier and age group during 2007–2016, informing policymakers about broader social consequences of housing regulations.

The literature links housing markets to marital behavior through wealth, affordability, and transaction costs. Becker (1973, 1974) posits marriage decisions as cost–benefit choices; shifts in costs/benefits alter stability. Empirical evidence is mixed: in the UK, negative house price shocks increased divorce (Rainer & Smith, 2010); in the US, declining prices lowered divorce risk due to high transaction costs (Farnham et al., 2011); Klein (2017) found no significant link to slumps but stronger marital commitments with rising prices; in Iran, higher housing costs destabilized marriages (Farzanegan & Gholipour, 2016). In China, unaffordable housing is associated with reduced initial marriage rates and higher first-marriage age (Wrenn et al., 2019). Zheng et al. (2018) found rising prices increased divorces in the short run in eastern China, with reverse causality in the long run. Policy-focused work is scarce; Alm et al. (2022) linked HPR to strategic divorce propensity using Baidu search behavior rather than administrative divorces. Methodologically, difference-in-differences (DID) is common but vulnerable to selection bias and heterogeneous treatment effects; matching and newer estimators (De Chaisemartin & D’Haultfoeuille, 2020; Sun & Abraham, 2021; Goodman-Bacon, 2021) are recommended. This study adds a policy-oriented, causal perspective using administrative outcomes on marriage and divorce-related events.

Design: City-level staggered difference-in-differences (DID) for 2007–2016 across Chinese cities in tiers 1–3. The study first predicts policy adoption propensity using pre-policy (2010) socioeconomic indicators to construct comparable treatment and control groups, then estimates HPR’s impact on outcomes. Sample: Initial 125 cities (tiers 1–3). After propensity-based selection emphasizing the 2010 affordability index (ratio of housing expenditure to income), 71 cities remain (40 treated HPR cities, 31 controls) for main DID analysis. Policy background: Nationally promoted in 2010, HPR limits purchase eligibility by hukou/status and imposes stricter down payments/mortgage terms (e.g., 30% first home, 50% second; 70% second in tier 1). A loophole allowed divorced ex-spouses to each qualify for a first-home purchase, enabling paper divorces and later reconciliation. Many cities rescinded HPR in 2014 and some reinstated in 2016; some tightened anti-loophole provisions in 2021 (outside study window). Empirical steps: 1) Policy adoption model (Logit): HPR_i = a0 + a1 X_i + ε_i, with 2010 ln(GDP), ln(Population), ln(Fixed asset investment), ln(Revenue), ln(H_Price), and Affordability. Affordability strongly predicts adoption (pseudo-R^2 up to 0.628), guiding construction of comparable controls. 2) Main DID/TWFE: Y_it = α0 + α1 HPR_it + trend_i + μ_i + τ_t + ε_it, where Y_it includes ln house prices, ln marriages (overall, first marriages, by age: 20–24, 25–29, 30–34, 35–39, 40+), ln divorces, ln remarriages, ln reconciliations; μ_i and τ_t are city and year fixed effects; trend_i is a city-specific linear trend. 3) Event study: Y_it = β0 + Σ_k β_k (HPR_it × 1[Year_it = k]) + controls, with Year_it = −1 as baseline, to test parallel trends and dynamic effects. 4) Heterogeneous treatment effects: Assess potential TWFE bias; verify positive weights following De Chaisemartin & D’Haultfoeuille (2020), and estimate HTE-robust effects for robustness. Baseline results use the 71-city sample due to better comparability. Data sources: HPR timing/status from municipal websites and Baidu Baike; marriage/divorce/remarriage/reconciliation counts from China Civil Affairs statistical yearbooks; socioeconomic indicators and average commodity housing prices from city yearbooks; affordability index from Li et al. (2020). Study period chosen to cover 4–5 years pre- and 5–6 years post-HPR rollout and before widespread post-2014 rescission/restoration dynamics fully evolve. City tiers: Tier 1 (Beijing, Shanghai, Guangzhou, Shenzhen), New Tier 1 (15 fast-growing cities per Yicai), Other Tiers (remaining tier 2 and 3).

- HPR reduced house prices: DID estimates show average declines of about 5.8% (full sample, basic TWFE), 7.9% (full sample with city trends), and 9.1% (matched 71-city sample), all statistically significant at 1%. Event study indicates immediate post-HPR price drops and a rebound roughly four years later, coinciding with 2014 rescissions. - Marriage entry overall: No significant average effect on total marriages or first marriages across all cities. - Age-specific marriage entry: Significant 9.9% reduction in marriages among ages 20–24 in HPR cities (5% level). Other age groups show negative but insignificant averages. - Heterogeneity by city tier: - Tier 1: Marriages decreased by 13% for ages 25–29 (5% level), but increased by 18.4% for ages 30–34 (1% level) and 14.5% for ages 35–39 (10% level), suggesting postponement to older ages. - New Tier 1: No significant effects on marriage entry by age. - Other tiers (remaining tier 2–3): Broadly negative coefficients; marriages among ages 20–24 fell by about 11% (5% level). - Strategic divorce proxy (reconciliation): Overall, HPR did not significantly change total divorces or remarriages. However, reconciliations (remarriage to original spouse) rose markedly: - Tier 1: +55.4% (1% level). - New Tier 1: +39.4% (5% level). - Other tiers: Not significant. This pattern is consistent with strategic divorces to bypass purchase/mortgage constraints in high-return markets. - Mechanisms and policy design: The down payment gap for second homes (up to 70% in tier 1 vs 30% for first homes) creates incentives for paper divorces and later reconciliation, undermining demand-suppression goals.

The findings confirm that a non-family-oriented housing policy can meaningfully reshape family formation. HPR succeeded in cooling housing prices, but it also delayed marriages among younger cohorts—especially in first-tier cities—and fostered strategic divorce/reconciliation behavior in stronger markets. These patterns align with Becker’s framework: couples respond to economic incentives, weighing housing affordability, eligibility, and potential wealth gains from additional property. The age and tier heterogeneity indicates that constraints tied to hukou and tenure requirements disproportionately affect young non-hukou residents in expensive cities, shifting marriage timing into the 30s. The spike in reconciliations in higher-tier cities shows that households optimize around policy rules, exploiting loopholes to expand property holdings, which can counteract policy aims by sustaining demand. These results are salient for societies experiencing the second demographic transition, where marriage and fertility declines are policy concerns; housing regulations may inadvertently exacerbate these trends if broader demographic impacts are not anticipated.

The study demonstrates that China’s home purchase restriction policy reduced housing prices but had unintended demographic consequences: delayed marriage among younger cohorts (notably ages 20–24, and 25–29 in tier 1) and increased strategic divorce behavior evidenced by higher reconciliation rates in tier 1 and new tier 1 cities. These outcomes support the economic view of marriage decisions and reveal a policy loophole enabling couples to bypass restrictions to acquire additional properties, potentially undermining HPR’s demand-suppression effectiveness. Policy implications include anticipating behavioral responses and closing loopholes that incentivize paper divorces, while balancing affordability goals with demographic objectives. Future research should employ micro-level household data on property ownership, hukou status, mortgage terms, and family circumstances to unpack mechanisms and heterogeneity within cities and households.

- Data granularity: City-level aggregates limit insight into household-level mechanisms (e.g., specific eligibility constraints, hukou status, and property portfolios). Microdata would enable causal mediation analysis and finer heterogeneity. - Policy endogeneity: Although adoption propensity is modeled and matched, residual selection on unobservables may remain. - Measurement constraints: Administrative reconciliation as a proxy for strategic divorce may not capture all strategic behavior; timing misclassification is possible. - External validity: Results pertain to Chinese city tiers and the 2007–2016 policy regime; effects may differ under later rule changes (e.g., 2021 anti-loophole provisions) or in other institutional settings.