partner-search strategy, the red line on pestering, signs of interest, relationship quality, long-distance relationships, the registration process, the premarital health check, the health accounting, the time accounting, the money accounting, parents paying toward a home and joint marital debt, exit costs, going to couples therapy together when the relationship is stuck. For the long-form piece, see Is marriage worth it?.
A 14B 5C 1
Lifespan 3Time and energy 9Money 6Personal freedom 23 contested
This section breaks “is marriage worth it” into several separate accounts: money, time, health, legal risk and relationship quality, and works each one out on its own. Each account uses only numbers whose sources can be traced, and the sources are limited to official statistics or meta-analyses (which pool the data from multiple studies and recalculate it all together). No moral judgments are made here, and no conclusion is drawn for you. The health accounts are measured in risk of death, the time accounts in minutes per day, and the legal accounts by what the statutes say. These three measures are not converted into one another. For how to break down the accounts, a fill-in checklist and common misconceptions, see the long-form piece “Is Marriage Worth It?”.
The items in this section are grouped by topic below; the numbers in parentheses are item numbers.
Finding a partner and dating: meet more people (Item 1), stop once the other person says no (Item 2), judge interest by behavior (Item 3), take turns answering questions that go from light to deep (Item 5), long distance is no worse than the same city (Item 6).
Whether the relationship is good: look at how you yourself feel in the relationship (Item 4), argue about the issue, not the person (Item 17), go to couple therapy together (Item 20).
The accounts of whether marriage is worth it: look at the registration data first (Item 7), discount the health benefits as observational data (Item 8), count the time account as unpaid labor (Item 9), work out the exit costs (Item 15), keep a separate account for marrying for your elders (Item 16).
Registering the marriage and the premarital checkup: registration needs only your ID card (Item 13), get a premarital checkup once before marrying (Item 14).
Money and debt: for the money account, look at the legal default rules first (Item 10), when parents pay toward a home, put in writing whether it is a loan or a gift (Item 11), large sums borrowed by one spouse (Item 12).
Same-sex partners and sexual orientation: same-sex partners should set up voluntary guardianship and wills (Item 18), don't marry by hiding your sexual orientation (Item 19).
Sources are reproduced exactly as they appear in the Chinese original, including Chinese titles of laws and quoted statutory text. Original Chinese text of this section at upstream commit bb25081, on GitHub.
Meet more people instead of fixating on one: whether two people click can't be predicted before they meet
Value for cost Standard
In plain terms
Two speed-dating studies had people fill in more than 100 mate-preference questionnaire items before meeting. The questionnaires could only work out who tends to like others in general and who is popular in general; they could not work out whether these two particular people would spark. Whether two people click is known only once they meet. So getting to know a few more people is a better deal than betting on one person.
Cost
Free. Spend your time on getting to know a few more people, not on going over the same person again and again. The hard part is holding back from fixating.
No moneyA few hoursSome willpowerBenefit size medium
Benefit
In two speed-dating studies, participants first filled in more than 100 mate-related self-report scales (questionnaires where you rate yourself) before meeting. The researchers ran the numbers with a random forest model (an algorithm that finds patterns). It could work out “how readily a person likes others overall”: these questionnaires explain 4%–18% of the variance in that, where variance is the differences between people, and the questionnaires explain only a small part of it. It could also work out “how popular a person is overall”, explaining 7%–27%. But “whether these two specific people spark” could not be worked out: no combination of pre-meeting traits and preferences could predict it. The same researchers had earlier run another speed-dating experiment. Participants reported their ideal-partner preferences beforehand, for example whether they valued looks or financial prospects more. These preferences did not predict their interest in a specific person at the event.
Evidence grade
A
Notes
The two studies show only that “it can't be predicted before meeting”, not that “no amount of effort helps”. Whether two people click can only be known after they meet, so spending your time on getting to know a few more people, and actually meeting each one, is a better deal than betting everything on one person. The participants were all college students in Europe and North America. China's dating and marriage environment is different, so discount the findings when applying them to yourself.
Sources
Joel S, Eastwick PW, Finkel EJ (2017). Is romantic desire predictable? Machine learning applied to initial romantic attraction. Psychological Science, 28(10), 1478-1489. https://doi.org/10.1177/0956797617714580；Eastwick PW, Finkel EJ (2008). Sex differences in mate preferences revisited: do people know what they initially desire in a romantic partner? Journal of Personality and Social Psychology, 94(2), 245-264. https://doi.org/10.1037/0022-3514.94.2.245
Stop once the other person has clearly said no: pestering them further is a public-security offense, not “sincerity”
Value for cost Standard
In plain terms
If the other person has clearly said no and you keep messaging them or going to wait for them, from January 2026 that is a public-security offense. You can be detained for up to 5 days or fined up to 1000 yuan. In more serious cases, detention is 5 to 10 days. The police can also order you not to contact the other person, and breaking that order means another 5 to 10 days of detention.
Cost
Free, and it takes no time. The hard part is holding back: no more messages, no more going to wait for them.
No moneyDone in passingLots of willpowerBenefit size medium
Benefit
The Public Security Administration Punishments Law (治安管理处罚法), in force from January 1, 2026, lists pestering and stalking as a separate offense. The text reads: “repeatedly sending obscene, insulting, threatening or other such messages, or using harassment, pestering, stalking or other such methods, to disturb another person's normal life”. This conduct is punished with detention of not more than five days or a fine of not more than one thousand yuan. In more serious cases, it is punished with detention of not less than five days and not more than ten days, and a fine of not more than one thousand yuan may be imposed as well. This provision adds one more layer: with the approval of the head of the public security authorities (police), the offender may be ordered not to contact the victim for a set period. Violating the no-contact order is punished with a further detention of not less than five days and not more than ten days.
Evidence grade
A
Notes
No research at all supports the idea that “pursuing someone to the bitter end will win them over”, but detention and no-contact orders are written in black and white. The record left behind will also affect civil service exams, exams for public-institution posts, political background checks, and the professional qualifications for some occupations. The line between pursuing someone and harassing them does not depend on what you think; it depends on whether the other person has clearly said no.
Judge whether someone is interested by behavior, not by “signals”: who initiates, who makes plans, who follows up
ContestedValue for cost Standard
In plain terms
Researchers had 52 pairs of strangers, a man and a woman each, chat for ten minutes or so and then judge whether the other was flirting. The result: “they're not flirting” was judged fairly accurately, while “they're flirting”, by contrast, was not picked up. When 261 people judged from videos instead, the conclusion was the same. So don't guess from expressions and tone; look at behavior: who initiates, who arranges the next meeting, and whether the other person follows up when you don't push.
Cost
Free, and it takes little time. The hard part is changing a habit: stop puzzling over the other person's expressions and tone, and record the facts instead.
No moneyDone in passingSome willpowerBenefit size medium
Benefit
Two studies measured how accurately people detect flirting. The first had 52 pairs of opposite-sex strangers talk one-on-one for 10 to 12 minutes. Afterward each reported whether they themselves had been flirting, and whether they thought the other had been flirting. The result was that “interactions without flirting were judged more accurately, while interactions with flirting were judged worse”. The second showed 26 one-minute videos to 261 observers to judge, with the same conclusion. People can fairly reliably tell that “the other person isn't flirting”, but find it hard to tell that “the other person is flirting”.
Evidence grade
B
Notes
Contested. Other research argues that it is not that men overestimate women's sexual intent, but that women report their own intent lower on questionnaires. Which side judges more accurately is still unsettled. But both camps agree on one thing: inferring what someone is thinking from subtle cues such as expressions and tone is very inaccurate. Visible behavior is far more reliable: whether the other person initiates, whether they arrange the next meeting, and whether they follow up even when you don't push.
Sources
Hall JA, Xing C, Brooks S (2015). Accurately detecting flirting: error management theory, the traditional sexual script, and flirting base rate. Communication Research, 42(7), 939-958. https://doi.org/10.1177/0093650214534972；Perilloux C, Kurzban R (2015). Do men overperceive women's sexual interest? Psychological Science, 26(1), 70-77. https://doi.org/10.1177/0956797614555727（争议方）
Whether a relationship is good depends mainly on how you yourself experience it, not on the other person's attributes on paper
Value for cost Standard
In plain terms
What best shows whether a relationship is good is how you yourself feel, not the other person's attributes on paper. For example: whether you feel the other person is committed and appreciates you, and how much the two of you fight. Your own perceptions are two to four times as useful as your partner's reports. But whether this relationship gets better or worse from here, essentially no questionnaire can work out.
Cost
Free, and it takes little time. The hard part is changing perspective: put away the partner checklist and look at how the two of you get along instead.
No moneyDone in passingSome willpowerBenefit size medium
Benefit
A machine-learning analysis of 43 longitudinal couples datasets from 29 labs; longitudinal means following the same couples and recording along the way. The strongest predictors of relationship quality all came from the person's own perceptions of the relationship: perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, and conflict. Among the items about the individual, the strongest were life satisfaction, negative affect, depression, avoidant attachment and anxious attachment. These relationship items explained at most 45% of the variance at the start of the surveys. Variance here means the differences in relationship quality from couple to couple, and 45% is the part of it these items can account for. By the end of the studies they explained at most 18%. Moreover, the predictive power of “the items you report yourself” was two to four times that of “the items your partner reports”. Beyond your own reported perceptions of the relationship, adding individual differences and the partner's reports added no predictive power at all.
Evidence grade
A
Notes
The same study also found that whether the relationship got better or worse over the study period could essentially not be predicted by any set of self-report measures in the questionnaires. So to judge whether your current relationship is good, asking yourself how you feel is more accurate than checking the other person against a checklist of attributes, but don't expect it to predict how things will go later. For the account linking relationship quality and physical health, see Item 17 (treat relationship quality as a health account).
Sources
Joel S, Eastwick PW, Allison CJ, et al. (2020). Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies. Proceedings of the National Academy of Sciences, 117(32), 19061-19071. https://doi.org/10.1073/pnas.1917036117
To get closer, take turns answering a set of questions that go from light to deep; don't learn scripted lines
Value for cost Standard
In plain terms
Two strangers took turns answering 36 questions that go from light to deep, and after 45 minutes they felt much closer than people who made small talk for the same length of time. What works is taking turns telling the truth and going deeper layer by layer, not scripted lines. This shows only that they felt closer at the time, not that it can grow into a long-term relationship.
Cost
Free. It takes 45 minutes.
No moneyDone in passingNo willpowerBenefit size small
Benefit
Pairs of strangers took turns answering three sets of progressively deeper questions, 36 in all. The questions range from “whom would you most like to treat to a meal” to “when did you last cry in front of another person”. Closeness after 45 minutes was significantly higher than in a control group that made small talk for the same length of time. It was also higher than the average rating that college students in the study gave their closest relationship in real life.
Evidence grade
B
Notes
The experiment has been run only once, with college students, and it measured only closeness over a short time, not whether that could grow into a long-term relationship. The closeness in the experiment came from the two people taking turns, layer by layer, saying true things about themselves, not from scripted lines, tricks or a one-sided performance. So in conversation you can do “I say something true, then ask something just as deep”, with both sides taking turns.
Sources
Aron A, Melinat E, Aron EN, Vallone RD, Bator RJ (1997). The experimental generation of interpersonal closeness: a procedure and some preliminary findings. Personality and Social Psychology Bulletin, 23(4), 363-377. https://doi.org/10.1177/0146167297234003
A long-distance relationship is no worse than a same-city one; the real hurdle is the first three months after you move to be together
Value for cost Standard
In plain terms
Long-distance relationships are actually more stable while the two are apart, because seeing each other less makes it easy to imagine the other person as better than they really are. The real hurdle comes after moving to be together. In the tracking data, about half of long-distance couples got as far as being in the same city, and the other half ended while still apart. Of the half that reunited, one-third broke up within 3 months.
Cost
It costs travel fares, plus the time spent traveling back and forth. You see each other less often than in the same city. The hard part is holding up through long stretches without seeing each other.
Real moneyTime every dayLots of willpowerBenefit size medium
Benefit
Two studies reach the same conclusion. Long-distance relationships are actually more stable than same-city ones while the partners are apart. The reason is that idealizing the other person as better than they really are is markedly stronger in long-distance relationships; the studies measured idealized distortion, romantic love, relationship memories and perceived agreement. And the less often they meet, the rosier the picture. But this stability holds only while the two are still apart; once they are in the same city, long-distance couples are more likely to break up. The study that tracked the move from long-distance to same-city gives these proportions: about half of long-distance couples got as far as moving to be together, and the other half ended while still apart. Of the half that reunited, one-third broke up within 3 months of reuniting. The longer the gaps between visits while apart, and the stronger the idealization, the less stable the relationship after reuniting.
Evidence grade
B
Notes
The subjects were dating relationships of American college students, not marriages, and the studies are fairly old, so don't apply the absolute proportions directly. Another study surveyed 717 people in long-distance relationships and 425 in same-city ones. The difference in relationship quality between the two kinds was very small; whether a relationship is good depends mainly on each person's own circumstances and on the relationship itself. So when deciding whether to go long distance, don't use “long distance always falls apart” as a reason. What to prepare for in advance is the few months after reuniting, when you have less personal space, learn more about each other, and have more conflict and jealousy.
Sources
Stafford L, Merolla AJ, Castle JD (2006). When long-distance dating partners become geographically close. Journal of Social and Personal Relationships, 23(6), 901-919. https://doi.org/10.1177/0265407506070472；Stafford L, Merolla AJ (2007). Idealization, reunions, and stability in long-distance dating relationships. Journal of Social and Personal Relationships, 24(1), 37-54. https://doi.org/10.1177/0265407507072578；Dargie E 等 (2015). Journal of Sex & Marital Therapy, 41(2), 181-202. https://doi.org/10.1080/0092623X.2013.864367（备注里那项异地研究）
Look at the registration data before listening to your elders: what the marriage rate and the divorce rate each actually measure
Value for cost Standard
In plain terms
In 2024, 6,106,000 couples registered marriages nationwide, roughly two-tenths fewer than the year before, and 3,513,000 couples divorced. The claim online that the divorce rate is over half comes from dividing one of these two numbers by the other: about 57.5%. But the people divorcing and the people marrying are not the same group, so it cannot be read as half of marriages ending in divorce. The bulletin's marriage rate of 4.3‰ and divorce rate of 2.5‰ are calculated by dividing by the national population.
Cost
Free. Read the marriage registration passage of the Ministry of Civil Affairs annual bulletin once; it takes a few minutes.
No moneyDone in passingNo willpowerBenefit size small
Benefit
In 2024, 6,106,000 couples completed marriage registration according to law nationwide, down 20.5% from the previous year; the marriage rate was 4.3‰. In the same year, 3,513,000 couples completed divorce procedures according to law, of which 2,622,000 couples registered their divorce with civil affairs departments and 891,000 couples divorced by court judgment or mediation; the divorce rate was 2.5‰. The bulletin notes that marriage (divorce) rate = number of couples marrying (divorcing) in the year / average total population in the year × 1000‰. This is a “crude rate”: the denominator is the whole population, not the married population.
Evidence grade
A
Notes
The “divorce rate of 50% or more” often seen online divides the number of couples divorcing in a year by the number of couples marrying that year: 351.3 / 610.6 ≈ 57.5%. This is called the “divorce-to-marriage ratio”. The two numbers being divided are not the same group of people, and the bulletin contains no such indicator, so it cannot be read as “half of marriages end in divorce”. To work out “the probability that a marriage ends in divorce in the end”, you would have to follow the same group of people who married and keep counting all the way through; this section found no official data of that kind.
Count the health benefits of marriage, but discount them as observational data
ContestedValue for cost Standard
In plain terms
Married people's probability of dying over the same period is about 12% lower than that of people who are not married, and people who have never married have a probability about 24% higher than married people. But this only means the two things occur together, not that marriage has the effect. People who are healthy, earn more and socialize more to begin with are both more likely to marry and more likely to live long. These numbers come basically from Europe and North America and serve only as an order-of-magnitude reference for China.
Cost
Free, and it takes no time. Just remember one point: these numbers say that death rates are lower among married people; the two things occur together, it does not mean marriage makes you live longer.
No moneyDone in passingNo willpowerBenefit size small
Benefit
A meta-analysis of older adults, pooling multiple studies and recalculating, covered 53 comparisons and more than 250,000 people. For married versus not married, the RR for all-cause death was 0.88, 95% CI 0.85–0.91 (RR is the risk ratio, CI is the confidence interval): the probability of dying over the same period is about 12% lower. In the same analysis, widowed was 1.11, divorced or separated 1.16 and never married 1.11, that is, about 11%, 16% and 11% higher respectively. There was no difference between men and women. The most conservative estimate was RR 0.94, only about 6% lower. Another meta-analysis compared the never-married (people who have never married) with the married, using 95 publications and more than 500 million people. The result was HR 1.24 (1.19–1.30): the probability of dying over the same period is about 24% higher. And the older people are, the smaller the gap. A further meta-analysis broken down by sex covered 21 studies and more than 7,890,000 people: for both men and women, not being married occurs together with all-cause death, cancer death and cardiovascular death. On the men's side, the gaps in cardiovascular death and all-cause death were larger. Never-married men, compared with never-married women, had 31% more stroke deaths and 9% more all-cause deaths.
Evidence grade
A (contested)
Notes
Contested. The issue is that people who marry have already been selected: people who are healthy, earn more and socialize more are both more likely to marry and more likely to live long. All three meta-analyses use observational data from existing populations, only followed and recorded, with no assignment to comparison groups. The authors themselves also wrote that there was publication bias or methodological limitations. Roelfs also found that the more rigorous the study, the smaller the calculated gap. These numbers come mainly from European and North American populations and serve only as an order-of-magnitude reference for China.
Sources
Manzoli L, et al. (2007). Marital status and mortality in the elderly: a systematic review and meta-analysis. Soc Sci Med 64:77–94. https://doi.org/10.1016/j.socscimed.2006.08.031；Roelfs DJ, et al. (2011). The rising relative risk of mortality for singles: meta-analysis and meta-regression. Am J Epidemiol 174(4):379–389. https://doi.org/10.1093/aje/kwr111；Wang Y, et al. (2020). Sex differences in the association between marital status and the risk of cardiovascular, cancer, and all-cause mortality: a systematic review and meta-analysis of 7,881,040 individuals. Glob Health Res Policy 5:4. https://doi.org/10.1186/s41256-020-00133-8
Count the time account as “unpaid labor”, and agree on the division of work before registering the marriage
Value for cost High
In plain terms
In the national time-use survey, men spend an average of 1 hour 52 minutes a day on unpaid work such as housework and accompanying and caring for family members, and women 3 hours 29 minutes. The survey did not split people by whether they were married, so there is no official figure for “how much more housework comes with marriage”, and this gender gap can serve only as an upper-bound reference. The division of work can be negotiated into something different before you register.
Cost
Free. Before marrying, spend an hour or two turning the unpaid work (housework, caring for the old and the young, whose job gives way) into a clear division of labor, then follow that division afterward. The hard part is laying it all out and talking, instead of putting it off.
No moneyA few hoursSome willpowerBenefit size large
Benefit
The 2024 national time-use survey. Unpaid labor is housework, accompanying and caring for family members, and other such work that is not paid. People who did this kind of work that day spent an average of 2 hours 45 minutes a day on it: men 1 hour 52 minutes, women 3 hours 29 minutes. The share of people who did this kind of work at some point in the day was 67.5% for men and 83.9% for women. Residents' housework averaged 1 hour 17 minutes a day, or 1 hour 59 minutes counting only those who did it, 28 minutes less than in 2018. Accompanying and caring for family members averaged 30 minutes a day, or 1 hour 46 minutes counting only those who did it. The 2018 figures were: housework, men 45 minutes and women 2 hours 6 minutes; accompanying and caring for children, men 17 minutes and women 53 minutes.
Evidence grade
A
Notes
Both surveys split people only by sex, urban or rural, and age, not by whether they were married. So there is no official figure for “how much more housework comes after marriage”, and the gender gap can serve only as an upper-bound reference. The numbers are national averages; the two of you can absolutely negotiate a different division of work.
For the money account of marriage, look at the legal default rules first, then decide whether to make a written agreement
Value for cost High
In plain terms
Without a written agreement, the law's default is that wages and bonuses earned after marriage, business and investment income, and property inherited or received as a gift all belong to the two of you jointly. Property from before the marriage, and property that a will or gift contract states is for one spouse only, are that spouse's own. The two of you can agree in writing that each keeps their own, and once signed the agreement is binding. The spouse who gave more because of raising children or caring for older relatives can ask for compensation at divorce.
Cost
Free. Read through the articles on property in the Marriage and Family Book of the Civil Code (民法典婚姻家庭编). If needed, the two of you sign a written property agreement; the law only requires it to be in writing, and it does not have to be notarized. The hard part is talking openly about money in advance.
No moneyA few hoursSome willpowerBenefit size large
Benefit
The default rules when no other agreement is made: wages, bonuses and remuneration for services, income from production, business and investment, income from intellectual property, and property inherited or received as a gift during the marriage belong to husband and wife jointly (Article 1062). Pre-marital property, compensation for personal injury, property that a will or gift contract specifies as belonging to one spouse only, and articles for one spouse's exclusive daily use are personal property (Article 1063). The two parties may agree in writing that property acquired before and during the marriage belongs to each separately, to both jointly, or partly separately and partly jointly, and this agreement is legally binding on both of them (Article 1065). The spouse who bears more obligations because of raising children, caring for older relatives or assisting the other spouse's work is entitled to request compensation at divorce (Article 1088).
Evidence grade
A
Notes
There are no official statistics on betrothal gifts (caili), weddings or marital homes, so this item gives no numbers. How much compensation is paid under Article 1088 is first negotiated by the two of you, and if you can't agree, the court decides; there is no fixed formula. A written property agreement works for both sides: the higher earner can use it, and so can the spouse who does more of the unpaid work such as housework and caregiving.
When parents pay toward buying a home, put in writing at the time of the transfer whether it is a loan or a gift
Value for cost Very high
In plain terms
When parents pay for a home before the marriage, by default it is a gift to their own child only. Money paid after the marriage with no agreement is treated as the couple's joint property. From February 2025, if after marriage one side's parents pay the full amount and do not state in writing that it is for their own child only, at divorce the court can award the home to the paying parents' child, and then decide whether to compensate the other spouse. If it is to count as a loan, write it down clearly on the day, with both sides signing.
Cost
Free. On the day of the payment, write a one-page statement saying whom the money is a gift to, or whom it is lent to, signed by both the paying parents and their child. If it says loan, have both spouses sign, and put the same sentence in the transfer note.
No moneyDone in passingNo willpowerBenefit size large
Benefit
When parents contribute toward buying a home for the couple before the marriage, it is deemed a gift to their own child personally, unless the parents clearly state that it is a gift to both. For contributions after marriage, any agreement is followed. Where there is no agreement or the agreement is unclear, it is handled under the principle of the couple's joint property. From February 1, 2025, there is one more rule. Where one side's parents paid the full amount during the marriage and the gift contract does not state that it is a gift to their own child only, the court may, when dividing property at divorce, award the home to the paying parents' child. After awarding it, the court weighs several factors together to decide whether to compensate the other spouse, and how much. These factors are the circumstances of the couple's life together and of having children together, fault in the divorce, the size of each person's contribution to the family, and the market price of the home at the time of divorce. Where one side's parents paid part of the cost, or both sides' parents contributed, the home is awarded to one of the spouses on the basis of the source and proportion of the contributions, with reasonable compensation (nationwide).
Evidence grade
A
Notes
Writing it down clearly helps both sides. The parents who pay no longer have to count on drawing up an IOU after the fact at divorce to protect their contribution. Since Interpretation (II) (解释（二）) took effect, who paid is itself the basis for dividing the home, and whose name is on the property certificate no longer decides who gets it. The spouse receiving the contribution also doesn't have to wait until divorce to find out what the money counts as. If it is to be a loan, treat it as a real loan: both sides agree that it is a loan, agree how it will be repaid, say so in the transfer, and keep repayment records if needed. If any one of these is missing, you will have to explain it in court.
If one spouse borrows a large sum and you neither signed nor ratified it afterward, it does not automatically become your debt
Value for cost High
In plain terms
Money your spouse borrowed behind your back does not automatically become your debt, as long as it clearly exceeds the household's everyday spending and you neither signed nor acknowledged it afterward. The lender has to prove for themselves that the money went to your shared life or shared business. But if money such as home-purchase money really did become a home the two of you live in or own jointly, the lender can often prove that much.
Cost
Free. Don't sign your spouse's IOUs or guarantee documents as “co-borrower” or “guarantor”. Also don't write things like “we'll pay this back” in chats; it may be taken as you acknowledging the debt after the fact. The hard part is not engaging when you are being pressed for payment.
No moneyDone in passingSome willpowerBenefit size large
Benefit
Money borrowed in the following situations is joint marital debt. First, when both spouses signed together. Second, when one spouse ratified it afterward, that is, acknowledged the debt after the fact, so that both acknowledge it. Third, when one spouse borrowed in their own name during the marriage for the family's everyday living needs. Money one spouse borrowed in their own name that clearly exceeds the household's everyday spending is not joint marital debt. The exception is when the creditor (the side owed the money) can prove that the money was used for the couple's shared life or shared production and business, or that it was a debt both acknowledged. Home-purchase money clearly exceeds the family's everyday living needs, and the burden of proof lies with the party claiming the debt: the lender has to produce the evidence. But if the money really did become a home the couple live in together or own jointly, the creditor can often prove the point that it was “used for the couple's shared life”. So the more effective step is the one in Item 11 (put in writing whether parents' money is a loan or a gift) (nationwide, from 2021).
Evidence grade
A
Notes
If, when property is being divided at divorce, your spouse's parents suddenly produce a loan agreement signed only by your spouse, look at it in two steps. Step one: whether the loan actually happened. If you say the money was never handed over and can give a coherent reason, the court will judge by considering several things together. First the loan amount, how the money was handed over and the two sides' financial means, then their usual dealings, whether property changed correspondingly, and what witnesses say. If the other side can produce only a transfer record, it is still up to them to prove that the loan was made. Step two: once the loan is established, look at whether it is joint debt. For the general rules on IOUs and guarantees, see Section 8, Item 18.
Registering a marriage needs only your ID card plus a signed declaration: no household registration booklet, no fee, certificate issued on the spot
Value for cost Standard
In plain terms
From May 2025, registering a marriage needs only two things: your own ID card plus a signed declaration. The declaration states that you have no spouse, and no lineal blood relationship or collateral blood relationship within three generations with the other person. You don't need to bring the household registration booklet, and you don't need to go back to the place of either person's household registration (hukou) to do it. Registration costs nothing at all, and if you meet the conditions the certificate is issued on the spot.
Cost
Free. Both people must be there in person. It usually takes a few dozen minutes.
No moneyDone in passingNo willpowerBenefit size small
Benefit
The new Regulations on Marriage Registration (婚姻登记条例), in force from May 10, 2025, changed the documents mainland residents must submit to marry. The text says only two things need to be presented: “(i) the person's own resident ID card; (ii) a signed declaration that the person has no spouse and has no lineal blood relationship or collateral blood relationship within three generations with the other party”. The household registration booklet is no longer a required document. The registration authority is “the civil affairs department of a county-level people's government, or a township (town) people's government designated by the people's government of a province, autonomous region or municipality directly under the central government on the principle of convenience for the public”. The old regulations' restriction of going “to the place of either party's permanent household registration” has been removed. The regulations also state that “marriage registration authorities shall not charge fees for handling marriage registration”, and that “where the parties meet the conditions for marriage, registration shall be completed on the spot and a marriage certificate issued”.
Evidence grade
A
Notes
The registration authority refuses only in four situations: someone has not reached the legal marriage age, it is not fully voluntary on both sides, one or both already have a spouse, or the two are lineal blood relatives or collateral blood relatives within three generations. The only things that may cost money are side services such as photos and photocopies. Registration itself costs nothing at all; if you are asked for a “production fee” or a “rush fee”, refuse outright and file a complaint. For the 30-day cooling-off period for divorce registration, see the item on exit costs in this section.
Get a premarital checkup once before marrying; a serious illness must be truthfully disclosed to the other person before registration
Value for cost Standard
In plain terms
If you have a serious illness and don't truthfully tell the other person before registration, they can ask a court to revoke the marriage. They must do so within one year of the day they find out. The premarital checkup itself has not been a precondition for registering a marriage since 2003; going is voluntary. The cost depends on local policy; it takes half a day.
Cost
Most regions make it free or reduced-cost for people of the eligible age. Even paying yourself, it is only a few hundred yuan. It takes half a day.
A little moneyDone in passingNo willpowerBenefit size medium
Benefit
The legal consequences of concealment are clear. The new Regulations on Marriage Registration (婚姻登记条例) provide that “where one party has a serious illness, they shall truthfully inform the other party before marriage registration; if they do not, the other party may request a people's court to revoke the marriage under Article 1053 of the Civil Code (民法典)”. The state treats the premarital checkup as the first checkpoint in preventing birth defects. The current target is that “the premarital medical examination rate and the coverage of the target population for pre-pregnancy eugenic health checks are kept at 70% or more and 80% or more respectively”. The document also calls for “locating premarital medical examination institutions close to marriage registration sites for the public's convenience”, and promotes “‘one-stop’ services for marriage registration, premarital medical examination and childbearing guidance”. The Civil Code provides that a party with a serious illness shall truthfully disclose it before marriage registration, and if they do not, the other party may request a people's court to revoke the marriage. A request to revoke a marriage shall be made within one year from the day the party knew or should have known of the grounds for revocation.
Evidence grade
B
Notes
The premarital checkup has not been a required first step for registering a marriage since 2003; going is voluntary. But the rule about “not truthfully disclosing a serious illness” does give the other person the right to go to court to revoke the marriage, within one year of the day they find out. There is no nationwide rule that the premarital checkup is free. The 2020 notice from the National Health Commission and four other departments says only “actively encourage localities to introduce free premarital checkups in light of local conditions”. So whether it is free and which tests it includes are decided locally; ask at your local maternal and child health institution or at the one-stop window of the marriage registry.
Work out the exit costs: divorce by agreement has a 30-day cooling-off period, and divorce by lawsuit has statutory conditions
Value for cost High
In plain terms
Divorce by agreement isn't over once you sign. Within 30 days of filing the application, either person can withdraw it. After that period, the two of you must also go in person within 30 days to collect the divorce certificate; if you don't, it counts as withdrawn. If the other person doesn't agree, the only route is a lawsuit. The court mediates first, and grants a divorce when mediation fails only if there are circumstances such as domestic violence, bigamy, or living apart for a full two years.
Cost
Free. Before marrying, spend a little time reading through how long it would take, and what procedure you would go through, if you ever needed to get out. The hard part is thinking about splitting up before you even marry.
No moneyA few hoursSome willpowerBenefit size large
Benefit
Divorce by agreement requires signing a written divorce agreement and applying in person at the marriage registration authority (Article 1076). Within 30 days from the day the registration authority receives the application, either party may withdraw it. Within 30 days after that period ends, both parties must apply in person for the divorce certificate; not applying counts as withdrawal (Article 1077). When one party does not agree, the route is a lawsuit, and the court shall mediate first. Where there is bigamy or cohabiting with another person, domestic violence or maltreatment or abandonment, repeated gambling or drug use despite admonition, living apart for a full two years because of breakdown of affection, or similar circumstances, and mediation fails, a divorce shall be granted. If a divorce is refused and the couple then live apart for a further full year, a new lawsuit shall be granted (Article 1079). In 2024, 3,513,000 couples divorced nationwide, of which 891,000 went through court judgment or mediation.
Evidence grade
A
Notes
The cooling-off period applies only to divorce by agreement. Divorce by lawsuit has no cooling-off period, but it takes longer. Court fees and lawyer fees vary by place, so this item gives no numbers. High exit costs don't mean you shouldn't marry, but you should know these things before you marry.
Keep a separate account for “marrying for your elders”; don't mix it with your own
Value for cost Standard
In plain terms
Your elders' share of the benefits, that is, peace of mind, face and no more pressure to marry, is real. But it does not automatically turn into health benefits or relationship quality for you. And the three accounts of time, money and exit costs are borne entirely by you. Write two columns on paper, separating which benefits are yours and which are your elders', and you will see at a glance whether this deal is worth it.
Cost
Free. Write two columns on paper. One column is which benefits of this marriage are your own; the earlier items in this section let you work these out. The other column is which benefits are your elders': peace of mind, face, an end to the pressure to marry. The hard part is keeping the elders' column separate and not merging it into your own.
No moneyDone in passingSome willpowerBenefit size medium
Benefit
The elders' benefits are real. But they do not automatically turn into your health benefits (Item 8) or relationship quality (Item 4: whether a relationship is good depends on your own experience of it). And the time account (Item 9), the money account (Items 10 to 12) and the exit costs (Item 15) are borne entirely by the person who marries. Once the accounts are kept separately, you can see for yourself whether the value for cost of “marrying as a task” is high or low.
Treat relationship quality as a health account: when you argue, address the issue, not the person; no insults, no sarcasm
Value for cost Standard
In plain terms
Couples with good relationships are somewhat healthier and have a somewhat lower risk of death. The association is not large; it is about as large as the association between diet and health. In one experiment, small wounds made as part of the experiment healed only six-tenths as fast in couples who stayed hostile while arguing as in calm couples.
Cost
Free. When you argue, talk only about the matter at hand: no insults, no sarcasm, no dragging up old grievances. The hard part is holding your tongue in the heat of anger.
No moneyDone in passingSome willpowerBenefit size small
Benefit
Pooling 126 studies and more than 72,000 people (a meta-analysis): the higher the marital quality, the better the health, with correlation coefficients r between .07 and .21. A correlation coefficient of 0 means no relationship at all, and 1 means moving perfectly in step. Within this, the correlation with lower risk of death was r = .11, and with smaller cardiovascular reactions during arguments, r = -.13. The authors say these effects are all small, about as large as the association between health habits such as diet and health. In another crossover experiment, 42 healthy couples were each admitted to hospital twice, for 24 hours each time. One time they had a mutually supportive conversation; the other time they discussed a matter they disagreed on. After the argument, the blister wounds made in the experiment healed more slowly. Couples who showed more hostility both times healed at only 60% of the rate of low-hostility couples. The morning after the argument, the inflammation markers in their blood (IL-6, tumor necrosis factor) also rose more.
Evidence grade
B
Notes
Why grade B: most of the pooled studies only followed and recorded people, with no assignment to comparison groups. So it is impossible to tell whether a bad relationship makes people ill or poor health makes the relationship worse. In some studies the association remained after accounting for age and income. Robles's paper also states that some of the results show publication bias. The studies that explicitly compared men and women found basically no difference between them. This section found no pooled analysis directly comparing the death rates of “a bad marriage” and “not marrying”, so “a bad marriage harms your health more than not marrying” can only count as an inference. The wound experiment had only 42 couples. “Address the issue, not the person; no insults, no sarcasm” is a practice the author inferred from the “hostile behavior” in the experiment; no study has compared it directly. If your fights are fierce and you can't change that, go to couple therapy together; see Item 20 (couple therapy). If fights turn physical, or there is long-term verbal abuse or threats, it is no longer a communication problem; it is domestic violence; see Section 8, Item 43 (if you are being abused at home). To think through whether to leave, see Item 15 (exit costs) for the exit costs.
Sources
Robles TF, Slatcher RB, Trombello JM, McGinn MM (2014). Marital quality and health: a meta-analytic review. Psychological Bulletin. https://doi.org/10.1037/a0031859；Kiecolt-Glaser JK 等 (2005). Hostile marital interactions, proinflammatory cytokine production, and wound healing. Archives of General Psychiatry. https://doi.org/10.1001/archpsyc.62.12.1377
Same-sex partners, while you are both lucid, put in place a power of attorney, voluntary guardianship and wills: legally you are not close relatives, and without these you have no right to sign and no right to inherit
Value for cost Standard
In plain terms
Close relatives in law are people such as your spouse, parents, children, brothers and sisters; a partner doesn't count. Without documents prepared in advance, when one of you is seriously ill, the other cannot make decisions for them. When one of you dies, the other cannot inherit anything at all from the estate. The three documents each cover one stage: while alive, the power of attorney; after losing the capacity to make decisions, voluntary guardianship; after death, the will.
Cost
Voluntary guardianship and wills are done at a notary office; notary fees vary by place, starting from a few hundred yuan. A will you handwrite yourself costs nothing. Altogether it takes one or two half-days. The hard part is that the two of you have to sit down and talk about illness and death.
A little moneyA few hoursSome willpowerBenefit size large
Benefit
The close relatives listed in Article 1045 of the Civil Code (民法典) are spouse, parents, children, brothers and sisters, paternal grandparents, maternal grandparents, paternal grandchildren and maternal grandchildren. The statutory heirs under Article 1127 are first spouse, children and parents, and second brothers and sisters, paternal grandparents and maternal grandparents. A partner appears in neither place. Article 1219 covers surgery, special examinations and special treatments. Where it is impossible or inappropriate to explain to the patient, medical staff must explain to the patient's “close relatives” and obtain their consent. There are three routes to fill the gap. First, Article 33 allows an adult to agree in writing in advance with “other individuals willing to serve as guardian” on who will be their guardian. After the person loses all or part of their capacity for civil conduct, this individual becomes the guardian, and the guardian is the legal representative. Second, Article 165 provides that a written power of attorney must state the agent's name, the matters delegated, the scope of authority and the term, and be signed or sealed by the principal. Once the principal dies, the authorization ends. Third, Article 1133 allows making a will that gives personal property to individuals other than the statutory heirs. An article by the Yichang Municipal Justice Bureau published on the website of the Hubei Provincial Department of Justice states that same-sex partners can have a voluntary guardianship agreement notarized, naming each other as their future guardian (nationwide).
Evidence grade
A
Notes
A partner receives the estate as a “legatee”, not as an heir. After learning of the legacy, the partner must clearly state acceptance within 60 days; not stating it counts as giving it up. The partner, and people whose interests are bound up with the partner's, cannot serve as witnesses to the will. If several wills were made, the last one prevails; see Section 17, Item 2 (a later will overrides an earlier one). Whether a voluntary guardian can directly sign surgical consent forms varies from hospital to hospital; ask the hospital you usually use in advance. If family members don't recognize these documents, it means going to court: first-instance ordinary procedure takes 6 months or more, and you pay the lawyer's fees yourself. So have the documents done at a notary office wherever possible. For a home bought together or money saved together, write a separate written agreement based on what each person actually paid. For how to set up voluntary guardianship, see Section 17, Item 1 (name your future guardian in writing). The people who benefit are you and your partner.
Don't marry by hiding your sexual orientation, and don't treat a cooperative marriage as an easy way out: legally it is a real marriage, and the only way out is divorce
Value for cost Standard
In plain terms
Marrying while hiding your sexual orientation is not legally void, nor can it be revoked; the only way out is divorce. A cooperative marriage (形婚) is also a real marriage: money earned and debts incurred after marriage are all treated as the couple's. At divorce, the spouse who was deceived can ask for a larger share of the property and compensation for emotional harm, but whether courts will support this is not yet settled.
Cost
Free. The hard part is withstanding your family's pressure to marry, which may mean holding out for many years.
No moneyDone in passingLots of willpowerBenefit size medium
Benefit
Article 1051 of the Civil Code (民法典) provides only three situations in which a marriage is void: bigamy, a kinship relationship within the prohibited degrees, and not having reached the legal marriage age. Articles 1052 and 1053 provide only two grounds for revoking a marriage: being coerced into marriage, and concealing a serious illness before marriage. Article 17 of Interpretation (I) on the Marriage and Family Book (婚姻家庭编解释（一）) provides that a request to declare a marriage void on grounds other than these three is dismissed by court judgment. So the only way out is divorce. A cooperative marriage, once registered, is also a lawful marriage. Article 1062 provides that wages, bonuses, remuneration for services and so on earned after marriage are the couple's joint property. Article 1064 provides that debts the two signed for together, and debts one spouse incurred for the family's everyday living needs, are joint marital debt. Divorce damages fall under Article 1091, limited to bigamy, cohabiting with another person, domestic violence, maltreating or abandoning family members, and “other serious fault”. Article 2 of Interpretation (I) limits “cohabiting with another person” to cohabiting with “an extramarital person of the opposite sex”. So a same-sex relationship can only be argued as “other serious fault”. Damages include compensation for emotional harm; the deceived spouse must raise the claim together with suing for divorce, and if no divorce is granted, it is not supported. Since the Civil Code took effect, no published court judgments of this kind could be found. Under the old law, an article published in 2013 on the website of the Jiangsu Provincial High People's Court discussed a case in which a wife sued for divorce because her husband had hidden that he was gay. The article set out two views. One held that homosexuality was not among the faults defined by law at the time, so neither a larger share of property nor compensation for emotional harm should be supported. The authors endorsed the other view: treat the deceived spouse as the “party not at fault”, favor her when dividing the property, and also award her compensation for emotional distress (nationwide).
Evidence grade
A
Notes
On a larger share of property and compensation for emotional harm there is only one analytical article, and it does not say how the case was finally decided, so whether you can get them depends on the particular court. Compared with the old law, the Civil Code (民法典) adds “other serious fault”, which gives the deceived spouse an opening to argue. The statutes themselves are grade A. Where both spouses in a cooperative marriage have same-sex partners, the relationships with those partners are not protected by marriage law, and property must be put in writing separately; see Item 18 (same-sex partners set up voluntary guardianship and wills). Divorcing out of a cooperative marriage likewise means going through the 30-day cooling-off period or a lawsuit; see Item 15 (divorce by agreement has a 30-day cooling-off period). If a divorce agreement states that compensation is waived, a later lawsuit for it will not be supported. The person who benefits is first the spouse who would be deceived, and second you yourself: once the concealment comes out, divorce, property division and child-rearing arrangements all land on you at once.
If your relationship is stuck in repeated fights and cold wars, go to couple therapy together: just waiting basically won't make it better on its own
ContestedValue for cost Standard
In plain terms
Stuck couples who complete couple therapy together are clearly much more satisfied with the relationship, and about seven in ten improve. Couples who just wait on a waiting list barely change. But the effect fades: at the 5-year follow-up, about half are still clearly better, and about a quarter have separated or divorced.
Cost
It costs money. It is charged per session and varies by place and hospital, so this item gives no numbers. In the studies it was usually once a week, and one trial ran about 8 months in all. The hard part is that both of you have to go, and keep at it until it is finished.
Real moneyA few hoursLots of willpowerBenefit size medium
Benefit
A 2020 meta-analysis pooled 58 studies and 2092 couples. After couple therapy, the before-and-after change in relationship satisfaction was g = 1.12 (confidence interval 0.92–1.31), which the authors call a large effect. g is an effect size; see the glossary in the README. Couples assigned to a waiting list had g = 0.12 (−0.04 to 0.29) over the same period, no clear improvement. The worse the relationship at the start, the more it actually improved. A 2012 review states that couple therapy leads to improvement in 70% of the couples treated. A randomized trial took 134 couples in long-term, serious distress and gave them about 8 months of one of two behavioral couple therapies. At the end of therapy, the change in satisfaction was d = 0.90 and 0.71. By year 5, 50.0% and 45.9% of the couples in the two approaches respectively still had clear improvement; the authors call this clear improvement “clinically significant improvement”. At the same time, 25.7% and 27.9% had separated or divorced.
Evidence grade
A (contested)
Notes
Contested. The issue is how long the effect lasts. In another meta-analysis that included only randomized trials, g = 0.60 at the end of therapy and 0.44 six months later. At 12 months only behavioral couple therapy had data, with g = 0.06: the effect had barely held. Another review pooling 20 meta-analyses says clinically meaningful effects appear in 40% to 50% of people. “Seven in ten improve” and “four to five in ten clearly improve” use different standards, so don't compare them with each other. The 5-year trial had no untreated control group, so it cannot say whether divorce became more or less common. The 2020 analysis pooled only opposite-sex couples, and this item found no randomized trials in China either, so discount the findings when applying them to yourself. Where to go in China: the Mental Health Law (精神卫生法) requires psychotherapy to be done in medical institutions, and psychological counselors may not provide psychotherapy. You can ask the psychiatry or psychology department of a hospital whether they offer couple therapy. If you look for an organization outside a hospital, first ask about its qualifications and what training its staff have had. If there is domestic violence, don't do it together for now. The 2016 meta-analysis states that its conclusions apply only to mild to moderate situations in which both partners turn physical because of conflict. Violence marked by controlling the other person or by serious injury is not covered. Practitioners also worry about retaliation at home after a session. If you have been hit, first see Section 8, Item 43 (if you are being abused at home). For how to talk during fights, see Item 17 (address the issue, not the person); to think through whether to leave, see Item 15 (exit costs). The people who benefit are you and your spouse.
Sources
Roddy MK, Walsh LM, Rothman K, Hatch SG, Doss BD (2020). Meta-analysis of couple therapy: effects across outcomes, designs, timeframes, and other moderators. Journal of Consulting and Clinical Psychology, 88(7), 583-596. https://doi.org/10.1037/ccp0000514；Lebow JL, Chambers AL, Christensen A, Johnson SM (2012). Research on the treatment of couple distress. Journal of Marital and Family Therapy, 38(1), 145-168. https://doi.org/10.1111/j.1752-0606.2011.00249.x；Christensen A, Atkins DC, Baucom B, Yi J (2010). Marital status and satisfaction five years following a randomized clinical trial comparing traditional versus integrative behavioral couple therapy. Journal of Consulting and Clinical Psychology, 78(2), 225-235. https://doi.org/10.1037/a0018132；Rathgeber M, Bürkner PC, Schiller EM, Holling H (2019). The efficacy of emotionally focused couples therapy and behavioral couples therapy: a meta-analysis. Journal of Marital and Family Therapy, 45(3), 447-463. https://doi.org/10.1111/jmft.12336（争议方）；Shadish WR, Baldwin SA (2003). Meta-analysis of MFT interventions. Journal of Marital and Family Therapy, 29(4), 547-570. https://doi.org/10.1111/j.1752-0606.2003.tb01694.x（争议方）；Karakurt G, Whiting K, van Esch C, Bolen SD, Calabrese JR (2016). Couples therapy for intimate partner violence: a systematic review and meta-analysis. Journal of Marital and Family Therapy, 42(4), 567-583. https://doi.org/10.1111/jmft.12178；全国人大常委会 (2018 修正). 精神卫生法（第二十三、五十一条）. https://flk.npc.gov.cn/detail?id=2c909fdd678bf17901678bf7448a066d（国家法律法规数据库）
