A Bleak Economy Story Lowered Marriage Plans in Japan
In a December 2025 experiment, 2,270 unmarried adults in Japan rated marrying in the next three years lower after a fake article about layoffs and falling wages. A hiring-boom version raised that rating. The predicted scores were 2.05 and 2.47 on a 0-to-6 scale.
The article looked like news. It told unmarried adults in Japan that the next three years would bring bankruptcies, layoffs, fewer regular jobs, and wages falling as unemployment climbed.
Asked if they would marry inside that window, they scored the idea lower than when the article said companies would hire and pay would rise faster than prices.
Shigeki Matsuda, a sociologist at Chukyo University in Toyota, published the experiment in Genus on October 9. The journal is still labeling it an article in press.
That early version is already citable. The journal says a later edit can still replace it.
A story about the next three years can cool a wedding plan. Ask which forecast the household is using, and whether anyone is actually dating.
What they actually read
The survey ran online in December 2025 through Cross Marketing, a Japanese web-survey firm. It reached 2,270 people ages 23 to 32 who were unmarried, childless, and not students.
Quotas matched the 2020 census on sex and age among unmarried people. Schooling, jobs, and who lives at home can still look different from the country as a whole.
Each person read six fake articles, picked at random from 27 mixes. That is 13,620 readings. Every article combined a forecast for the economy, a forecast for births, and a line on government policy.
The policy line is a package: expand, keep, or cut youth job help together with child allowances and childcare. The study cannot separate those pieces.
After each article, Matsuda asked whether they would intend to marry in that version of the next three years. A second question asked about having a child. Both run from 0, strongly disagree, to 6, strongly agree.
The average marriage answer, across every article, was 2.3. The average answer on having a child was 2.0. The middle of the scale is 3.
A hiring story moved the score. A baby boom did not.
On marriage, the predicted score was 2.05 after the bleak economy and 2.47 when the article said hiring would jump and pay would rise ahead of inflation.
The middle story landed at 2.26. That was the article he used as the baseline: growth holds, regular jobs rise a little, wages rise a little and still lag prices, and unemployment eases. From the bleak article to the hiring boom is 0.42 points.
Plans for a child moved with the economy too, from 1.81 to 2.16. Expanding the job-and-allowance package landed at 2.15, and cutting it landed at 1.82.
The middle article sat at 1.97 in both of those comparisons.
A darker birth forecast moved the scores by less. Births falling as low as 660,000, with the population under 120 million, lowered both intentions. A recovery to 850,000 births and about 123 million people did not score higher than the middle article, where births fall to 700,000.
Matsuda's explanation is that a sudden turnaround may not have sounded real. When he asked what they expect, on a scale from rosy at 1 to bleak at 7, the population answer averaged 4.7.
45.6% leaned bleak, 37.1% sat in the middle, and 17.4% leaned toward a recovery.
Half had never had a partner
Just over half the sample, 52.7%, had never had a romantic partner. In the same model, people with a current partner scored 0.92 points higher on marriage than people who never had.
People living with their parents, 37.8%, scored about 0.16 points lower on marriage. People without a job, compared with people in regular jobs, scored about 0.43 points lower on marriage and about 0.51 lower on having a child.
Matsuda says those background gaps cannot be set next to the article effect. The article was assigned at random. A partner, a job, or a parent's house was already true of the person.
He also writes that the size of the article effect should not be read as the size a real news week would produce. A check that reweighted the panel toward the census left the conclusions in place. Those weights used unmarried adults ages 25 to 34, because the census tables he needed do not split out ages 23 to 32.
When relatives treat a gloomy headline as a reason to wait, separate the forecast from the paycheck. Then ask whether anyone is actually dating.