The Gender Wage Gap in International Comparison – Three Institutional Designs and the 134-Country Child Penalty Atlas
Japan's gender wage gap is roughly 1.9× the OECD average, ranking second worst among member countries. But the central question of this article is not how far behind Japan stands. Using OECD, ILO, and WEF comparative data alongside Kleven et al.'s (2024) Child Penalty Atlas covering 134 countries, we read the differences in how countries chose to answer the same problem – mandatory certification, index disclosure, voluntary disclosure – and structure what institutional design can and cannot change.
TL;DR
- Japan's 21.3% gender wage gap is approximately 1.9× the OECD average of 11.3%, the second worst among member countries. Yet 'gap size' and 'female labor force participation' operate as independent dimensions – closing the gap is not simply a matter of women working
- Institutional designs cluster into three types: mandatory certification (Iceland IS 85), index disclosure (France Index Pénicaud / EU Directive 2023/970), and voluntary disclosure (Japan). The allocation of burden of proof and the presence of financial penalties determine intervention intensity
- Kleven et al.'s (2024) Child Penalty Atlas, covering 134 countries, shows long-term post-childbirth income loss of 21-26% in the Nordic region versus 51-61% in Continental Europe. As economies develop, gaps in education and labor force participation narrow, but the Child Penalty persists and becomes the dominant driver of gender inequality
What Is Happening
Japan's position read through four data systems – OECD, ILO, WEF, and the Cabinet Office Gender Equality White Paper
21.3%. This is Japan's 2024 gender wage gap as calculated by the OECD using the median of full-time employees. The OECD average is 11.3%, meaning Japan's figure is roughly 1.9× that benchmark, second-worst among member countries after South Korea (approximately 29.3%). Considering that Belgium reports 1.1%, Italy roughly 4.0%, and France roughly 9.6%, Japan's level is plainly an outlier within the OECD.
* Source: OECD Gender Wage Gap indicator (2024). Defined as the difference between male and female median earnings of full-time employees, divided by male median earnings. Japan's 21.3% is ~1.9× the OECD average of 11.3%, second-worst among OECD countries.
Domestic indicators in Japan corroborate this international position. According to the Reiwa 7 Gender Equality White Paper, when the median full-time male wage is set at 100, the female median is 88.7 for the OECD average and 78.0 for Japan, consistent with the OECD indicator. The ILO's Global Wage Report 2024-25 reports that globally, women earn approximately 78 cents for every dollar earned by men – and Japan, despite OECD membership, sits close to this global average.
The World Economic Forum's Global Gender Gap Report 2025 places Japan at 118th out of 148 countries, the same rank as the previous year and the lowest among G7 nations. The overall score is 66.6%, below the world average of 68.8%; the Economic Participation and Opportunity subscore is 61.3% (112th place). Compared to the previous year, the economic subscore improved through a rise in the share of women managers and parliamentarians from 14.6% to 16.1%. Yet in absolute terms, Japan remains in the lower group of member countries.
Compressing this international comparison into a single number can obscure something important: the relationship between the gender wage gap and female labor force participation. Reading the ILOSTAT 2023 analysis of equal pay for work of equal value alongside national data, a distinctive distribution emerges. The Nordics (Iceland, Norway, Finland, Sweden) show gaps of 5-15% combined with high female labor force participation. Southern Europe (Italy, Portugal) shows gaps of 4-6% but lower participation than the Nordics. Japan and Korea show gaps of 22-31% with mid-range participation. The intuitive correlation that "if women work, the gap closes" – or its inverse – does not hold in the data. The gender wage gap and female labor force participation operate as independent dimensions.
Background & Context
Three institutional types – mandatory certification, index disclosure, voluntary disclosure – and the corresponding country designs
Three Institutional Designs
Faced with the same problem, countries have chosen different institutional responses. This article organizes them into three types, ordered by intervention intensity.
Intervention Intensity: Strong ←→ Weak
Mandatory Certification
Iceland (IS 85)
Coverage
25+ employees
Mechanism
Third-party certification required, burden of proof on employer, re-certified every 3 years
Index Disclosure
France Index Pénicaud / EU Directive 2023/970
Coverage
50+ (France) / 150+ (EU)
Mechanism
Annual composite index of 5 metrics, mandatory improvement if <75/100, 5%+ unexplained gap triggers joint evaluation
Voluntary Disclosure
Japan (Act on Promotion of Women's Active Engagement)
Coverage
301+ → 101+
Mechanism
Wage gap disclosure required; certification (Eruboshi) is voluntary; enforcement limited to reporting and guidance
* Intensity scored on three axes: (a) certification mandate, (b) financial penalty cap, (c) burden of proof allocation. Iceland concentrates all three on employers; Japan retains voluntary certification, no monetary penalty, and burden of proof on claimants.
First, mandatory certification (Iceland IS 85). Iceland enacted an equal pay law in 1961, but for decades it remained largely declaratory. In 2018, the Government of Iceland introduced the Equal Pay Standard (IS 85), requiring employers with 25 or more staff to certify their equal pay systems (third-party certification by accredited bodies for 50+ employees; firms with 25-49 may alternatively submit documentation directly to the Directorate of Equality). Certification must be renewed every three years. According to a European Parliament briefing from 2021, the core of IS 85 is the shift in the burden of proof. Previously, the person alleging wage discrimination had to prove the case; under IS 85, employers must demonstrate that they pay equally for work of equal value. This removes the burden on individual workers to litigate why "the same work is paid differently."
Second, index disclosure (France's Index Pénicaud and the EU Directive 2023/970). France introduced the Index Égalité Professionnelle (commonly Index Pénicaud), named after Minister of Labor Muriel Pénicaud, in 2019. It requires companies with 50 or more employees to publish a composite index (out of 100) of five metrics by March 1 each year: pay gap, raise rate gap, promotion rate gap, salary raises upon return from maternity leave, and the gender ratio among the top 10 highest paid. Companies scoring below 75 must improve within three years; non-compliance can trigger fines of up to 1% of payroll. In its first year of operation, French government statistics showed many large companies scoring below 75 and being forced to file improvement plans.
The EU is extending this approach with the Pay Transparency Directive 2023/970, adopted in 2023. Member states must transpose it by June 7, 2026. In its first phase, employers with 150 or more workers must report gender pay gaps; gaps exceeding 5% that cannot be objectively justified trigger mandatory joint pay assessments with worker representatives. The directive embeds a shift in the burden of proof, requires salary range disclosure to applicants, and bans gender-differentiated wages in job advertisements. France's Index Pénicaud is being realigned in 2024-2025 to integrate with the EU directive.
Third, voluntary disclosure (Japan's Act on Promotion of Women's Active Engagement). Under a notification linked to the Act on Promotion of Women's Active Engagement, from July 2022 companies with 301 or more regular employees became required to disclose gender wage gaps. The scope of this disclosure requirement is scheduled to expand to companies with 101 or more employees in April 2026. Certification schemes ("Eruboshi" and "Platinum Eruboshi") exist but remain voluntary, and mechanisms tying certification directly to subsidies or procurement preferences are limited. Sanctions extend only to reporting, advice, and guidance – there is no direct financial penalty for non-compliance. Across the three axes of intervention intensity (certification mandate, financial penalty, burden of proof), Japan applies the lightest load on employers.
Quotas as Complement (the Norwegian Type)
Distinct from the three types above, quotas intervene directly in the upstream opportunity structure. In 2003, Norway legislated a minimum 40% female representation on listed company boards, with enforcement from 2005. Non-complying firms face deregistration and dissolution risk. Norway also introduced a father's quota for parental leave, raising paternal take-up from 4% to 90%. Quotas intervene in upstream decision-making rather than downstream wages, and so they differ in character from mandatory certification and index disclosure, which target wages directly. The Nordic countries run both systems in parallel.
Reading the Structure
The limits of institutional design read through the Child Penalty Atlas and Goldin's 'greedy work'
What the Child Penalty Atlas Reveals
What do differences in institutional design produce? On this question, Henrik Kleven, Camille Landais, and Gabriel Leite-Mariante delivered decisive data in their Review of Economic Studies article "The Child Penalty Atlas" (2024). First released as NBER Working Paper w31649 in 2023 and formally published in ReStud Vol. 92, Issue 5, the study applies a pseudo-event-study method to cross-sectional data from 134 countries, estimating the impact of childbirth on female employment and earnings by country.
Public childcare + parental leave + quotas
Private childcare + short parental leave + disclosure
Single-earner model + tax incentives
Long-hours norm + post-leave career penalty
* Child Penalty = long-term reduction in female labor income relative to pre-birth, estimated via pseudo-event study (Kleven et al. 2024). The 21-26% Nordic vs 51-61% Continental Europe gap reflects differences in childcare provision, parental leave, and working-hour norms. Japan values are estimates from the same dataset.
The headline results are clear. Long-term post-childbirth income loss stays at 21-26% in the Nordic countries (Denmark, Sweden), rises to 31-44% in Anglo-American countries (US, UK), reaches 51-61% in traditional breadwinner economies (Germany, Austria), and 55-70% in East Asia (Japan and Korea estimates). (The companion piece by Kleven, Landais, and Søgaard (2019), focused on Denmark alone, observed a long-term ~20% reduction in female income immediately after childbirth, providing the benchmark against which the Atlas estimates are calibrated.)
The Atlas's broader implication, visible on the childpenaltyatlas.org dashboard, is that as economies develop, gaps in education and labor force participation narrow, but the Child Penalty persists and increasingly becomes the dominant driver of gender inequality. For countries like Japan – OECD members with persistently large wage gaps – the primary driver is no longer differences in education or labor force participation but the divergence of income trajectories around childbirth.
Goldin's "Greedy Work" and National Labor Practices
What pathway produces the Child Penalty? Claudia Goldin proposed the concept of greedy work in her 2014 American Economic Review article "A Grand Gender Convergence: Its Last Chapter." Within a single occupation, workers able to put in long hours are paid non-linearly more, which Goldin identifies as the central driver of the gap. The parent shouldering childcare (overwhelmingly women) cannot meet the time demands of greedy work, leaving a residual gap even on hourly wages.
This concept maps sharply onto Japan's regular employment model. As Kazuo Yamaguchi showed in his September 2024 briefing at the Prime Minister's Office, capacity for long overtime functions as an effective criterion for promotion to management in Japan. The L-shaped curve – women's regular employment ratio peaking at 25-29 and declining thereafter – reflects a structure where women in childrearing years cannot meet the time profile demanded by greedy work and are pushed into non-regular employment.
The Limits of Institutional Design
We return to the starting question. Why did each country choose its design, and how does that choice act on the Child Penalty?
What emerges from the three-type comparison is that disclosure-based systems (voluntary and index disclosure) do not by themselves reduce the Child Penalty. Japan's Act on Promotion of Women's Active Engagement has substantially expanded the number of disclosing firms since 2022, but the L-shaped curve and the slow rise in women's management share have not changed structurally. France's Index Pénicaud imposes improvement obligations on firms scoring below 75, but it is not designed to intervene in the divergence of career trajectories around childbirth.
In contrast, Iceland's IS 85 and the EU Directive 2023/970 build in a shift in the burden of proof. Because employers must demonstrate that they pay equally for work of equal value, the criteria for evaluating positions and occupations become subject to ex-post audit. This is not a measure that directly reduces the Child Penalty, but it makes it difficult to maintain greedy-work norms while leaving an "unexplained gap" untouched.
Cukrowska-Torzewska and Matysiak's (2020) meta-analysis in Social Science Research synthesized 208 estimates across 34 countries from 1995 to 2018. The motherhood wage penalty averaged 3.6% but was significantly smaller in countries with developed childcare provision and higher paternal leave take-up. The motherhood penalty is not a fixed demographic fact but a function of institutional design.
A RIETI study comparing Japan and Korea found that Japan's occupational segregation index exceeds Korea's, with women over-represented in low-wage occupations and under-represented in high-wage ones. The interlocking of occupational segregation and statistical discrimination narrows opportunity at the entry point, before greedy work even comes into play.
Options Japan Could Learn From
If Japan were to move from voluntary disclosure, what are the realistic options?
First, fully importing Iceland-style mandatory certification would require substantial changes in labor practices. Mandating certification for employers with 25 or more workers (with third-party certification for those with 50 or more) would require simultaneous development of job evaluation systems, accreditation schemes for certification bodies, and administrative oversight. The implementation cost is high.
Second, France's index disclosure type extends naturally from the current Japanese framework. Disclosure obligations already exist, so converting to a composite index of five metrics, adding improvement obligations for low-scoring firms, and introducing financial penalties involves a smaller institutional leap.
Third, alignment with EU Directive 2023/970 is already necessary for Japanese firms operating EU subsidiaries with 150 or more workers. Movements to reverse-import EU subsidiary practices into headquarters have already begun at some companies, and over time this may exert pressure to dissolve the dual structure with domestic rules.
Fourth, Nordic-style board quotas remain an option for staged introduction. Japan has already set a target of 30% female board representation in the Prime Market by 2030, but without sanctions for non-compliance it falls short of Norway's enforcement model.
What the international comparison suggests is that "disclosure alone" cannot reduce the Child Penalty. Disclosure is effective as a first step toward visibility, but for visible numbers to translate into changes in position evaluation, hiring criteria, and post-leave return conditions, additional institutions that shift the burden of proof are needed. While the companion piece "Inside Japan's 22.1% Gender Wage Gap" dissected the domestic structure, this article reads the differences in institutional responses to the same problem in international comparison. Laid one over the other, Japan's challenge is not that "the numbers are bad" but that "the intervention intensity is low."
Related Articles
- Inside Japan's 22.1% Gender Wage Gap – The Structure That 'Equal Pay for Equal Work' Cannot Explain
- 'Not Having Time' Is Not a Personal Problem – The Structure of Time Poverty Created by a 5.5× Unpaid Labor Gap
- The Structure of the Women's Active Engagement Act Revision – What Numerical Targets Do and Do Not Change
References
OECD Gender Wage Gap indicator — OECD. OECD
Gender Equality White Paper Reiwa 7 (Part 1, Figure 2-3: International Comparison of Gender Wage Gap) — Cabinet Office Gender Equality Bureau. Cabinet Office of Japan
Global Gender Gap Report 2025 — World Economic Forum. World Economic Forum
Global Wage Report 2024-25: Is real wage growth sustaining the recovery? — International Labour Organization. ILO
The Child Penalty Atlas — Kleven, H.; Landais, C.; Leite-Mariante, G.. The Review of Economic Studies, Vol. 92, Issue 5, pp.3174-3207
Children and Gender Inequality: Evidence from Denmark — Kleven, H.; Landais, C.; Søgaard, J. E.. American Economic Journal: Applied Economics, Vol. 11, No. 4, pp.181-209
A Grand Gender Convergence: Its Last Chapter — Goldin, C.. American Economic Review, Vol. 104, No. 4, pp.1091-1119
The Motherhood Wage Penalty: A Meta-Analysis — Cukrowska-Torzewska, E.; Matysiak, A.. Social Science Research, Vol. 88-89
A Comparative Study of Gender Inequality: Occupational Segregation in Japan and Korea — RIETI. RIETI
Main Determinants of the Gender Wage Gap and Policy Measures (briefing material) — Yamaguchi, Kazuo. MHLW / Prime Minister's Office briefing material
Equal Pay Certification (IS 85) — Government of Iceland. Government of Iceland
Pay Transparency in the EU: A new directive — European Parliament. European Parliamentary Research Service
Directive (EU) 2023/970 on pay transparency — European Union. EUR-Lex
ILOSTAT Statistics on Women — International Labour Organization. ILOSTAT
Equal Pay for Work of Equal Value: Where do we stand in 2023? — International Labour Organization. ILOSTAT
