VoxEU Column Gender Labour Markets

Explaining the persistence of the gender wage gap in Japan: The ‘glass ceiling’ and the ‘sticky floor’

Although the gender wage gap in Japan has been decreasing over the last 15 years, it remains large. This column shows that both the ‘glass ceiling’ and the ‘sticky floor’ exist in the Japanese labour market. The country’s human resource management system and a culture which rewards those who are willing to work outside of regular hours are to blame.

As in many countries, the gender wage gap has long drawn the attention of policymakers, researchers, and business people in Japan. Figure 1 shows that male and female wages in Japan have been converging since 1990.

However, as of 2014, the Japanese wage gap was the third highest after those of Korea and Estonia (among OECD countries), as shown in Figure 2. In short, the gender wage gap in Japan has been decreasing but is still large.

Figure 1 Change in the gender wage gap index in Japan, 1990–2016

Source: Japanese Ministry of Health, Labour and Welfare, Basic Survey on Wage Structure.
Note: Gender wage gap index = {(male monthly average wage) – (female monthly average wage)}/ (male monthly average wage)}×100 for full-time workers.

Figure 2 Gender wage gap index of OECD countries, 2014

Source: http://stats.oecd.org/
Notes: 1. Gender wage gap index = {(male monthly median wage) – (female monthly median wage)}/ (male monthly median wage)} ×100 for full-time workers. 2. Data for Israel are for 2011, for Spain for 2012 and for Sweden, France, and Chile for 2013.

What, then, impedes a further reduction? To address this question, in a recent paper I look beyond the average wage to examine the gender gap in Japan across the wage distribution (Hara 2016).

What is a ‘glass ceiling’ and a ‘sticky floor’?

The metaphors of the ‘glass ceiling’ and the ‘sticky floor’ represent the presumed barriers that exist at the extremes of the female wage distribution. These barriers prevent some women from reaching the upper echelon in business, government, and academia while keeping others in jobs with low wages. These metaphors imply that there might be invisible bottlenecks impeding the occupational advancement of women, which I call ‘subtle barriers’.

Recent advances in econometric methods, together with a popular notion of these subtle barriers in the labour market, has led to renewed interest in the topic, with recent studies building upon the extensive research on the gender gap at the mean to explore how and why the gender gap varies throughout the wage distribution.1

In order to explore the nature of the gender wage gap in Japan and its change over time, I adopt a decomposition method following Firpo et al. (2009) and Fortin et al. (2011),2 which enables me to separate the observed gender wage gap at each percentile of the wage distribution into two parts – one part due to differences in human capital attributes such as education, tenure or experience in the labour market (the compositional effect), and another part that cannot be explained by these human capital attributes (the wage structure effect). My interest here is in the latter, which represents gender-based differences in the returns to those attributes.

Focusing on this unexplained part of the gender gap, I define the ‘glass ceiling’ as the situation in which the wage structure effect is larger at the top of the wage distribution than in the middle range, implying that women who receive high wages (e.g. those in the 90th percentile among women) still receive lower wages than men with high wages (e.g. those in the 90th percentile among men). I define the ‘sticky floor’ as a wage structure effect that is larger at the bottom of the distribution than in the middle (e.g. the tenth percentile among women versus the tenth percentile among men).

Glass ceiling or sticky floor?

The results of the Firpo et al. (2009) decomposition are shown in Figure 3, and the blue dashed line shows the raw gender wage gap (calculated without controlling for human capital attributes). The red line, which shows the wage structure effects (i.e. the unexplained portion of the gender gap) at each percentile with controls for human capital attributes, is of interest here. Larger wage structure effects are observed at the bottom and top of the wage distribution than in the middle range throughout the period under study, indicating that both the glass ceiling and the sticky floor phenomena exist in the Japanese labour market. The figure also shows that both phenomena have occurred since at least 1990.3

Figure 3 Across- and within-establishment gender gap caused by the wage structure effect, 2014

Source: Hara (2016). Data are from the Japanese Ministry of Health, Labour and Welfare, Basic Survey on Wage Structure.
Notes: 1. ‘Across’ is estimated using OLS estimation, and ‘within’ is estimated using Mundlak’s estimation to control for fixed effects across establishments. Both estimations control for tenure and its square, potential experience and its square, and years of education. 2. The analysis sample is full-time workers at establishments employing both female and male workers.

Within a company or between companies?

Meanwhile, there could be two pathways for generating a gender wage gap: gendered job allocation into relatively lower paying jobs within a single company, and into low paying companies. The green line in Figure 3 shows wage structure effects while controlling for establishment fixed effects. The gender gap at each percentile reduces in relation to the red line, but still exists, indicating that both the glass ceiling and sticky floor phenomena exist even within an establishment.

However, the decline in the gender gap at each percentile (i.e. the difference between the red and green lines) is quite small. This suggests that job allocation into relatively lower paying jobs within a single company could be more significant for the gender wage gap in Japan.

What causes both phenomena?

Why do both the glass ceiling and the sticky floor exist in Japan? One factor behind the gender wage gap in Japan could be the characteristic human resource management system (that also has been adopted by companies worldwide) aimed at promoting employee participation and productivity in order to improve economic performance. The Japanese human resource management system, also known in the literature as the ‘innovative human resource management model’ or ‘high-performance work practices’, has attracted great interest from business leaders and researchers.4 The positive aspects of the Japanese human resource management system have long been emphasised, and this is bolstered by the ample empirical evidence that it actually improves a company's productivity.

However, a dark side to the Japanese human resource management system also exists. When companies introduce the system, gender segregation tends to take place because incentive schemes which prompt workers to willingly perform firm-specific human capital investment and remain at the company are key for making the system work. Theoretically, the company expects, ex ante, a high probability that a female worker will leave because she has a better outside option, which is her higher value of household production. Consequently, male workers are likely to be assigned to career-track jobs with a large emphasis on firm-specific skills that are highly valued and rewarded disproportionately while female workers are more likely to be assigned to non-career- track jobs with little opportunity to accumulate firm-specific skills. Such gendered job segregation could contribute to the gender wage gap.5

In fact, such a system, called the ‘career track-based management system’, exists in the Japanese labour market.6 This could cause quite a few women to be stuck in low-paying jobs, which might cause the sticky floor phenomenon.

In addition, Japanese companies have developed a wage determination system which rewards disproportionately workers who work long hours and particular hours (e.g. late at night and during holidays).7 Women are likely to choose jobs without long working hours or opt for flexible work. This could explain why women may not receive very high wages, which might be one of the reasons behind the glass ceiling.

To check whether these hypotheses are valid, I also perform the same analysis by company size and obtain different results for large and medium-size companies.8 When I restrict the analysis sample to workers at large companies with over 1,000 employees, the sticky floor is still observed, but the glass ceiling disappears. This suggests that the sticky floor phenomenon can be explained by the fact that female workers tend to be segregated into non-career-track jobs and thus do not accumulate higher-value, firm-specific skills. In contrast, when the analysis sample is restricted to workers at medium-size companies (100-999 employees), the glass ceiling is observed clearly, which could be explained by the wage determination system, although further research is necessary to conclude that.

We can say that the coexistence of the glass ceiling and sticky floor phenomena appears to be related to the Japanese human resource management system, and that restructuring it might be necessary for further reduction in the gender wage gap in Japan.

Editors' note: The main research on which this column is based appeared as a Discussion Paper of the Research Institute of Economy, Trade and Industry (RIETI) of Japan.


Albrecht, J, A Bjorklund and S Vroman (2003), “Is there a glass ceiling in Sweden?”, Journal of Labor Economics 21(1): 145-177.

Arulampalam, W, A L Booth and M L Bryan (2007), “Is there a glass ceiling over Europe? Exploring the gender pay gap across the wage distribution”, Industrial & Labor Relations Review 60(2): 163-186.

Carrillo, P, N Gandelman and V Robano (2014) “Sticky floors and glass ceilings in Latin America”, Journal of Economic Inequality 12(3): 339-361.

Chi, W and B Li (2008), “Glass ceiling or sticky floor? Examining the gender earnings differential across the earnings distribution in urban China, 1987-2004”, Journal of Comparative Economics 36(2): 243-263.

Chiang, H-Y and F Ohtake (2014), “Performance-pay and the gender wage gap in Japan”, Journal of the Japanese and International Economies 34: 71-88.

Christofides, L N, A Polycarpou and K Vrachimis (2013), “Gender wage gaps, “sticky floors” and “glass ceilings” in Europe”, Labour Economics 21: 86-102.

De la Rica, S, J J Dolado and V Llorens (2008), “Ceilings or floors? Gender wage gaps by education in Spain”, Journal of Population Economics 21(3): 751-776.

Firpo, S, N M Fortin and T Lemieux (2009), “Unconditional quantile regressions”, Econometrica 77(3): 953-973.

Fortin, N M and T Lemieux (1998), “Rank regressions, wage distributions, and the gender gap”, Journal of Human Resources 33(3): 610-643.

Fortin, N M, T Lemieux and S Firpo (2011), “Decomposition methods in economics,” in Handbook of Labor Economics, ed. by O Ashenfelter and D Card, vol. 4A, chap. 1, 1-102, Amsterdam: Elsevier.

Goldin, C (2014), “A grand gender convergence: Its last chapter,” American Economic Review 104(4): 1091-1119.

Hara, H (2016), “Glass ceilings or sticky floors? An analysis of the gender wage gap across the wage distribution in Japan”, RIETI Discussion Paper Series, 16-E-099.

Javdani M (2015), “Glass ceilings or glass doors? The role of firms in male-female wage disparities”, Canadian Journal of Economics 48(2): 529-560.

Kawaguchi, A (2015), “Internal labor markets and gender inequality: Evidence from Japanese micro data, 1990-2009”, Journal of the Japanese and International Economies 38: 193-213.

Lazear, E P and S Rosen (1990), “Male-female wage differentials in job ladders”, Journal of Labor Economics 8(1 Part2): S106-S123.

Moriguchi C (2014), “Japanese-style human resource management and its historical origins”, Japan Labor Review 11(3): 58-77.


[1] These include Fortin and Lemieux (1998) for the United States, Javdani (2015) for Canada, Arulampalam et al. (2007) and Christofides et al. (2013) for Europe, Albrecht et al. (2003) for Sweden, De la Rica et al. for Spain, Chi and Li (2008) for China and Carrillo et al. (2014) for Latin America.

[2] The Firpo et al. (2009) decomposition is a decomposition method for distributional statistics on the recentered influence function regression proposed by Firpo et al. (2009). See also Fortin et al. (2011) for details.

[3] See Hara (2016) for more information.

[4] See Moriguchi (2014).

[5] The theoretical model is provided in Lazear and Rosen (1990).

[6] Chiang and Ohtake (2014) and Kawaguchi (2015) show the potential adverse gendering effects of Japanese high-performance work practices, and Chiang and Ohtake (2014) examine the gender wage gap across the wage distribution. The focus of that study is on gender gap differences between two specific groups of workers (performance-pay and non-performance-pay), and they find that a glass ceiling exists for female white-collar workers under a non-performance-pay system.

[7] See Goldin (2014) for the United States.

[8] See the revised version of Hara (2016) for details (https://sites.google.com/site/hiromihara3/).

2,415 Reads