VoxEU Column Education

The reduction of school days in Japan increased educational inequality

Japan switched to five-day weeks for its primary and junior high schools and saw an increase in educational inequality. This column discusses new evidence suggesting a loose tie between number of days at school and inequality. Importantly, this tie reflects the fact that homes with university-educated parents tend to offset the official reduction in hours with additional tuition.

One of the major objectives of compulsory education is to assure uniform educational opportunities for all children regardless of their socioeconomic background. For that reason, most advanced countries provide compulsory education as well as textbooks free of charge. Getting education policy right for those at an early age is also important for competitiveness in the global knowledge economy (Murtin and Viarengo 2013).

It has also been widely observed that children’s educational attainment is closely linked to the educational attainment of their parents, giving rise to intergenerational persistence in educational attainment. As a result, one would expect that an increase in the number of years of compulsory education will raise the years of schooling mainly of children born to parents with low educational attainment and hence weaken the link between parents’ and children’s educational attainment. Empirical studies using data for Sweden and Norway suggest that this indeed appears to be the case.

However, the total amount of compulsory education children receive depends not only on the number of years of compulsory education, but also on the amount of lessons within a year. Intergenerational persistence in educational attainment may also be influenced by the number of school days a year that a country adopts.

Studies suggest that the socioeconomic gap in students’ achievement tends to widen after summer breaks, which is thought to reflect heterogeneity in the home environment, because not all students continue study during the summer break. These findings provide indirect evidence regarding the interplay among schooling time, socioeconomic background, and academic achievement.

The number of days at school

However, it appears that to date no studies have directly examined this issue. I seek to examine how the decrease in school days in Japan through the introduction of the five-day school week in public primary and junior high schools has affected the intergenerational persistence in educational attainment.

Alongside the 1988 revision of the Labor Standards Act that reduced the legal work hours from 48 to 40 hours a week, increasing demands for schools to follow suit (in order to give teaching staff two days off in line with other employees and to allow children to spend Saturdays with their parents), the five-day school week was gradually introduced:

  • Initially, the second Saturday of each month became a holiday at public primary and junior high schools from September 1992;
  • From April 1995, the second and the fourth Saturday then also became holidays, and finally. From April 2002, all Saturdays became holidays.

Kawaguchi (2013) examines how the fact that in 2002 the first and the third (and then the fifth) Saturday became additional school holidays affected children’s study time and academic performance, and how changes therein differed according to children’s socioeconomic background measured in terms of parents’ educational attainment.

To examine changes in children’s study time, the study focuses on the time use of ninth graders – third-year junior high school students – using a large-scale survey on time use. The survey, conducted every five years, surveys approximately 200,000 persons over the age of nine, asking respondents about their time use on two designated consecutive days among nine days from the second Saturday to the third Saturday of October. Respondents are asked to report what they did on those two days in 15 minute intervals by choosing among 20 different options such as ‘school work’, ‘sleeping’, etc.

Figure 1 shows ninth graders’ average study time by day of the week and the household head’s educational attainment.

Figure 1. Ninth graders’ average study time (in minutes)


I compare the results for 2001 and 2006, which lie on either side of 2002 when, in addition to the second Saturday, the third Saturday of every month became a holiday.

The figure shows that ninth graders’ Saturday study time decreased irrespective of the household head’s educational attainment. The decrease in the case of junior high school-educated heads of household was 153 minutes, but university-educated heads of household was only 106 minutes. Moreover, although the weekday study time increased between 2001 and 2006 for all households, the increase for ninth graders in junior high school-educated households was only eight minutes, while it was 17 minutes in university-educated households. In addition, whereas the study time on Sundays decreased by eight minutes for ninth graders in junior high school-educated households, it increased by ten minutes in university-educated households.

Do fewer days of school have an effect?

Thus, looking at the changes in study time by day of the week, the results suggest that in the case of households that have university educated parents, the decrease in ninth graders’ Saturday study time was compensated for by an increase in study time on weekdays and Sundays. For ninth graders in households headed by junior high school graduates, almost no such compensatory behaviour can be observed.

Of course, a possible reason for this divergence could be an underlying long-term trend in growing inequality in study time. But this is not the case, as illustrated by the fact that between 1996 and 2001 the average study time of ninth graders increased in both junior high school-educated and university-educated households, and the increase was very similar: 24 minutes in the former case, 35 minutes in the latter.

According to regression analysis, it seems parents’ educational attainment is a statistically significant determinant of the change in study time between 2001 and 2006, while this is not the case between 1996 and 2001. The results remain qualitatively unchanged when including prefecture and year fixed effects in the estimation.

Overall, the various results show that the introduction of the five-day school week increased the dependence of children’s study time on parents’ educational attainment.

Inequality in study time and academic performance

How does this increase in inequality in study time among ninth graders with different socioeconomic backgrounds affect their academic performance? I analysed (2013) microdata of the 1999 and 2003 waves of the Trends in International Mathematics and Science Study (TIMSS), which assesses the mathematics and science competencies of eighth graders in several dozen countries around the world, as well as the 2000 and 2003 waves of the Programme for International Student Assessment (PISA), which assesses the competencies in reading, mathematics, and science of 15-year-olds in OECD countries.

These datasets do not contain information on parents’ educational attainment in its 2003 wave. However, they do contain information on the number of books in the home, which tends to be strongly correlated with parents’ educational attainment and can therefore be used to predict parents’ educational attainment.

The results for eight graders show that the average test score between children with high school-educated parents and with university-educated parents differed by about 4.0 standard deviations. This result in itself is not particularly remarkable, since it is well-known that children’s academic achievement is closely correlated to parent’s educational attainment. What is noteworthy, however, is that for 2003, i.e., after the full implementation of the five-day school week, the coefficient for an additional year of predicted parental education rises to approximately 0.12 standard deviations, which is about 20% higher than the value for 1999. In other words, the gap between children with high school-educated parents and with university-educated parents rose to about 0.48 standard deviations.

Taken on their own, these results for test scores do not necessarily provide conclusive evidence that the full implementation of the five-day school week resulted in an increase in educational inequality across socioeconomic groups. Yet, taken together with the above-mentioned increase in inequality in study time, it seems natural to assume that the two are causally related.


The full implementation of the five-day school week has led to an increase in socioeconomic disparities in both study time and test scores. Using the relationship between the two, it is possible to indirectly estimate the effect of study time on academic results. Doing so suggests that an additional minute of daily study time increases test scores by about 0.15 standard deviations. In other words, an additional hour of study a day raises the test score by nine standard deviations, indicating that study time is a major determinant of academic achievement.

This suggests that providing an environment that increases the study time of children whose parents have only completed junior high school themselves and whose study time has actually decreased as a result of the full implementation of the five-day school week would help to reduce socioeconomic inequality in academic achievement.

In Japan, many are concerned about a decline in the demands of compulsory education in Japan due to the full implementation of the five-day school week and the ‘thinning out’ of the curriculum leading to a decline in average academic achievement. Consequently, an increasing number of schools have started to offer supplementary classes on Saturdays. Attempts such as these not only can help to raise average academic achievement, but also are likely to raise the achievement of students from a low socioeconomic background and thus potentially help to decrease inequality between students from different socioeconomic groups.

A Japanese version of this column appeared in Hitotsubashi University's Hi-Stat Vox 25.


Kawaguchi, Daiji (2013), “Fewer School Days, More Inequality”, Hitotsubashi University Global COE Hi-Stat Discussion Paper Series 271, Hitotsubashi University.

Murtin, Fabrice and Martina Viarengo (2013), “The expansion and convergence of compulsory schooling: Lessons for developing countries”, VoxEU.org, 18 January.

4,934 Reads