The causal effects of working time on mental health: The effectiveness of the law reform raising the overtime wage penalty

• Published date: 01 December 2023
• Author: Miki Kohara,Taisei Noda
• Date: 01 December 2023
• DOI: http://doi.org/10.1111/1468-0106.12441

ORIGINAL ARTICLE

The causal effects of working time on mental

health: The effectiveness of the law reform

raising the overtime wage penalty

Miki Kohara

1

| Taisei Noda

2

1

Osaka School of International Public

Policy, Osaka University, Osaka, Japan

2

Department of Economics, Rice

University, Houston, USA

Correspondence

Miki Kohara, Osaka School of

International Public Policy, Osaka

University; 1-31 Machikaneyama,

Toyonaka, Osaka, 560-0043, Japan.

Email: kohara.miki.osipp@osaka-u.ac.jp

Abstract

The present paper reexamines the causal effect of

working hours on workers’mental health. We utilize

Japan's 2010 reform of the Labor Standards Act as a social

experiment to examine how the increased wage penalty

for long overtime work affects working hours and

workers’mental health. Utilizing a unique panel dataset

containing health behaviours as well as individual, house-

hold and workplace characteristics of male workers, we

find that the wage penalty reform indeed succeeded in

reducing overtime hours and total working hours and

that the reductions contributed to better mental health of

workers. Further empirical investigation suggests that the

reduction effect of the reform on working time is homo-

geneous among age groups; however, the harmful effect

of working time on mental health is large and statistically

significant among young workers. Our results suggest

that setting a high wage penalty for long overtime work

effectively reduces overtime work and improves workers’

health outcomes, particularly for young people.

1|INTRODUCTION

The link between working hours and mental health has attracted both academic and social

attention. It is well known that there is a strong correlation between these two factors.

However, the existence of causal relationships remains ambiguous. This is because a reverse

causality from mental health to working hours can exist. The healthier the individuals become,

the longer they work.

Received: 5 November 2023 Accepted: 5 November 2023

DOI: 10.1111/1468-0106.12441

638 ©2023 John Wiley & Sons Australia, Ltd Pac Econ Rev. 2023;28:638–664.wileyonlinelibrary.com/journal/paer

Is this relationship causal? Even though there is a tremendous amount of research examining

the correlation between the two, few empirical studies dealing with the causal effects of working

hours on workers’mental health have been conducted. Using data from 15 European countries,

Cottini and Lucifora (2013) show a causal relationship between adverse working conditions and

workers’mental health. They show that workers’mental health deteriorated with an increase in

job demands. Their measure of job demands reflects severe working conditions and work envi-

ronments, including long working hours.

Using a regression discontinuity design and difference-in-differences (DID) analysis with

the Swedish registered data, Hofmarcher (2021) shows that an increase in paid vacations in

Sweden does not improve young workers’health. Cygan-Rehm and Wunder (2018) conducted

a fixed-effect-instrumental-variable (FE-IV) estimations using statutory workweek regulations

in the German public sector as an instrument with a panel dataset to report that increased

working hours led to lower subjective and objective health measures.

Berniell and Bietenbeck (2020) examine the effect of French law reform on workers’health

by reducing the standard workweek at constant earnings. They conducted a DID analysis utiliz-

ing the fact that the targeted reform had different deadlines for implementation depending on

firm size. The study reports that the reform-induced reduction in working hours changed

workers’health behaviours (e.g., smoking) and improved body mass index and self-reported

health. A similar implication, but on worker satisfaction, was derived by Lepinteur (2019), who

applied the DID estimation to the reduction in standard working hours in Portugal and France.

For Asian countries, Nie et al. (2015) used the China Health and Nutrition Survey and

found little evidence that long work hours affect either workers’health or their health-related

activities. Taking the endogeneity problem more seriously, Ma (2023) conducted a dynamic

panel estimation with the China Family Panel Studies in the 2010s and showed that long work-

ing hours have positive and significant effects on the risk of mental illness. Ahn (2016), using a

panel dataset on workweek reduction in South Korea, showed that work-hour reduction

encourages workers to engage in more healthy behaviours.

Although health is not a target to be examined, Lee and Lee (2016) focus on the regulations to

reduce standard work hours and the differences among industries and establishment sizes. They

exploit the DID-IV estimation to show that a reduction in weekly working hours significantly

decreases the injury rate in South Korea. Using historical changes in the standard workweek in

Japan and Korea, Hamermesh et al. (2017) show that workers’satisfaction, which is strongly

related to mental and physical health, increases with a reduction in statutory working hours.

Focusing on workers’mental health in Japan, Kuroda and Yamamoto (2016) examine

Japanese employee-level panel data and show that the deterioration of mental health is attrib-

uted to increases in long working hours in Japan, mainly by using panel estimation to remove

unobserved heterogeneity among workers. Using the work records of a Japanese manufacturing

company and the original panel survey of employees in that company from 2015 to 2016, Sato

et al. (2020) also show that long and irregular working hours for Japanese workers deteriorate

their mental health, The results are noteworthy in the sense that they use the company's objec-

tive record of working hours but not a respondent's objective self-reported answer to the length

of work. They also control for individual fixed effects in the estimation.

This study addresses the endogeneity problem by utilizing Japan's 2010 reform of the Labor

Standards Act as a social experiment. The reform targeted workers of large firms in several indus-

tries and raised the wage penalty from 25% to 50% for overtime hours of more than 60 h per month.

The increase in the wage penalty might incentivize employers to reduce overtime by more than

60 h per month, which corresponds to approximately 15 h per week. We use this as a possible

KOHARA and NODA 639