JILPT Research Eye
Industry-Specific Characteristic in Overwork by Young Regular Employees

September 28, 2017
(Originally published on December 5, 2016 in Japanese)


Tomohiro TAKAMI

Researcher, Department of Labour and Economy

The publication of Japan’s first White Paper on Prevention of Karōshi (Death from Overwork) this year underlines how serious the issue of preventing overwork has become in Japan. There are undeniably problems in Japan’s work environment, where deaths and suicides from overwork still occur. It is also beyond debate that Japan’s norm of grueling long working hours needs to be corrected above all.

When discussing the problematic nature of long working hours in this way, we usually focus on the question of how long working hours must be in order to be considered as a problem. But another question that should not be overlooked is in what way working hours must be long enough in order to be a problem. Even if we disregard extremely long working hours that make it difficult to maintain physical health, there would still be other working styles that cause problems for physical and mental health. For example, working without taking days off or constantly working overtime at home is thought to have a particularly significant damaging effect on mental and physical health. In such cases, not only extended working hours but the “content of the extended hours” itself causes problems for physical and mental health, and particularly for the latter. Meanwhile, even if working hours are not long, pressures of the work (such as always having to meet strict quotas) must place stress and burden on the working person if they are markedly intense. In some cases, the working style might have a more direct impact on mental health than simply the length of “working hours.” Here, I would like to introduce some results of research undertaken based on this perspective.

This research arose from a re-analysis of data from a questionnaire survey previously conducted by JILPT. [Note 1] Specifically, we focused on work-related stress (= “psychological stress”) that could lead to mental disorders in young (aged under 35) regular employees, [Note 2] and studied the problem of overwork while trying to discover what was involved in that stress. In terms of the characteristics of the analysis, we examined the difference among industries and studied how forms of overwork differ from industry to industry.[Note 3] Obviously, if working styles differ from industry to industry, aspects of work perceived as “demanding” would also of course differ. That is, it seemed inadequate to conclude uniformly that “working too hard” in connection with psychological stress was “a problem of long working hours,” but that there could be various forms of problems depending on the characteristics of the industry.

In the analysis, three characteristic industries were analyzed: “Education, learning support,” “Accommodations, eating and drinking services” and “Finance and Insurance.” This is because these three industries showed particularly high levels of psychological stress among working people, according to the data.

So, what kind of working styles cause psychological stress? Interestingly, workplace problems perceived by working people differed greatly by industry. ,Table 1 shows “Problems in the workplace or company” by industry. “Working hours” were widely brought up as a problem in all industries, but this was especially pronounced in “Education, learning support.” Conversely, “Holidays and time off” were most commonly cited in “Accommodations, eating and drinking services,” and the percentage citing “Management by performance quotas, results or progress” was conspicuously high in “Finance and Insurance.” Meanwhile, though omitted in the Table, these industry-specific characteristics could also be seen as differences in the reasons for working overtime. That is, while a large work volume, a sense of responsibility toward the work and others are cited as reasons for overtime in “Education, learning support,” staff shortages were given as the largest reason for working overtime in “Accommodations, eating and drinking services.” In “Finance and Insurance,” the main reason was that targets and performance quotas are too high.

Table 1 Problems in the workplace or company (by industry)
Working hours Holidays, time off Management by performance quotas, results or progress Bullying, harassment Induced voluntary retirement Others No particular problem N
Construction 22.0% 19.3% 5.3% 4.9% 1.8% 7.5% 39.3% 883
Manufacturing 20.1% 11.7% 8.9% 5.5% 1.8% 10.0% 41.9% 886
Electricity, gas, heat supply and water 18.0% 12.8% 7.9% 6.6% 1.1% 7.1% 46.4% 366
Information and communications 26.1% 7.6% 9.1% 3.2% 2.2% 9.7% 42.1% 877
Transport and postal services 27.3% 15.2% 5.7% 6.0% 1.3% 7.9% 36.6% 831
Wholesale trade 21.5% 16.2% 5.7% 6.7% 2.8% 9.8% 37.3% 864
Retail trade 21.4% 22.1% 6.6% 4.6% 2.2% 6.8% 36.2% 908
Finance and Insurance 16.0% 11.2% 21.9% 5.3% 1.6% 6.7% 37.2% 489
Real estate and goods rental and leasing 22.7% 18.8% 4.6% 3.5% 2.1% 9.0% 39.4% 480
Scientific research, professional and technical services 23.0% 13.3% 7.8% 4.6% 1.3% 12.0% 38.0% 474
Accommodations, eating and drinking services 27.7% 25.3% 4.7% 2.7% 1.4% 6.9% 31.3% 364
Living-related and personal services and amusement services 21.6% 20.1% 6.7% 5.2% 1.1% 8.9% 36.4% 269
Education, learning support 32.8% 15.1% 4.5% 6.3% 1.0% 6.3% 33.9% 872
Medical, health care and welfare 21.2% 21.5% 5.6% 7.1% 1.1% 10.9% 32.6% 978
Miscellaneous services 25.2% 14.8% 6.4% 5.2% 2.1% 11.8% 34.6% 485

Industry-specific characteristics are also significant causative factors behind psychological stress. Let us look at this from the example of “Education, learning support”, which is known as an industry with long working hours, but at the same time, is also known for its conspicuous volumes of “take-home overtime” (Figure 1). This should come as no surprise when we consider the classic image of the schoolteacher going home with test papers to mark, or the next day’s lessons to prepare.

Figure 1. Ratios of take-home overtime, by industry

Click to expand

In this industry, moreover, the frequency of take-home overtime was strongly related to psychological stress. Table 2 shows the result of analysis on factors involved in psychological stress. If we first analyze this without considering the level of “take-home overtime” (Model 1), psychological stress seems greater when overtime hours are longer. [Note 4] But if we take the frequency of “take-home overtime” into our analysis (Model 2), the level of take-home overtime directly makes an impact on psychological stress, while the length of overtime hours becomes less significant. In other words, the actual nature of problem lies in the amount or frequency of take-home overtime rather than “the length of overtime hours” as we previously saw.

Table 2 Determinants of psychological stress at work: Education, learning support (ordinal logistic regression analysis)
Analytical Model 1 Analytical Model 2
B Standard error B Standard error
Gender (Male=1, Female=0) 0.318 0.137 * 0.31 0.137 *
Age .062 .028 * .055 .028 *
Final educational level (BM: Junior/senior high school)
Vocational school, junior college, college of technology .254 .383 .147 .385
University / graduate school .494 .330 .429 .332
Marital status (BM: Unmarried)
Married -.134 .135 -.166 .136
Divorced / widowed -.026 .549 -.052 .551
Scale of regular employees in the business (BM: 300 or more)
Fewer than 10 -.140 .241 -.239 .243
10-99 .329 .202 .241 .203
100-299 .165 .277 .097 .278
Recruitment format (BM: New graduate recruitment)
Mid-career recruitment -.286 .156 -.280 .157
Internal promotion from non-regular employee -.289 .278 -.224 .280
Year joined company (BM: Up to 2004)
2005-2007 .119 .214 .059 .215
2008-2010 .222 .221 .168 .222
2011-2014 .269 .246 .169 .247
Managerial position (Yes=1, No=0) .041 .167 .020 .168
Annual income (BM: Less than 3 million yen)
3.00-3.99 million yen -.004 .177 .053 .178
4.00-4.99 million yen -.313 .193 -.315 .194
5 million yen or more -.298 .215 -.352 .216
Actual working hours per day .150 .041 ** .072 .043
Holidays per month (BM: 9 days or more)
Up to 4 days .967 .219 ** .777 .221 **
5-8 days .261 .142 .180 .143
Take-home overtime (BM: Never)
Often 1.094 .174 **
Sometimes .536 .153 **
χ2 value 54.862 ** 95.504 **
-2 log likelihood 2236.743 2197.488
Cox-Snell R2 0.061 0.103
Nagelkerke R2 0.065 0.111
N 877 877

**significant at 1% level *significant at 5% level

Then, why does take-home overtime cause stress? The answer cannot be found in the analysis. There are various possibilities; one interpretation is that taking work home on a daily basis makes it harder to detach the mind from one’s work even after going home. When this accumulates, it causes significant psychological stress. As said, this is just one interpretation, and a full elucidation is a task for the future.

In JILPT Research Report No.185, two other industries (“Accommodations, eating and drinking services” and “Finance and Insurance”) were analyzed in detail. The main points of the results are below.

  • A characteristic problem with “Accommodations, eating and drinking services” is that it is difficult for employees to take holidays and time off. When an employee cannot take enough holidays, especially no annual paid leave, psychological stress mounts up. Although staff shortages and personnel management are partly to blame for the difficulty in taking holidays and time off, it is possible that the increasing ratio of non-regular employees in the workplace could also affect this as a background factor.
  • In “Finance and Insurance,” significant problems are associated with results-based management. Although applying a system of target management is not a problem in itself, psychological stress increases when employees have too little discretion over the amount of work to be carried out. Also, when an individual’s results and performance are reflected in the monthly salary, it leads to fierce competition among employees. This raises psychological stress among working people.

Thus, we may surmise that directions for measures to protect the mental health of working people would be somewhat different from industry to industry. Naturally, this does not only concern the three industries discussed above. Based on the findings from the analysis, three types of overwork problem and measures in response to each of them are conceivable. First, in industries and workplaces where regular employees tend to have difficulty in taking holidays and time off owing to staff shortages and the growing trend of non-regular employment, the greatest priority must surely lie in measures to secure holidays, such as reviewing personnel management. Next, in industries and workplaces where employees tend to be psychologically worn out by harsh results-based management and competition among individuals, steps should be taken to optimize workloads so that employees are not burdened with excessive performance quotas. Management skills are also required of managers to prevent the workplace from becoming exhausted due to excessive competition. Finally, in industries and workplaces where long overtime hours are prone to occur due to a sense of responsibility and insistence on the job, as well as the amount of work, attention should be paid to management strategies, detailed counseling, and health management systems in order to prevent workers from taking on too much responsibility or duties.

To summarize, the analysis has led to the suggestion that measures in line with each individual industry and workplace are needed as well as correcting long working hours. When considering measures to prevent overwork, there would be practical significance in identifying (= classifying) points that require care in the characteristics of the individual industry or workplace. This research is one attempt to that end.

Note 1. The main text of this article has been published as Chapter 5 Causes of Psychological Stress in the Work of Young Regular Employees (PDF:2.1MB)(in Japanese) in the JILPT Research Report No. 185 [Polarization of working styles: Results of secondary analysis of JILPT questionnaire Surveys.]

Note 2. In this analysis, responses to “May become mentally ill” in the survey headings (four options from “Feel very strongly” to “Do not feel”) were used as indicators of psychological stress.

Note 3. The data used in the analysis were designed with a fixed sample assigned to each industry, and were thus suited to comparative analysis of industries. For details of the survey, see Research Series No. 136 [Survey on the workload on regular employees and the present situation of the workplace.]

Note 4. Table 2 appears to show that psychological stress is greater when the coefficient of a given variable is more clearly positive. Statistically significant figures are marked with *. Here, the discussion focuses particularly on results in the shaded parts.