Impact of Commuting Time on Employees’ Job Satisfaction—An Empirical Study Based on China’s Family Panel Studies (CFPS)

Abstract

In China, job satisfaction has become a significant concern. Previous research has mainly focused on the impact of working conditions and personal characteristics on job satisfaction, neglecting the influence of commuting. This study utilized the ordered logistic (Ologit) regression model and demonstrated that commuting time negatively affects job satisfaction. Through additional analysis, it was found that increasing the duration of nap time can enhance employees’ job satisfaction. Additionally, providing a monthly commuting allowance of more than CNY 40.00, along with increased nap time, can help alleviate the negative impact of commuting on job satisfaction. Moreover, a heterogeneity analysis was conducted to explore the potential variations in this impact by gender, marital status, the employer’s nature, and region. The outcomes indicated that gender does not significantly affect job satisfaction in relation to commuting. However, individuals who are married or cohabiting, those employed in the private sector, and those residing in the northeast or central regions of China experience a negative impact on job satisfaction due to commuting. Finally, we propose relevant suggestions to improve employees’ job satisfaction and enhance their work efficiency in order to achieve the sustainable development of the company.

1. Introduction

Improving job satisfaction among workers is important for both companies and individuals. With the demographic changes, economic growth, and social development in China, the labor market in China has started to leave the surplus phase and labor shortages are frequent. Improving employees’ job satisfaction has emerged as a top priority for corporate human resource management, as high levels of satisfaction play a crucial role in attracting, recruiting, motivating, and retaining a company’s workforce [1,2]. Additionally, job satisfaction serves as a vital indicator of one’s life status and is strongly linked to an individual’s mental well-being, encompassing issues such as burnout, anxiety, and low self-esteem [3,4]. In modern times, individuals are no longer content with seeking employment but instead strive towards achieving a rewarding and gratifying profession. Due to this shift in perspective, job satisfaction has gained prominence as a critical social concern.

Job satisfaction has been extensively studied since the 1930s, and is generally defined as an employee’s emotional or attitudinal response towards their job [5,6]. Job satisfaction has been a topic of great interest since its inception. Currently, research on job satisfaction can be categorized into three main areas: firstly, research on the influencing factors of job satisfaction [1]; secondly, research on the impact of job satisfaction on other variables [7,8]; and, thirdly, research on the effects of job satisfaction as a mediating variable [9,10]. Among these, understanding the factors that influence job satisfaction is important for enhancing organizational performance and improving employees’ well-being.

This study focused on the factors influencing job satisfaction. Previous research has categorized these factors into work conditions and personal characteristics [1,3,11,12,13]. Commuting, being an essential part of work, is often overlooked when analyzing the influences on job satisfaction.

Commuting holds great significance in both daily life and work, particularly in China, a rapidly developing country. Following the trend of urban growth in developed countries, developing countries are experiencing a rapid expansion of urban space and an increase in the urban population, with many growth-related urban problems [14]. Since the reform and opening-up, China has made considerable strides through accelerated industrialization and urbanization [15]. In China, the proportion of urban populations has risen from 35.4% to 65.2% between 2000 and 2022 [16,17]. However, despite these accomplishments, a multitude of issues have arisen, with traffic congestion being a prevalent symptom of unbalanced urbanization [18,19]. Commuting time for employees has increased and the quality of their journey has decreased due to traffic congestion [19]. According to the 2022 Commuting in China’s Major Cities Monitoring Report, over 14 million people have a one-way commute exceeding one hour. The issues related to “extreme commuting” have garnered widespread social attention. Despite commuting not being officially recognized as part of working hours and lacking legal regulations, its undeniable impact on employees’ physical and mental well-being, workflow, and, ultimately, job satisfaction cannot be dismissed. Moreover, the influence of commuting on job satisfaction is complex. On one hand, commuting imposes physical and psychological stress on individuals [20,21]. On the other hand, commuting acts as an essential link between work and personal life, bringing positive impacts to both realms [22,23]. Thus, conducting research on the relationship between commuting and job satisfaction holds great significance in the practical and theoretical domains. This study focuses on Chinese residents as the research subjects and utilizes a large sample of microdata to examine the impact of commuting on job satisfaction. The aim was to draw more specific and precise conclusions and recommendations that can assist employers and employees in addressing the challenges posed by commuting and provide guidance for enhancing employees’ job satisfaction.

2. Theoretical Background and Hypothesis

Commuting has a significant impact on job satisfaction, but only a few scholars have delved into this aspect, leaving numerous unanswered questions. Traditional economics assumes that rational economic individuals weigh the costs of commuting against their income, assuming that long commutes will inevitably be compensated for by low rent or housing costs and higher-paying jobs. However, the reality often contradicts this traditional assumption of rationality, and this discrepancy in utility is often referred to as the “commuting paradox” [24]. Some studies support this paradox by suggesting that commuting is the most distressing time of the day, negatively affecting commuters’ physical and mental well-being and reducing their overall satisfaction [20,21,25,26,27]. Moreover, increased commuting stress leads to diminished employee satisfaction, increased fatigue, absenteeism, and even the intention to quit [28,29,30]. However, the existence of the commuting paradox has been challenged by several studies. It was discovered that commuters generally felt positive or neutral during their journey, and no evidence was found to link longer commutes to decreased subjective well-being [31,32]. It is worth noting, though, that a thorough analysis of these studies has uncovered fascinating insights. The positive or neutral emotions experienced during commuting may result from the satisfaction derived from employment during a recession [31]. Meanwhile, from a methodological standpoint, linear and ordered fixed-effects models are in line with previous research, demonstrating a negative association between the commute’s duration and subjective well-being [32]. Overall, most studies have concluded that commuting has both physically and psychologically negative effects, resulting in reduced subjective satisfaction [27,33]. Nonetheless, some scholars have argued that commuting serves as a psychological threshold or transition zone between home and work, assisting individuals in completing role transitions, and that few people desire to eliminate commuting entirely [22,23]. As such, the impact of commuting on subjective well-being is intricate. Job satisfaction, a crucial facet of subjective well-being, lacks consistent findings and extensive discussions regarding its relationship to commuting.

As is evident, there has been a lack of consistent conclusions regarding the effects of commuting on job satisfaction. Most scholars believe that commuting time has a significant negative effect on job satisfaction [20,33,34], while others argue against a direct relationship [29,35].

With limited research and inconsistent findings on the influence of commuting on job satisfaction, it is crucial to examine in depth how commuting impacts job satisfaction. We posit that commuting, being a necessary but unpaid and insecure part of work, can lead to physical and mental stress for employees, ultimately affecting their job satisfaction. On the basis of these insights, and in conjunction with previous research findings, we propose the following hypothesis:

Hypothesis 1.

Job satisfaction decreases as commuting time increases.

Furthermore, several studies have demonstrated that a reasonable amount of napping can enhance alertness, increase positive emotions, reduce fatigue and stress, and enhance an individual’s ability to handle complex work tasks, thereby increasing productivity [36,37]. Hence, in order to mitigate the negative impact of commuting on job satisfaction, we introduced nap time as an integral component of our study, aiming to investigate the significant role that naps play in influencing the impact of commuting on job satisfaction. Therefore, we propose the following hypothesis:

Hypothesis 2.

Increased nap time increases job satisfaction and mitigates the negative effects of commuting on job satisfaction.

We also selected the commuting allowance as a threshold variable to assess the threshold effect. Research has shown that commuting imposes both physical and psychological burdens on employees, leading to a decrease in their job satisfaction [27]. Therefore, companies can mitigate the negative effects of commuting on job satisfaction through in-kind compensation and cash subsidies. In-kind compensation includes company-allocated cars, shuttles, and parking spaces. Cash subsidies generally refer to the payment of commuting allowances. However, the in-kind compensation considered in the questionnaire only includes company cars and shuttle buses, resulting in a limited scope of in-kind compensation. Hence, we utilized the variable “monthly commuting allowance” in the questionnaire to analyze its threshold effect.

Hypothesis 3.

There is a threshold effect of the commuting allowance when analyzing the effect of commuting on job satisfaction.

3. Materials and Methods

3.1. Data

This study aimed to examine the impact of commuting on job satisfaction using the open-access data from China’s Family Panel Studies (CFPS) from 2020. Launched in 2010, the CFPS data follow more than 16,000 respondents from different regions of China every year or every two years. The CFPS sample is a multi-stage probability sample extracted using the implicit stratification method. The initial sample comprised administrative districts at the first stage (PSU), followed by administrative villages or neighborhood committees at the second stage (SSU), and family households at the final stage (TSU). This database focuses on the economic and non-economic well-being of the Chinese population, and examines various research topics, including economic activity, educational outcomes, population migration, and health. It covers 31 provinces in China, which provides comprehensive support for our analysis.

In addition, considering the factors that influence job satisfaction and aligning with China’s 2020 statutory retirement age, which is generally set at a maximum of 55 for women and 60 for men, except in exceptional circumstances, we chose a sample of individuals aged 16 years or older who were below 55 years for women and below 60 years for men, and removed the unemployed sample. As this study focused on the impact of commuting, we excluded individuals who do not have a fixed place of work from the sample. According to [38], self-employed individuals have access to more information, including commuting data, compared with employed individuals, and often make adjustments in their choices to maximize their utility. Therefore, commuting time has no effect on job satisfaction for self-employed individuals. Thus, this study focused on the effect of commuting on job satisfaction for the employed group and only retained the sample of the employed group. Additionally, in order to mitigate the influence of extreme values, we also winsorized wages and working hours at the 2% and 98% quartiles. Additionally, to facilitate the analysis of employers’ characteristics, we removed samples where the employers were individuals, families, private non-profit enterprises, organizations, associations, guilds, foundations, village councils, or samples where the nature of employers could not be determined. We found that these exclusions did not significantly alter the overall characteristics of the sample. After deleting samples with missing key variables, 6786 valid samples from 31 provinces were retained.

3.2. Variables

The dependent variable was job satisfaction. In this study, individuals’ subjective satisfaction with their jobs was selected as the dependent variable, with 1 corresponding to “very dissatisfied”, 2 to “not very satisfied”, 3 to “average”, 4 to “fairly satisfied”, and 5 to “very satisfied”.

The independent variable was commute time. In this study, commute time measured in minutes was selected as the variable for measuring commuting.

For the control variables, referring to previous literature, we classified the control variables into two broad categories: personal and work-related variables [20,33]. Personal control variables included gender, age and secondary terms of age, health, and current marital status. It is widely believed by scholars that the relationship between age and job satisfaction follows an inverted U-shape [33,38]. Thus, the secondary term of age was included as a control variable in this study. For the binary variable representing health status, individuals who considered themselves to be “very healthy”, “healthy”, or “relatively healthy” were assigned a value of 1, while individuals who considered themselves average or unhealthy were assigned a value of 0. Current marital status was also treated as a binary variable in this study. The marital status of individuals who were married or cohabiting was set to 1; in all other cases, it was set to 0. Work-related control variables included wages, working hours, job security, and educational attainment. Wages were defined as monthly after-tax earnings. Working hours were measured as the number of hours worked per week. Job security was based on whether the job provided a pension, and health, unemployment, work injury, and maternity insurance, with a value of 1 assigned if at least one of these types of insurance was provided, and 0 assigned otherwise. Educational attainment was captured by comparing the highest level of education completed by the individual to the educational level required for the job, with a value of 1 assigned if the individual’s highest level of education completed was greater than or equal to the level required for the job; otherwise, a value of 0 was assigned. The workplace was the primary site of employment for the respondent, with 0 representing outdoor work and 1 representing non-outdoor work.

The threshold variable was the transport allowance. In this study, the threshold variable was the amount of commuting allowance (in CNY, the official currency of China) per month).

The descriptive statistics of the sample are presented in

Table 1. Descriptive statistics.

Variable	Symbol	Mean	Type of Variable	Percentage (%)
Job satisfaction	Satisfaction	3.72	Very unsatisfied	0.94
			Unsatisfied	9.14
			Average	19.22
			Satisfied	58.69
			Very satisfied	12.01
Commute time	Commute	21.42	0–8 min	20.54
			9–10 min	19.6
			11–20 min	28.32
			21–30 min	16.46
			31–210 min	15.08
Gender	Gender	0.58	Female	41.63
			Male	58.37
Age	Age	36.94	16–28	23.33
			29–32	16.92
			33–38	19.86
			39–47	20.06
			48–60	19.83
Subjective health status	Health	0.87	Healthy	87.37
			Unhealthy	12.63
Marriage Status	Marry	0.76	Married	75.63
			Unmarried	24.37
Monthly wage (CNY)	Wage	4304.79	150–2500	21.9
			2501–3500	22.62
			3501–4500	19.22
			4501–6000	21.1
			6001–12000	15.16
Weekly worktime	Worktime	52.48	8–40 h	31.11
			40.1–48 h	14.9
			48.1–56 h	20.79
			56.1–70 h	21.6
			70.1–98 h	11.6
Job security	Secure	0.58	Provided	57.63
			Unprovided	42.37
Job education requirement	Edu	0.86	Achieved	86.37
			Unachieved	13.63
Workplace	Workplace	0.19	Outdoor	81.36
			Others	18.64

. It is evident that the mean value of job satisfaction for the entire sample is 3.72, with 70.7% of respondents expressing relatively high or very high job satisfaction. The mean one-way commute time for the entire sample was 21.41 min. Notably, Beijing and Shanghai, being the most popular cities for Chinese employees, exhibited a comparatively more stressful commute. Descriptive statistics for these two cities revealed a mean one-way commute time of 32.42 min, approximately 51.4% higher than the mean for the whole sample. Regarding the personal characteristics of the sample, 58% of the employees identified as male, while 42% identified as female. The average age of the employees was 36.94 years. Additionally, 87% of the employees considered themselves to be in relatively good health, and 76% of employees were married or cohabiting. Regarding the job characteristics of the sample, the employees had an average monthly salary of CNY 4304.79 and worked an average of 52.48 h per week. Furthermore, 58% of the employees had at least one type of insurance provided by their employer, 86% of employees had education levels meeting or exceeding the requirements for their respective jobs, and just 19% of respondents were engaged in outdoor work.

3.3. Methods

The main objective of this study was to examine the effect of commuting on job satisfaction. As job satisfaction is an ordinal variable, the study utilized the ordered logistic (Ologit) regression model, which is widely used in the literature for such analyses [39]. This model is an extension of the Logit model and is specifically designed for situations where the explanatory variable exhibits a fixed order. To validate Hypothesis 1, we applied the Ologit model [40], using the following baseline model:

$${\mathrm{Satisfaction}}_{\mathrm{i}}=\mathrm{F}(\mathsf{\beta }{\mathrm{Commute}}_{\mathrm{i}}+\mathsf{\gamma }{\mathsf{{\rm X}}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}})$$

(1)

In Equation (1), Satisfaction_i represents the explanatory variable, which corresponds to the job satisfaction of Sample i. Commute_i is the explanatory variable of primary interest, indicating the time spent by Sample i on a one-way trip to and from work. X_i includes a set of control variables that correspond to the personal and job characteristics of the sample, which are described in detail in Table 1; ε_i denotes the random error term. F(·) denotes a non-linear function, which takes a specific form:

$$\mathrm{F}\left({\mathrm{y}}_{\mathrm{i}}^{*}\right)=\left\{\begin{array}{c}1, {\mathrm{y}}_{\mathrm{i}}^{*}<{\mathsf{\mu }}_1\\ 2, {{\mathsf{\mu }}_1<\mathrm{y}}_{\mathrm{i}}^{*}<{\mathsf{\mu }}_2\\ 3, {{\mathsf{\mu }}_2<\mathrm{y}}_{\mathrm{i}}^{*}<{\mathsf{\mu }}_3\\ 4, {{\mathsf{\mu }}_3<\mathrm{y}}_{\mathrm{i}}^{*}<{\mathsf{\mu }}_4\\ 5, {\mathrm{y}}_{\mathrm{i}}^{*}>{\mathsf{\mu }}_4\end{array}\right.$$

(2)

where ${\mathrm{y}}_{\mathrm{i}}^{*}$ represents the existence of unobservable continuous variables behind y_i, referred to as latent variables, which satisfy:

$${\mathrm{y}}_{\mathrm{i}}^{*}=\mathsf{\beta }{\mathrm{Commute}}_{\mathrm{i}}+\mathsf{\gamma }{\mathsf{{\rm X}}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(3)

μ₁ < μ₂ < μ₃ < μ₄ < μ₅ are the cutoff points, all of which are coefficients to be estimated.

In addition, to verify the threshold effect of commuting time on job satisfaction (Hypothesis 3), the commuting allowance was selected as the threshold variable in this study. The threshold regression model was constructed as follows:

$${\mathrm{Satisfaction}}_{\mathrm{i}}={\mathsf{\beta }}_1{\mathrm{Commute}}_{\mathrm{i}}(\mathrm{Allowance}\le \mathsf{\gamma })+{\mathsf{\beta }}_2{\mathrm{Commute}}_{\mathrm{i}}(\mathrm{Allowance}>\mathsf{\gamma })+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(4)

In Equation (4), Satisfaction_i represents the job satisfaction of Sample i; Commute_i denotes the time spent by Sample i on the one-way commute to and from work; Allowance_i signifies the monthly commuting allowance of Sample i, which also acts as the chosen threshold variable; ε_i represents the random error term. The parameter γ represents the threshold value of the model to be estimated. When Allowance_i ≤ γ, the coefficient of the effect of commuting time on job satisfaction is β₁. When Allowance_i > γ, the coefficient of the effect of commuting time on job satisfaction is β₂. If β₁ ≠ β₂, this indicates that Hypothesis 3 is valid, meaning there is a threshold effect of commuting allowance, resulting in a non-linear relationship between the effect of commuting time on job satisfaction when influenced by the commuting allowance.

4. Results

4.1. Preliminary Analysis

Model 1 in

Table 2. Results of Ologit and marginal effects.

Variables	Model 1	Marginal Effects
		Very Unsatisfied	Unsatisfied	Average	Satisfied	Very Satisfied
Commute	−0.005 ***	0.00005 ***	0.00040 ***	0.00055 ***	−0.00048 ***	−0.00052 ***
	(0.001)	(0.00001)	(0.00009)	(0.00013)	(0.00011)	(0.00012)
Gender	−0.183 ***	0.00171 ***	0.01461 ***	0.02019 ***	−0.01757 ***	−0.01895 ***
	(0.053)	(0.00054)	(0.00428)	(0.00587)	(0.00515)	(0.00554)
Age	−0.089 ***	0.00083 ***	0.00707 ***	0.00977 ***	−0.00850 ***	−0.00917 ***
	(0.020)	(0.00022)	(0.00165)	(0.00225)	(0.00198)	(0.00213)
Age2	0.001 ***	−0.00001 ***	−0.00010 ***	−0.00014 ***	0.00012 ***	0.00013 ***
	(0.000)	(0.00000)	(0.00002)	(0.00003)	(0.00002)	(0.00003)
Health	0.747 ***	−0.00697 ***	−0.05944 ***	−0.08217 ***	0.07148 ***	0.07710 ***
	(0.072)	(0.00109)	(0.00596)	(0.00777)	(0.00707)	(0.00769)
Marry	0.143 **	−0.00134 **	−0.01138 **	−0.01573 **	0.01369 **	0.01476 **
	(0.068)	(0.00066)	(0.00544)	(0.00750)	(0.00654)	(0.00705)
ln_wage	0.186 ***	−0.00173 ***	−0.01478 ***	−0.02043 ***	0.01777 ***	0.01917 ***
	(0.037)	(0.00041)	(0.00298)	(0.00406)	(0.00358)	(0.00384)
Worktime	−0.016 ***	0.00015 ***	0.00128 ***	0.00177 ***	−0.00154 ***	−0.00166 ***
	(0.002)	(0.00002)	(0.00013)	(0.00017)	(0.00015)	(0.00016)
Secure	0.142 ***	−0.00133 ***	−0.01131 ***	−0.01564 ***	0.01361 ***	0.01468 ***
	(0.052)	(0.00051)	(0.00417)	(0.00575)	(0.00502)	(0.00541)
Edu	−0.382 ***	0.00357 ***	0.03040 ***	0.04202 ***	−0.03656 ***	−0.03943 ***
	(0.071)	(0.00080)	(0.00575)	(0.00781)	(0.00692)	(0.00740)
Workplace	−0.091	0.00085	0.00727	0.01006	−0.00875	−0.00944
	(0.065)	(0.00062)	(0.00518)	(0.00716)	(0.00624)	(0.00672)
Observations	6786	6786	6786	6786	6786	6786

Notes: Robust standard errors are given in brackets. *** p < 0.01, ** p < 0.05.

utilized the Ologit model, which includes the complete range of variables. The analysis revealed that both commuting time and the selected personal and job-related control variables significantly influence job satisfaction.

In Table 2, we provide the regression results for Model 1 using the Ologit model. However, as the coefficients in the Ologit model may not have intuitive interpretations, the significance and sign of the coefficients alone provide limited information. Therefore, we also calculated the marginal effects on job satisfaction of all explanatory variables in Model 1. The results of these marginal effects are shown in Table 2.

In this study, our main focus was on the marginal effect of commute time as the main explanatory variable for job satisfaction. We examined how changes in the commute time affected the probability of job satisfaction, assuming that all control variables in the model had their mean values. These results of Model 1 are presented in

Table 3. Marginal effects of commuting and naps.

	Model 1	Model 2
Commuting
Very unsatisfied	0.00004 ***	0.00004 ***
Unsatisfied	0.00038 ***	0.00037 ***
Average	0.00060 ***	0.00058 ***
Satisfied	−0.00052 ***	−0.00051 ***
Very satisfied	−0.00050 ***	−0.00049 ***
Naps
Very unsatisfied		−0.00001 **
Unsatisfied		−0.00012 **
Average		−0.00018 **
Satisfied		0.00016 **
Very satisfied		0.00015 **
N	6786	6786
Pseudo-R2	0.023	0.023

Notes: *** p < 0.01, ** p < 0.05.

. From Table 3, we can observe that when all control variables had their mean values, a one-hour increase in commuting time led to a 0.0024 increase in the probability of job satisfaction being categorized as “very unsatisfied”, a 0.0288 increase in the probability of being categorized as “unsatisfied”, and a significant increase of 0.036 in the probability of being categorized as “average”. Furthermore, for every one-hour increase in commute time, there was a decrease of 0.0312 and 0.03 in the probability of job satisfaction being categorized as “satisfied” and “very satisfied”, respectively. Thus, the longer the commute, the lower the employee’s job satisfaction. Commuting, while a necessary part of everyday life, is not considered as work time, and it can infringe upon personal time without remuneration or guarantees. Therefore, commuting has a negative impact on job satisfaction. These results provide support for Hypothesis 1, which posits a negative relationship between job satisfaction and commuting time.

After examining the marginal effects of individual control variables on job satisfaction, we discovered that personal characteristics greatly influenced job satisfaction. Specifically, job satisfaction tended to be lower among men compared with women. This discrepancy can be attributed to the social stereotype placing men as the primary breadwinners in the family, resulting in greater pressure to perform and a desire for career advancement. Consequently, men generally experienced lower levels of job satisfaction than women [33]. Moreover, when examining the impact of job control variables on job satisfaction, we found that job characteristics also played a significant role. The coefficients related to job characteristics were all statistically significant at the 1% level. Employees with a perception that their educational level matched or exceeded the job requirements reported significantly lower job satisfaction compared with those who believed their educational level fell short of the requirements. Generally, individuals with higher educational backgrounds possess higher expectations and demands, which often remain unfulfilled in their current job roles. As a result, they tend to be less satisfied with their jobs [41].

4.2. Additional Analysis

To mitigate the negative impact of commuting on job satisfaction, this section incorporated naps and commuting allowances into the baseline regression analysis to observe any changes in the relationship between commuting and job satisfaction.

Firstly, we examined the impact of napping. Research suggests that taking a nap enhances memory and learning capacity, thereby aiding individuals in managing complex work tasks and increasing overall productivity [37]. On the basis of this, we propose Hypothesis 2, which states that increasing nap time can enhance job satisfaction and mitigate the negative effects of commuting on job satisfaction. Model 1 in Table 3 displays the marginal effect of commuting on job satisfaction in the baseline regression analysis. Model 2 in Table 3 introduces the nap variable to the baseline regression and shows the marginal effects of commuting and napping. As illustrated in the results, an increase in nap time improves the likelihood of employees feeling “very satisfied” and “satisfied” with their jobs while reducing the probability of them feeling “average”, “unsatisfied”, or “very unsatisfied”. It can also be observed that, with the addition of naps, the extent to which an increase in commuting time leads to a decrease in employees’ job satisfaction is mitigated. Specifically, the probability of employees feeling “very unsatisfied”, “unsatisfied”, or “average” about their jobs decreases as commuting time increases. At the same time, the probability of employees being “very satisfied” or “satisfied” with their jobs also decreases. This provides support for Hypothesis 2.

In this study, we conducted an empirical study with Equation (4) in order to verify the existence of the threshold effect of commuting allowance on job satisfaction (Hypothesis 3), and the results of the validity test for threshold estimation are presented in

Table 4. Validity test for the threshold estimates of transport allowance.

Model	F-Value	p-Value	BS	1%	5%	10%
Single threshold	19.532 ***	0.000	300	6.646	3.678	2.358
Double threshold	1.644	0.217	300	7.002	4.159	3.094
Triple threshold	0.515	0.460	300	5.501	3.893	2.680

Notes: (1) F-values and p-values were calculated through 300 repetitions of bootstrap resampling; (2) *** p < 0.01.

. Table 4 indicates that the commuting allowance successfully passed the single threshold test at the 1% level, but did not pass the double threshold and triple threshold tests. Thus, we can draw a preliminary conclusion that there is a single threshold for the commuting allowance. To further investigate this, we examined the confidence interval plot for the single threshold and observed that the commuting allowance also passed the likelihood ratio test, as demonstrated in Figure 1.

From Figure 1, it can be observed that when LR = 0, the threshold estimate was CNY 40.00, i.e., γ = 40.00. Subsequently, a threshold regression was conducted to analyze the empirical findings regarding the influence of commuting time on job satisfaction, using the commuting allowance as the threshold variable. The results of this regression are presented in

Table 5. Results for the threshold effect of transport allowance.

	(1)	(2)	(3)
	Single	Double	Triple
commute_1	−0.00231 ***	−0.00231 ***	−0.00231 ***
	(−4.44)	(−4.46)	(−4.46)
commute_2	0.00161 *	0.000875	0.0000564
	(1.88)	(0.85)	(0.04)
commute_3		0.00305 **	0.00148
		(2.15)	(1.11)
commute_4			0.00305 **
			(2.15)
Constant	3.754 ***	3.754 ***	3.754 ***
	(259.30)	(259.30)	(259.30)
r2	0.00428	0.00452	0.00459
r2_a	0.00398	0.00408	0.00401
F	14.57	10.26	7.823
N	6786	6786	6786

Notes: Robust standard errors are in brackets. *** p < 0.01, ** p < 0.05, * p < 0.1.

. Since the double threshold and triple threshold did not pass the significance tests, only the results from the single threshold regression were analyzed. From the threshold regression results in Table 5, it is evident that when employees receive a monthly commuting allowance below CNY 40.00, the impact of commuting time on job satisfaction is significantly negative at the 1% level of significance. Conversely, when employees receive a monthly commuting allowance exceeding CNY 40.00, the impact of commuting time on job satisfaction shows a significant positive trend at the 10% level of significance. These findings indicate that the commuting allowance effectively mitigates the negative impact of commuting time on job satisfaction. Therefore, Hypothesis 3 is supported.

4.3. Heterogeneity Analysis

Previous studies have revealed variations in the impact of commuting on job satisfaction across individuals [28,33,34]. In order to gain a more comprehensive understanding of this variation, this study aimed to conduct a heterogeneous analysis of the sample based on gender, marital status, the nature of the employer, and region.

Traditionally, heterogeneity has often been analyzed in terms of gender, but the conclusions have been inconsistent [26,33,34]. Some researchers have found that longer commutes are more strongly correlated with job satisfaction for women in comparison with men [26,34]. They proposed that women’s heightened sensitivity to commuting time is likely due to their greater responsibility for daily household tasks [26,34]. However, some scholars have also discovered that commuting time has a greater impact on male employees’ job satisfaction. They argue that, on one hand, men demonstrate relatively less tolerance, making their psychological state more susceptible to environmental influences that act upon their job satisfaction. On the other hand, in terms of the division of household roles, men usually bear more stress, and commuting time indirectly affects their job performance and promotion opportunities, thereby influencing their job satisfaction as well [33]. Given the lack of consensus in the literature on this matter, this study initiated a heterogeneous analysis of the effect of commuting on job satisfaction, starting with gender. Ologit regressions were conducted on all variables, grouping the data on the basis of gender and subsequently calculating the marginal effect of commuting time on job satisfaction. The results are presented in

Table 6. Heterogeneity analysis of gender and marital status.

Satisfaction	Gender		Marital Status
	Female	Male	Unmarried	Married
	Commute	Commute	Commute	Commute
Very unsatisfied	0.00003 **	0.00005 ***	0.00002	0.00005 ***
Unsatisfied	0.00036 ***	0.00039 ***	0.00016	0.00041 ***
Average	0.00065 ***	0.00056 ***	0.00019	0.00070 ***
Satisfied	−0.00050 ***	−0.00053 ***	−0.00022	−0.00057 ***
Very satisfied	−0.00054 ***	−0.00047 ***	−0.00015	−0.00059 ***
N	2825	3961	1654	5132
Pseudo-R2	0.021	0.025	0.012	0.028

Notes: *** p < 0.01, ** p < 0.05.

.

It can be seen that as commuting time increased, the probability of job dissatisfaction among employees increased. The probability of this increase was smaller for female employees compared with males, indicating that males who are less satisfied with their jobs are more likely to be affected by commuting, while females are more likely to be influenced by other factors. Similarly, among employees with average job satisfaction, females were more likely to be affected by commuting. When the commuting time increased, the probability of a decrease in overall job satisfaction among male and female employees was roughly the same. However, the decrease was greater for males who were “satisfied” with their jobs, while it was greater for females who were “very satisfied” with their jobs. This suggests that the group of females with the highest job satisfaction was more susceptible to the negative impact of commuting, while males with relatively good job satisfaction were more prone to the negative effects of commuting. Therefore, we can conclude that there is not a significant difference in heterogeneity among the samples based on gender.

In previous studies, only a few scholars analyzed heterogeneity from the perspective of marital status. Unmarried employees tend to have weaker boundaries between work and life, while married employees face additional family roles and demands, which require them to transition and adjust their states for different domains [42,43]. Commuting plays a significant role in facilitating role switching for married individuals [42,43]. However, ref. [22] believed that married employees face relatively greater life pressures, so they have a higher tolerance for commuting compared with job characteristics such as salary and job security. It can be observed that the academic community has not yet reached a consensus on this issue. Therefore, we categorized the samples on the basis of marital status to observe the impact of commuting on job satisfaction for different groups.

We computed the marginal effects of commuting time on job satisfaction for different marital statuses, as shown in Table 6. It can be seen that an increase in commuting time did not result in a significant change in job satisfaction among unmarried employees. However, for married employees, an increase in commuting time significantly escalated the probability of job dissatisfaction and overall dissatisfaction, while significantly diminishing the probability of job satisfaction. Compared with unmarried employees, it is evident that the job satisfaction of married or cohabitating employees is more susceptible to the negative impact of commuting. We believe that this is because unmarried employees do not strictly separate their work and personal lives, thereby experiencing a minimal effect of commuting on their job satisfaction. On the other hand, married employees have additional family responsibilities, and an increase in commuting time can disrupt their work–life balance, leading to prolonged periods of high pressure in both domains, ultimately affecting their job satisfaction.

Due to the unique nature of civil service work and the differences in management systems compared with employees in state-owned enterprises and private enterprises, there are differences in income, benefits, and working hours [44,45]. Examining the descriptive statistics of employees from different types of employers in

Table 7. Descriptive statistics stratified by the nature of employers.

Variable	Civil Servants			State-Owned Enterprise			Private Enterprise
	Obs	Mean	SD	Obs	Mean	SD	Obs	Mean	SD
Satisfaction	1202	3.87	0.79	1000	3.71	0.80	4584	3.68	0.84
Commute	1202	19.36	17.26	1000	24.14	22.93	4584	21.36	20.79
Wage	1202	4053.58	2182.77	1000	4550.09	2459.21	4584	4317.14	2391.96
Worktime	1202	45.03	14.24	1000	49.20	15.31	4584	55.15	16.94
Secure	1202	0.82	0.38	1000	0.80	0.40	4584	0.46	0.50
Edu	1202	0.91	0.29	1000	0.90	0.31	4584	0.85	0.36
Workplace	1202	0.16	0.37	1000	0.22	0.41	4584	0.19	0.39

, we observed that overall job satisfaction was highest among civil servants. Although employees in state-owned enterprises had the longest commuting time, their job satisfaction was second only to that of civil servants and higher than that of employees of private enterprises. Looking at their job characteristics, we found that although civil servants had the lowest average salary level, their average working hours were shorter, and they had the highest insurance coverage, resulting in better overall treatment. Conversely, although employees in private enterprises had a higher average salary level, second only to employees in state-owned enterprises, they had the longest average working hours and the lowest insurance coverage, which was less than 50%, resulting in poorer overall treatment. Therefore, we believe that the impact of commuting on job satisfaction differs among employees of different types of employers, and we conducted a heterogeneity analysis, the results of which are shown in

Table 8. Analysis of the heterogeneity by the nature of employers.

Satisfaction	(1)	(2)	(3)
	Civil Servants	State-Owned Enterprise	Private Enterprise
Very unsatisfied	0.00002	0.00001	0.00005 ***
Unsatisfied	0.00022	0.00020	0.00041 ***
Average	0.00037	0.00029	0.00062 ***
Satisfied	−0.00012	−0.00029	−0.00060 ***
Very satisfied	−0.00050	−0.00022	−0.00048 ***
N	1202	1000	4584
Pseudo-R2	0.033	0.020	0.020

Notes: *** p < 0.01.

.

It can be observed that job satisfaction among civil servants and employees in state-owned enterprises was not adversely affected by commuting time. However, employees in private enterprises were negatively impacted by commuting time, with an increased likelihood of experiencing dissatisfaction or an average level of job satisfaction as the commuting time increased, and a decreased probability of experiencing satisfaction. Only the job satisfaction of employees in private enterprises was negatively affected by commuting time, meaning that as the commuting time increased, the probability of experiencing dissatisfaction or average job satisfaction increased, while the probability of experiencing job satisfaction decreased.

Due to the variations in economic development across China’s provinces, the National Bureau of Statistics divides China into four regions: the northeast region (incorporating Liaoning, Jilin, and Heilongjiang), the east region (including Beijing, Tianjin, and Shandong), the central region (comprising Shanxi, Henan, Anhui, Hubei, Jiangxi, and Hunan), and the west region (including Guangxi, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang, and Tibet). The eastern regions include Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan; the central regions include Shanxi, Anhui, Jiangxi, Henan, Hubei, and Hunan; and the western regions include Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Tibet, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang. Commuting patterns differ significantly across regions with variable levels of economic development [46]. Hence, the sample was partitioned into four regions to conduct a heterogeneity analysis.

Table 9. Descriptive statistics by region.

Variable	Northeast			East			Central			West
	Obs	Mean	SD	Obs	Mean	SD	Obs	Mean	SD	Obs	Mean	SD
Satisfaction	856	3.76	0.88	2716	3.70	0.82	1570	3.69	0.81	1644	3.75	0.83
Commute	856	23.12	21.21	2716	22.06	20.75	1570	19.49	18.69	1644	21.29	21.57
Wage	856	3822.38	2008.76	2716	4869.40	2645.32	1570	4036.62	2206.59	1644	3879.28	1997.87
Worktime	856	52.28	18.02	2716	51.72	16.08	1570	52.61	16.92	1644	53.71	16.90
Secure	856	0.58	0.49	2716	0.63	0.48	1570	0.54	0.50	1644	0.51	0.50
Edu	856	0.89	0.31	2716	0.88	0.33	1570	0.86	0.34	1644	0.82	0.38
Workplace	856	0.20	0.40	2716	0.15	0.36	1570	0.20	0.40	1644	0.23	0.42

presents the overall work-related features of the four regions. The data show that job satisfaction was higher on average in the northeast and west than in the east and central regions. Commuting time was longest in the northeast and shortest in the central region. Wages were highest in the east and relatively low in the northeast and west. Job security coverage was at its highest in the east, while it was relatively low in the west. Additionally, a greater proportion of participants in the west was employed in outdoor jobs, whereas only a small number were engaged in such work in the East.

Table 10. Heterogeneity analysis by region.

Satisfaction	(1)	(2)	(3)	(4)
	Northeast	East	Central	West
Very unsatisfied	0.00008 **	0.00004 **	0.00003 *	0.00005 *
Unsatisfied	0.00061 **	0.00032 **	0.00047 **	0.00035 **
Average	0.00079 **	0.00055 **	0.00076 **	0.00050 **
Satisfied	−0.00055 **	−0.00051 **	−0.00073 **	−0.00040 **
Very satisfied	−0.00094 **	−0.00041 **	−0.00053 **	−0.00049 **
N	856	2716	1570	1644
Pseudo-R2	0.024	0.020	0.028	0.029

Notes: ** p < 0.05, * p < 0.1.

displays the results of the regional heterogeneity analyses, demonstrating that longer commuting times had a more significant impact on job satisfaction in the northeast and central regions. Specifically, in the northeast region, the likelihood of experiencing extreme dissatisfaction, dissatisfaction or average satisfaction increased the most, while the probability of being extremely satisfied decreased the most. The likelihood of being satisfied showed a notable decline in the central region, with a negligible difference in the probability of average satisfaction compared with the northeast.

5. Conclusions

Overall, we used the Ologit model to analyze the impact of commuting on job satisfaction and found a significant negative correlation between them. Control variables related to individual and job characteristics also influenced job satisfaction. In terms of individual characteristics, overall job satisfaction was higher for female, healthy, and married or cohabiting employees compared with male, unhealthy, and unmarried employees. Additionally, job satisfaction displayed a downward-then-upward trend as the employees’ age increased. Regarding the job characteristics, employees with higher salaries and insurance coverage from their companies, and relatively lower educational levels exhibited higher job satisfaction compared with those with lower salaries, no insurance coverage, and relatively higher educational levels. Since the coefficients of the Ologit model provided limited information regarding the significance and coefficient signs, we also calculated the marginal effects of the explanatory variables on job satisfaction, with a specific focus on the marginal effect of commuting time. The results indicate that as the commuting time increased, the probability of employees feeling “very unsatisfied”, “unsatisfied”, or “average” rose, while the probability of feeling “very satisfied” or “satisfied” decreased, aligning with the overall negative impact of commuting time on job satisfaction.

To mitigate the negative effects of commuting on job satisfaction, we included variables such as nap time and commuting allowance in further analyses. By observing the impact of nap time on job satisfaction, we found that increasing nap time enhanced overall job satisfaction and, at the same time, reduced the negative impact of commuting time on job satisfaction. Additionally, we treated the commuting allowance as a threshold variable and analyzed its impact on the relationship between commuting time and job satisfaction. The results indicated that when the commuting allowance exceeded CNY 40.00 per month, it effectively alleviated the negative impact of commuting time on job satisfaction.

In this study, we analyzed heterogeneity by classifying the employees by gender, marital status, type of employer, and region. The findings suggested that the impact of commuting on job satisfaction did not significantly differ between the genders. However, married or cohabiting employees were more susceptible to the influence of commuting on job satisfaction, while unmarried employees were not affected by commuting. Additionally, the job satisfaction of civil servants and state-owned enterprise employees remained unaffected by commuting, with only employees in private enterprises experiencing the impact of commuting on their job satisfaction. Furthermore, individuals residing in the northeast and central regions tended to experience a more adverse impact on their job satisfaction as a result of commuting.

6. Practical Implications

On the basis of the analysis and research above, we can conclude that commuting time has a negative impact on job satisfaction. This negative influence is detrimental to both the physical and mental health of employees and reduces work efficiency, which is unfavorable for both employees and companies [25,28]. Therefore, we propose recommendations from the perspectives of employees and companies.

From the perspective of employees, since commuting time is not considered as working time, an increase in commuting time compresses private time. Therefore, we suggest that employees can engage in activities they enjoy during their commute, such as listening to music or podcasts, or watching videos. They can also utilize this time for personal growth, such as cultivating new hobbies or acquiring new skills. Research shows that higher autonomy leads to greater happiness and satisfaction while reducing stress levels [47]. Additionally, scholars suggest that social interactions have a positive impact on the emotional state during commuting, especially for long commutes, as social activities increase positive emotions and alleviate boredom and stress during the commute [31,48].

From the company’s perspective, the stress caused by commuting impacts employees’ morale, reduces work efficiency, and increases employees’ fatigue, absenteeism, and even attrition [28,29,30]. This is detrimental to the company’s long-term development. Therefore, companies should take appropriate measures to mitigate the negative impact of commuting on job satisfaction. For employee groups that are more susceptible to the negative effects of commuting on job satisfaction, such as married employees and employees in the private sector, companies should pay more attention to safeguarding their commuting rights to enhance their job satisfaction. Additionally, this study found that nap time and commuting allowances effectively mitigated the negative impact of commuting on job satisfaction. Therefore, we recommend that companies allocate time and space for employees to nap to alleviate fatigue and stress, thus improving work efficiency. Companies should also consider measures such as increasing commuting allowances to counteract the negative impact of commuting on job satisfaction.

7. Limitations and Directions for Further Research

Our study is subject to a few limitations. Initially, we studied commuting-related issues, and the matter of commuting should be classified by differentiating between exclusively personal transportation and a combination of public transportation with personal transportation. The latter mode of personal transportation represents the first and last miles. We only used the aggregate commuting time, which is the overall personal commuting time. In the future, it would be advantageous to differentiate between these two issues and to explore them further. In addition, it has been found when reviewing the literature on commuting that the choice of commuting mode can influence job satisfaction [49,50]. Promoting active commuting methods, such as cycling or walking, can enhance employees’ physical health and productivity [30,49]. Future studies could explore the impact of different commuting patterns on job satisfaction.