ANALYSIS OF THE INFLUENCE OF WORKLOAD AND SUPERVISION ON EMPLOYEE PERFORMANCE THROUGH WORK STRESS AS AN INTERVENING VARIABLE IN PRODUCTION DEPARTMENT OF PT YANMAR DIESEL INDONESIA

Work stress is often caused by work that is assigned to employees with excessive workload which can affect the quality of performance. Therefore supervision is very important to balance the two. This study aims to determine the effect of workload, work stress, and supervision on employee work performance at PT Yanmar Diesel Indonesia, and to determine work stress can mediate the effect of workload on employee performance at PT Yanmar Diesel Indonesia. This study uses a quantitative method with data collection using a Google form questionnaire. The results showed that Workload (WL) has a negative effect on employee performance (KK) with a coefficient value of -0.751. Supervision (PS) has a negative effect on employee performance (KK) with a coefficient value of -0.101. Workload (WL) has a negative effect on employee performance (KK) through work stress (SK) with a coefficient value of -0.080. Supervision (PS) has a negative effect on employee performance (KK) through work stress (SK) with a coefficient value of -0.39. Work stress (SK) has a negative effect on employee performance (KK) with a coefficient value of -0.136. Workload (WL) has a positive effect on work stress (SK) with a coefficient value of 0.587. Supervision (WL) has a positive effect on work stress (SK) with a coefficient value of 0.289.


INTRODUCTION
The development of the company, both in the product and service industries, has proven to be very high. Many companies want to survive and maintain their presence in order to capture the larger market effectively. Companies need to pay attention to their performance, so that companies can determine competitive strategies against their competitors. Fast-moving competition requires a more dynamic corporate strategy to deal with current competition. Competitors in business create a higher level of change and competition by prioritizing the quality of employees (Tulangow, Saerang, & Rumokoy, 2018).
Indonesia is known as an agricultural country with supporting tools (Agricultural Machinery Manufacturing) where these companies carry out tough competition. In order for businesses to continue to run, be sustainable, and have the courage to compete, they are always required to improve the quality of their services and products. The role of Agricultural Machinery Manufacturing companies is very large in carrying out the development, maintenance and maintenance of the agricultural industry in Indonesia. Based on information obtained through PT Yanmar Diesel Indonesia's website (www.yanmar.com), PT Yanmar Diesel Indonesia in carrying out its production process prioritizes the quality and capabilities of Human Resources (HR) compared to the use of modern technology. HR is an important asset that can create a company's reputation and success in the future (Manthosi and Makhubele, 2016).
Heavy equipment such as agriculture, construction, motor industry, electrical products, energy systems, and large generators are manufactured and exported by the company PT Yanmar Diesel Indonesia. The impact of the Covid-19 pandemic on PT Yanmar has also seen several changes. In 2019, an employee of the Production Department of PT. Yanmar has a total of 443 employees, while in 2020 the number of employees. The production department was reduced to 344 due to pandemic conditions and a decrease in the number of requests for machines, while in 2021 the demand for machine production increased so that the workload that had to be carried out by PT Yanmar's Production Department employees was increasing and getting heavier. The reduction in the number of employees due to the pandemic forced the surviving employees to eventually have to increase their workload to fill the vacancies left by previous employees. Too much or too little work causes physical fatigue, decreased concentration, self-control and thoroughness in work, resulting in substandard work.
Companies must be able to survive in a dynamic global market and PT Yanmar is no exception. The steps taken are to increase the duties of employees, increase the level of work pressure, in order to achieve the company's mission of improving company performance and quality. On the other hand it becomes a problem and burdens employees. Too much work is received even beyond the capacity of employees, causing work stress for employees. According to Omar, Aluwi, Fauzi, & Hairpuddin, (2020) a very dense and busy work environment can trigger mental and physical health problems, resulting in an increase in work stress. Workload can also be called workload referring to the amount of work allocated to an employee to do or the intensity of work tasks. Resilient employees do a good job and enjoy the workload. However, when the workload becomes redundant it has a negative impact on employees.
The workload for PT Yanmar's employees is also due to the limited number of employees. Figure  1 illustrates the comparison of normal working hours and overtime in January-March 2022. Employees must work overtime to meet and exceed work completion targets.

Figure 1. Normal Working Hours and Overtime Working Hours
Source: PT Yanmar employee overtime hours data Figure 1 above shows that the average overtime hours for production department employees in January-March 2022 is 65,288 hours.
Implementation of work also requires supervision. The key to successful human resource management is effective supervision. Supervision carried out by superiors/supervision/team leaders is not only to find fault with employees, but to ensure that work progress is in accordance with the agenda that has been determined by focusing and supervising subordinate employees (Gibson, 1996). Employees are more likely to make mistakes during the work process when there is not adequate supervision, or delay the completion of work. Employees are free to work whenever they like as a result of the absence of strict sanctions from the company and an ineffective work monitoring system. Employees experiencing stress at work may experience physical and mental symptoms such as fatigue, lack of sleep, impatience, headaches, insomnia, and being unable to show up for work until a potential accident occurs. The lack of human awareness of one's own limitations is basically the root cause of all forms of stress. The main forms of stress-frustration, conflict, anxiety, and guilt-will occur if you are unable to overcome these limitations (Luthans, 2012). Employee performance can be negatively affected by work stress, which can also result in decreased performance (Fonkeng, 2018).
Several previous studies presented in this study aim to be a comparison that this research has gaps or research differences. First, Rusda Irawati and Dini Arimbi Carollina (2017) explain research findings that internal workload has a positive impact on employee performance. The results showed that the performance of operator employees in the PT Molding Department was positively impactd by internal workload. Precision Giken Indonesia. Second, Rizky Aujada, Saryadi and Hari Susanta Nugraha (2015), also explained the results of the study, that there is a correlation between workload and work stress and has a positive impact and has significance on the stress variable, and there is a positive correlation between supervision and work stress. Subsequent research, conducted by Nurtika Sari, Darwin Lie, Efendi and Ady Inrawan (2015), Based on the results of the study, it can be seen that the performance of employees at the Pematangsiantar Auction Center and the State Property Service Office has increased significantly through supervision. Furthermore, Andri Wijaya (2018) explains the results of the study that job satisfaction is negatively affected by workload; when workload goes up, job satisfaction goes down, and vice versa. Job satisfaction is negatively affected by job stress. When jobrelated stress increases, job satisfaction decreases, and vice versa.
Therefore researchers are interested in doing this because the right distribution of workload in accordance with the abilities of employees and effective supervision can affect work stress and employee performance in the production section with the title "ANALYSIS OF  (2011) workload or workload is the result of interaction between task demands, workplace environment, worker skills, behavior, and perceptions of the workplace. According to Hancock & Meshkati in Zuraida (2020), states that working too much causes physical symptoms such as abdominal pain, difficulty sleeping, headaches, loss of appetite, and euphoria, etc. There are also mental symptoms such as forgetfulness, irritability, hopelessness, and difficulty concentrating. Finally, social behavior symptoms such as withdrawing, avoiding, smoking a lot, and drinking alcohol.
According to (Harry in Abang, 2018), the dimensions of the workload or the size of the workload, include: a) Load period, which shows how much time is available in organizing, carrying out and checking tasks or work. b) The mental effort load, or the amount of mental effort required to complete a task. c) Psychological stress load, which indicates the level of uncertainty, frustration, and job risk. On the other hand, there are four workload indicators according to Putra (2012), namely: a. Working Conditions Working conditions, such as making quick decisions when working on items and dealing with unforeseen events such as working later than necessary b. Standard and Quantity of Work The impression that an individual has about his work, for example the feelings that arise regarding the workload that must be completed within a certain period of time c. Use of Time Working Time, also known as standard or baseline time, is used for activities directly related to production). d. Targets to be achieved Individual views regarding the size of the work target given to complete the job. A view of the results of work must be completed within a certain period of time.

Supervision
Supervision according to Yahya (2006), can be defined as a procedure to ensure that management and organizational goals are met. This relates to the activity planning method. This understanding shows the close correlation between control and planning. In functional management, each leader from each unit or work unit is responsible for controlling or supervising employees who carry out their duties in accordance with their main responsibilities.
Supervisors must understand several aspects of supervision in order to function effectively. Handoko (2012:57) says there are 5 elements of supervision, namely: a. Setting standards for desired results; B. Determine how activities are carried out; C. Measuring how activities are carried out; D. Comparing standard implementation and deviation analysis; e. Take corrective action when necessary Robbins and Coulter (2012) say that there are four types of monitoring indicators: b. Setting guidelines (Principles, specifically setting benchmarks (goals) or desired outcomes, to be completed as a result check when hierarchical exercises occur. c. Measurements or procedures that are carried out repeatedly, continuously, and accurately in intensity in the form of daily, weekly, or monthly measurements so that the correlation between quality and number of results can be observed.
d. Comparing (Compare) means comparing the results with the goals or standards that have been set; it could mean the performance is higher, lower, or equal to the standard. The decision to take corrective action is called taking action. It is necessary to follow up in the form of corrections to deviations that occur when there is a deviation (deviation) between the standard and its realization. Manullang claims (Winardi, 2000: 224), there are four ways to collect information using monitoring techniques: personal review, oral or written interviews, reports, and special project supervision.

Employee performance
Work results that can be achieved by individuals or groups within an organization in accordance with their respective authorities and responsibilities, or how a person is expected to function and behave in accordance with the assigned tasks, are all examples of performance. It also means how unique a task is, how good it is, and how long it will take to complete it (Sutrisno, 2011).
The indicators of this study include the opinion of Robbins (2007) that the following aspects can be used to measure the performance of individual employees: a. Quality The nature of work is estimated by representative views of the nature of the jobs created and the completeness of the tasks on the abilities and capacities of workers. b. The quantity produced is referred to as the quantity expressed in units or activity cycles completed. c. Timeliness is measured by how well the activities are coordinated with the output results and how much time is left for other activities. d. Survival is how much various resource objectives (energy, cash, development, raw parts) are fully determined with the ultimate goal of growing each unit's symptoms in resource usage. Work commitment is the extent to which employees have work commitments with outside organizations and office responsibilities. e. Independence Commitment is the extent to which an employee can carry out his job responsibilities.

Work Stress
According to Triatna (2015), states that a person is in a state of stress when internal or external disturbances affect his physical and/or psychological condition, causing tension and unusual behavior (known as deviant behavior) at the physical, social, and psychological levels. According to Luthans (2006) there are several organizational stressors which are factors that come from within the organization and contribute to work-related stress. The company's strategy to compete with other businesses sometimes shifts as a result of internal changes.
The company's intervention then had several consequences, including: a. Employees are restless at work due to policies implemented by leaders who control their subordinates too much. b. Hazy Errands, the organization may have the option of assigning responsibility to representatives because of high organizational demands, this could be the fault of representatives for the effort derived by the organization which ultimately results in work pressure. c. Robbins and Coulter (2012) identified three dimensions of stress at work: d. Physical Signs and Symptoms Particularly in the field of medical research and health sciences, the initial symptoms that can be observed are physiological symptoms. Stress will generally cause changes in the body's digestion, increased heart rate and respiration, increased circulatory pressure, brain pain, and, surprisingly, more serious cardiovascular failure. e. Symptoms psychologically, dissatisfaction can be caused by stress. This is the most obvious and immediate psychological effect. On the other hand, other mental states such as tension, anxiety, irritability, boredom and procrastination can occur. Employees' levels of stress and dissatisfaction can impact their jobs if they are placed in positions that have multiple demands that conflict with one another or where the duties, powers and responsibilities of the incumbent are unclear. f. Indications of Behavioral Changes Eating habits, smoking and alcohol consumption, fast speech, restlessness, and sleep disturbances are symptoms of stress behavior. Other symptoms of behavioral stress include changes in productivity, absenteeism, and employee turnover rates.

RESEARCH METHODS
Because the data in this study comes from the results of events that have occurred, the researcher only reveals facts by measuring the symptoms of respondents. This research uses quantitative research methods (Suharsimi, 2010). Quantitative research is a research method based on the philosophy of positivism, which is used to research certain populations or samples (Sugiyono, 2017). This study aims to find the impact of the independent variables Workload Effect and Supervision on the dependent variable, namely Employee Performance through Job Stress as an intervening variable.
The populations in this study were 336 creative office representatives. An overview of the distribution of employees in job positions in the production section is done to make it easier for researchers to take samples as will be shown in this section to determine the population in table 1. The distribution of employees in the production department according to job title is shown in Table 1 above. The next step is sampling. A group of people, events, or other things with certain characteristics being investigated is called a population (Sugiyono, 2004). The distribution of samples for each field is as follows:  Table 2, the sampling was based on job title, namely staff 13 respondents, foreman 24 respondents, assistant foreman 33 respondents, and operators 266 respondents so that the total number of respondents taken as a sample was 336 respondents. The sample is part of the number and characteristics possessed by the population (Sugiyono, 2017).

Data
Information sorting methods should be possible through observation, interviews, or documentation. This study utilizes primary data types and Google form questionnaire data sources to select samples based on criteria. The Linkert scale was used to distribute the research questionnaire, which includes a number of statement items and questions. Respondents will respond to these statements and questions with closed answers-not essay answers-ranging from "strongly disagree" to "strongly agree". The scale used is 5 (odd) considering that several previous examinations often used this 5 scale, and during the use of 5 scales no significant problems were found. On an odd scale, category number three (neutral) or the middle is an option. The middle category was created to make it easier for respondents with moderate traits to participate.

Data analysis method
The SmartPLS program is used as a general approach for data analysis, and the Structural Equation Model with Partial Least Square (PLS-SEM) is used as a data analysis method in this study. The SmartPLS program is used in the Structural Equation Model, also known as component-based SEM or structural equation analysis.
Comprehensively the stages of the analytical method in this study consist of descriptive statistics and testing the research hypothesis as follows: 1. Descriptive statistics In research, descriptive statistics are used to describe and describe variables. Standard deviation, maximum-minimum, variance, and average value are examples of descriptive statistics that can provide an overview of the data set (Ghozali, 2017).

Partial Least Square -Structure Equation Modelling
The information handling strategy in this study uses the full underlying model with the Halfway Least Square (PLS) technique. Basically, World developed PLS to test weak theory and weak data, such as problems of data normality or small sample sizes. a. Outer Model In the evaluation section of the outer model, construct validity and reliability tests will be carried out. According to Ghozali (2011), the validity of the questionnaire can be evaluated using a validity test. The legitimacy test is estimated by concurrent legitimacy, discriminant legitimacy, and separation of normal change (AVE). While the results of the SmartPLS algorithm are used to measure the reliability test, the value of composite reliability and Cronbach's alpha 1) Convergent Validity The loading factor on latent variables and their indicators is the value of convergent validity. (Ghozali, 2011) The expected value is higher than 0.7 However, loading values between 0.50 and 0.60 are still acceptable in the scale development stage research (Ghozali and Latan 2015).

2) Discriminant Validity
In this section is the value of the cross loading factor, which can be used to determine whether a construct has sufficient discriminant by comparing the loading value of the construct in question with other constructs. In the PLS Algorithm report, the value of the cross loading factor can be seen by selecting discriminant validity and then cross loading.

3) Composite Reliability
Choose Matrix after selecting composite reliability in the PLS Algorithm report. For confirmatory research, a variable is considered reliable if its Cronbach alpha value is less than 0.7, while a Cronbach alpha value between 0.6 and 0.7 is still acceptable for exploratory research (Ghozali, 2011). b. Inner Model 1) R-Square The R2 value is used to measure how much the endogenous factors are impactd by different factors. The R2 value is 0.67 (strong), 0.33 (moderate) and 0.19 (weak) (Chin, 1998).

2) Q-Square
Using predictive-relevance measures, model fit in PLS analysis can also be evaluated by examining the percentage of variance explained by the R2 values for endogenous latent variables. (Q2). Its value can be calculated using the following formula: Path Coefficient Evaluation of the path coefficient is used to show how much impact the exogenous variables have on the endogenous variables. P-value ≤ α, it is said to have significance (Ghozali, 2013).

RESEARCH RESULT
The bootstrapping results in this study can be seen in the following figure: For more details, the results of path analysis using bootstrapping with the help of SmartPLS 3.2.9 software can be seen in the following table: The results of the path coefficient test in table 4.13 above, the original sample value of the workload policy (WL) on employee performance (KK) is -0.751 with significance of 0.000. These results mean that the workload has a negative impact and has significance on employee performance. Therefore, Ha1 is accepted.
The results of the path coefficient test in table 4.13 show that the original sample supervision (PS) value on employee performance (KK) is -0.101 with a significance level of 0.002. These results mean that supervision has a negative impact and has significance on employee performance. Therefore, Ha2 is accepted. The results of the path coefficient test in table 4.13 above, the original sample value of the workload (WL) policy on work stress (SK) is 0.587 with a significance of 0.000. These results mean that workload has a positive impact and has significance on work stress. Therefore, Ha3 is accepted.
The results of the path coefficient test in table 4.13 above, the original sample value of supervision policy (PS) on work stress (SK) is 0.289 with a significance of 0.000. These results mean that workload has a positive impact and has significance on work stress. Therefore, Ha4 is accepted. The results of the path coefficient test in table 4.13 above, the original sample workload (WL) value on employee performance (KK) mediated by work stress (SK) is -0.080 with a significance level of 0.000.
These results mean that workload has a negative impact and has a significance on employee performance mediated by work stress. Therefore, H5 is accepted.
The results of the path coefficient test in table 4.13 above, obtained the original sample supervision (WL) value on employee performance (KK) mediated by work stress (SK) of -0.039 with a significance level of 0.000. This result means that supervision has a negative impact and gives a negative impact. The significance of employee performance is mediated by work stress. Therefore, Ha6 is accepted. The results of the path coefficient test in table 4.13 show that the original sample work stress (SK) on employee performance (KK) is -0.136 with a significance level of 0.000. These results mean that work stress has a negative impact and has significance on employee performance. Therefore, Ha7 is accepted.

Effect of Workload on Employee Performance
Based on the research results, it is explained that a high workload will reduce employee performance, so it can be said that workload has a negative impact on employees. This research supports the research conducted by Putri and Rahyuda (2019) which states that workload has a negative impact and has significance on performance.

The Effect of Supervision on Employee Performance
Based on the research results, it is explained that high supervision will reduce employee performance, so it can be said that supervision has a negative impact on employees. This research supports the research conducted by Putri and Rahyuda (2019) which states that supervision has a negative impact and has significance on performance.

Effect of Workload on Work Stress
Based on the results of the study, it was explained that a high workload will increase employee work stress, so it can be said that workload has a positive impact on work stress. This research supports research conducted by Sugiarti (2018) which states that workload has a positive impact and has significance on stress.

Effect of Supervision on Work Stress
Based on the research results, it is explained that high supervision will increase employee work stress, so it can be said that supervision has a positive impact on work stress. This research supports research conducted by Putri and Rahyuda (2019) which states that supervision has a positive impact and has significance on stress. 5. The effect of workload on employee performance is mediated by work stress Based on the results of the study, it was explained that workload will reduce employee performance mediated by work stress, so that it can be said that work stress mediates the negative impact of workload on employees. This research supports research conducted by Putri and Rahyuda (2019) which states that supervision has a negative impact and has significance on performance mediated by work stress.

The Effect of Supervision on Employee Performance is Mediated by Work Stress
Based on the results of the study, it was explained that supervision will reduce employee performance mediated by work stress, so that it can be said that work stress mediates the negative impact of supervision on employees. This research supports the research conducted by Putri and Rahyuda (2019) which states that supervision has a negative impact and has significance on performance mediated by work stress.

Effect of Work Stress on Employee Performance
Based on the research results, it is explained that high work stress will reduce employee performance, so it can be said that work stress has a negative impact on employees. This research supports the research conducted by Putri and Rahyuda (2019) which states that work stress has a negative impact and has significance on performance.

CONCLUSION
This study draws conclusions and discusses the analysis of the effect of workload and supervision on employee work performance with work stress as an intervening variable in the