,

Specifically, this study aims to determine how much influence the recruitment of extrinsic reward and training on employee performance at PT. Bakrie Metal Industries Bekasi simultaneously. The population is 344 employees and a sample of 172 employees is taken. The writer uses quantitative methods using structural equation modeling (SEM) with the help of Lisrel 8.8 software (t-value > 1.96) and primary data was obtained from questionnaires whose measurements used a Likert scale that was tested for validity and reliability. The results of the study based on the validity and reliability test showed that the data was valid and reliable. Based on the t test and F test shows that recruitment, extrinsic rewards, and training have positive and significant influence on the performance of employees at PT. Bakrie Metal Industries Bekasi. The job description plays an important role for the recruitment process. The training participated by employees must be interesting so that the objectives and training material can be absorbed by the participants, and each of the recruitment variables, extrinsic rewards, and training had a positive and significant effect on employee performance.


I. INTRODUCTION
In the last 5 (five) years, the government has been very focused on infrastructure development, especially the construction of roads, both national roads and highways. One of the products of PT. Bakrie Metal Industries is needed to support government programs, especially the construction of road and bridge infrastructure. Infrastructure development in Indonesia is now prioritized by state-owned contractors and domestic private contracting companies. PT. Bakrie Metal Industries applies work hours at 07.30 WIB, but in reality, there are still employees who arrive late from the appointed hours. The discipline of employees decreased, as seen from employee absenteeism in 2019 is 7% and there was 16% of lost time due to the late attendance of employees. This phenomenon is very detrimental to PT. Bakrie Metal Industries, which includes many factors that cause a decline in employee performance.
Based on this, the study will look at: (1) whether there is a significant influence of recruitment on employee performance; (2) whether there is a significant effect of extrinsic rewards on employee performance; and (3) whether there is a significant influence training on employee performance at PT. Bakrie Metal Industries, Bekasi.

Employee Performance
According to Hasibuan (2012), performance is the result of work obtained by a person in carrying out given tasks that are based on skills, experience, and sincerity, as well as time. Siagian (2015) discusses the requirements needed to determine company performance in accordance with standards set by the organization for companies that match their job descriptions. Performance is the result of an employee's actions according to his role in the company or the responsibilities he carries. There are six indicators proposed by Robbins (2015), as follows: (1) quality; (2) quantity; (3) punctuality; (4) effectiveness; (5) independence; also (6) work commitments.

Recruitment
In recruitment, the recruitment indicators according to Mathis and Jackson (2012) are (1) basic recruitment rules; (2) recruitment sources; and (3) recruitment methods. According to Marwansyah (2012) recruitment is the process of getting qualified human resources (employees) to take a position or job in a company. Gaol (2014) opines the recruitment of its nature is a process of determining and attracting applicants who are able to work in a company. The starting process starts when applicants are sought and ends after their applications are accepted or collected. So, recruitment is a collection of activities that are used to get a collection of qualified job applicants.

Reward
According to Fahmi (2016), reward or what we often call compensation is a form of remuneration given to an employee for work performance, both financial (nonfinancial) and non-financial. Meanwhile, according to Bohlander (2013) there are two types of rewards that can be given to employees, namely intrinsic reward and extrinsic reward. Intrinsic Rewards are awards that are not visible (intangible) and are sourced from the person himself, for example status, job satisfaction, and informal awards. For extrinsic rewards are awards that are clearly visible in physical form (tangible) and given by others for an achievement, for example, certificates, job promotions, social relations, additional bonuses, and incentive payments. According to  indicators of rewards are as follows: (1) salary; (2) wages; (3) incentives; (4) allowances; and (5) awards.

Training
Harsono (2011) states that education and training can be interpreted as a teaching and learning process by using certain techniques or methods, in order to improve skills in handling tasks and functions systematically and in an organized manner. Meanwhile, Dessler's (2015), states that training is the process of teaching new employees or employees who have basic skills to carry out their work. Training is one way to improve human resources in the world of work as work demands change due to changes in the work environment, strategies, and so on. The training indicators according to Mangkunegara (2013) include (1) training objectives; (2) material; (3) method; (4) training methods; (5) participant qualifications; and (6) qualification of trainers (instructors).
According to Kerlinger (2014) a research design is a plan and structure (model/paradigm) of inquiry arranged in such a way as to obtain answers to research questions. Based on the above theoretical basis, the following research framework is developed: Hypothesis 1: there is an influence between recruitment (REK) and employee performance (KIN). Hypothesis 2: there is an influence between extrinsic rewards (RE) and employee performance (KIN). Hypothesis 3: there is an influence between training (DIK) and employee performance (KIN).

III. RESEARCH METHODOLOGY
This research was conducted at PT. Bakrie Metal Industries in Harapan Jaya district, Bekasi city, west Java for 5 months. The data used are primary data of the population is 344 employees and a sample of 172 employees is taken. The sampling method used was purposive sampling, where the sample was chosen by researchers from 9 existing departments (ppc, quality control, HSE, production, HRD, maintenance, engineering, project management, and finance).
This study consists of 4 research variables, where 3 between variables are exogenous latent variables (REK, RE and DIK), and 1 endogenous variable (KIN). Exogenous latent variables are independent variables that influence the dependent variable. In the SEM model, the exogenous latent variable is shown by the arrows coming from the variable towards the endogenous latent variable. Endogenous latent variables are dependent variables that are influenced by independent variables, and are indicated by arrows that lead to these variables.

IV. RESULTS AND DISCUSSIONS
Equations with SEM processed by Lisrel will meet the following test rules (Wijanto, 2008): (1) model match test; (2) validity test; and (3) reliability test with conditions according to the discussion in chapter III. The results of these tests in this study are good, starting from the suitability of the model, the validity test, to the reliability test. A detailed description of the results of data processing using Lisrel can be seen in the explanation below:

V. CONCLUSION
The results of testing with SEM software are as follows:

Recruitment Variable
Based on the theoretical framework built, the recruitment variable has 7 (seven) indicators, namely: the media used, applicant qualifications, job descriptions, education, expertise, work experience, and recruitment goals. After passing the factor load confirmation test on the 7 (seven) recruitment indicators, it turns out that there are 3 (three) invalid indicators. The step taken is removing those 3 (three) indicators from the model, while the other 4 (four) indicators are maintained because they meet the requirements. From this result, the factor confirmation analysis of recruitment yields 4 (four) indicators The highest factor load is "clear job description of applicants" then followed by the indicator "applicant qualifications according to expertise".
From the results of these respondents, it indicates that the job description plays an important role for the recruitment process to get candidates, and then the qualifications of the applicant's expertise in accordance with the vacancies needed.

Extrinsic Rewards Variable
Second, extrinsic reward variables are measured based on 7 (seven) indicators, namely: salary received, incentives, bonuses, promotions, company awards, appreciation, and allowances. Testing of the load of the seven factors in the extrinsic reward sub-factor turned out to only produce 5 (five) significant indicators and 2 (two) invalid indicators, so that they were excluded. The highest factor load for the extrinsic reward variable is "bonus giving aimed at improving performance", while "promotion" has the smallest factor load for the extrinsic reward variable.
Based on Lisrel data processing, positive and significant number data are obtained. This indicates that the better the rewards in the company, especially bonuses, the employee's performance will also increase. Conversely, the better the reward at the company, the employee performance will also decrease.

Training Variable
Training variables are measured based on 8 (eight) indicators, namely: training according to the target, training materials according to the needs of the job, trainers provide good training, interesting training methods, professional trainers, participants according to training criteria, participants improve their skills, and evaluate training according to the target. Testing of the eight-factor load in the training indicator were all accepted. The highest factor load for the training variable is "the training methods provided are quite interesting. The training process at PT. Bakrie Metal Industries received a response from good employees. The researcher concludes that the training participated by employees must be interesting so that the objectives and training material can be absorbed by the participants. While the lowest result is material that is suitable for the needs of the job, here is an input that employee responses want training to be practical and directly applicable to the job or more applicable.

Employee Performance
Employee performance variables are measured based on 8 (eight) indicators of observable variables, there are 7 (seven) indicators that are accepted/valid and 1 (one) indicator that is discarded because it is invalid. The highest indicator is "I am able to complete the work according to the time assigned". This means that respondents feel able to complete the work given by the company in accordance with the requested time. This is related to the characteristics of work at PT. Bakrie Metal Industries, as a contracting company that must complete work in accordance with the agreed contract. T-Test results concluded that each of the recruitment variables, extrinsic rewards, and training had a positive and significant effect on employee performance.