Supporting JIT Production with the Best Wage System

Michel Baudin

Draft 2 ó

This article was published in IIE Solutions in February, 1996. It is copyrighted material, and may not be used for any purposes other than review, without the written authorization of Michel Baudin.

Abstract

Conversion from mass to lean production must be accompanied by changes in the wage system for companies to send the right messages to their work forces. Straight time payment, measured day work, or piece rates offer no reward for flexibility or teamwork. The practitioners of JIT/lean production in Japan have developed wage systems that compensate operators based on the numbers of different jobs they can perform, and where their abilities are measured against both technical and people skills. The approach includes group productivity incentives that are just large enough to command the operatorsí attention without turning them into ìbounty hunters.î To operate this system, the human resources department be 60% larger than the current US average, and management must allocate the time needed to give operators the kind of periodic reviews that are traditionally reserved for professional staff. This is, however, necessary not only to support JIT/lean production today, but to prepare for future factories with few, highly-skilled employees.

Current wage systems ignore teams and multifunction operators

As half the automobile parts companies in the U.S. have programs to convert to production in cells with teams of multifunction operators, it is time to rethink the wage system -- without exaggerating its influence -- to reflect the new contributions companies expect of their shop-floor personnel. This is an area that has been given short shrift in the published literature, but is addressed by practitioners of JIT/lean production in Japan. The need for such an effort is readily understood by observing behavior on the shop-floor:

• The operator slams down the safety latch of the elevator on the side of his machine. It wonít start. He slams the safety latch harder, and curses. The elevator eventually starts, but the operator is frustrated and angry: he is paid on a piece rate, but this defective machine sets his speed.
• In another plant, an operator passes a defective component to her downstream colleague for assembly. It doesnít fit and he rejects it, but she continues producing more defective components and he continues to reject them until the pile of rejects fills up the space between them. The operator is paid by the hour, with a ìreasonable expectancyî of production, and an individual bonus for exceeding it. In no way is her performance evaluation linked to that of her colleague.
• In yet another plant, management is trying to implement cells, which require operators to rotate between jobs. But two operators who have been running a lathe and a grinder for 15 years are not inclined to learn each otherís trade. Among other reasons, they are in two of the plantís 70 different job categories, directly related to tasks, and with differences in pay. In the absence of any changes to that system, the operators are not convinced the company truly wants them to change.

The wage system does not and cannot drive the change from job-shops to flow lines, and from mass to lean production, but it must accompany it. If incentives are nonexistent or inconsistent with the new objectives, the operators will notice it and conclude that management is uncommitted.

The wage system as a communication tool

Our starting point is that the wage system is a means of communicating company expectations. The classical IE point of view on this subject is that the wage system must assure "equitable compensation," which is not easily translated into practice. In reality, the amount that can be paid to shop-floor personnel is dictated by the labor market. The size of the pie is fixed; the wage system can only influence how it is split. We will examine how the company can adjust the distribution of this pie so as to support its strategic objectives for the work force.

If what the company expects of an employee is simple, then the wage system can be simple too, as in the following examples:

• Pay for performance. Shoveling dirt can be paid strictly on the basis of the quantity of dirt moved. The operator controls how fast the work is done, and cannot influence quality. On a manufacturing shop floor, similar conditions are rare, and so are piece rates. Germany is the only major industrialized country where piece rates are still common, albeit in a regulated form.
• Pay for time. A night watchman deters intruders by his presence. He does not have to do anything but he must be there. His value to the company is proportional to the time he spends on the job, and therefore can be paid by the hour. Most American factory operators are paid by the hour, even though what they do matters.
• Pay for ability. Firefighters and equipment repair technicians are paid for their abilities and availability. What matters to the company is not how much they do on a daily basis but what they can do when the need arises. They are paid for their potential, not for their performance. More and more, shop floor operators are minding automatic machines, and their ability to respond to malfunctions becomes a more important part of their contribution. As a consequence, ability should be a factor in the wage system.

Pure piece rates or pure hourly rates are in fact not practical. The logic of a pure piece rate would, for example, preclude any paid time off. Given that Germany is the third largest economy in the world and that many management decisions in factories must be approved by "works councils" including elected labor representatives, one wonders how its industry could rely on such a throwback to Taylorism as piece rates. Looking closer, one finds that the German piece rate system is full of compromises and adjustments. The time standards are set not by employers but by a neutral organization called REFA, whose time studies are endorsed by unions, and these standards are loose enough that all workers can beat them every day.

Conversely, hourly rates almost always come in the form of ìmeasured day workî. They are always linked to "reasonable expectations" of productivity, set in much the same way as standards for piece rates. Companies sometimes pay bonuses for exceeding the standard, but these incentives are frequently offset by union agreements stipulating payments of "productivity bonuses" to everybody, regardless of performance.

However, even with these implementation adjustments, the two systems remain different in the messages they convey about the company's attitude to its operators. If, in a piece rate system, you produce too little, you make less money. In an hourly rate system, your pay is not immediately affected but your appraisal is. In the piece rate system, the company treats you like an outside contractor; in the hourly rate system, like a member of an organization.

In the hourly rate system, the operator is entitled to his wages on a regular basis as long as he is a member of the organization. His wages change based on a yearly or biyearly review schedule, but are not adjusted at every pay period, except for tardiness or overtime. Good performance results in promotions or higher than average raises at review time; bad performance, in remedial training, job transfers, or eventual dismissal , but not in reduced wages.

As part of the set of industrial engineering techniques now called ìJIT/lean production,î the Japanese automobile industry developed an elaborate approach to personnel administration and wages, involving base hourly rates, productivity incentives, payment for ability, and career planning for operators.

A wage system for JIT/lean production

Grade and ability rating

In cells, operator jobs are designed so that operators move in a loop to attend to several machines performing different operations. In this sense, they are all multifunctional. However, as the demand structure varies, so does the speed of cells, and they may either move to other jobs in other cells or see their set of machines expanded or shrunk. Clearly, the most valuable operators are those who can run the most different machines, not just within a cell but within the whole department. Many Japanese companies use a two-dimensional classification of operators by "grade" and "ability."

Grades are a formal ladder with rungs corresponding to defined levels of skills, giving access to specific training and promotion opportunities. The number of operators in each grade, however, is limited, and operators who satisfy all the requirements can only be promoted when there is an opening. Following is a possible grade structure:

• Rookie. Operator enter the company at this grade, and are rotated systematically between jobs for three years.
• Standard operator. Once the rookies have mastered half the jobs in the department, they are eligible for promotion to the next level.
• Flexible operator. Five to six years after joining the company, an operator should have mastered most of the jobs in the department, and be eligible for promotion to this level.
• High-level operator. The high-level operators are few in number, can perform and improve all the jobs in their department and train others, and can be promoted to first-line management.

The "ability" dimension is used to pay higher wages to operators in lower grades who are qualified for higher grades but are held back by lack of openings, or who have developed a deep knowledge of few skills instead of the operational knowledge of many skills needed for promotion. For each operator, the company measures two types of abilities:

• Technical ability. This measurement is posted on the shop floor for the whole department in a skills matrix (See Figure 1). Many U.S. companies who have begun implementing JIT/lean production have generated some form of skills matrices, but few have used them in the wage system yet.
• Managerial ability. The is the ability to work with, train and lead others, as assessed by superiors twice a year through the formal review process.

The skills matrix shows the technical proficiency of each operator for each task. As this information is on public display and is the justification for paying some operators more than others, its presentation is a sensitive matter. The following ways are use at different companies:

• A certification check mark, with possibly the year of certification attached. This matrix would say, for example, that operator Singh has been certified able to operate lathe LAT-06 since 4/94. It is simple, and easily accepted by operators, but the information is limited to ìgo/no-go,î which is insufficient.
• The cumulative time spent on the task. This matrix would say that operator Singh has run lathe LAT-06 for 853 hours. This is an objective measurement, requiring constant updates, but conveying different levels of proficiency in a non-controversial manner.
• Symbols formally representing levels of proficiency. In Japan, the symbols "," ",""," and "" are commonly used to represent ordered categories. American operators may be more comfortable with symbols that are more similar to those used to rate products in Consumer Reports, such as:

Cannot perform the task

Can perform the task at half speed under supervision.

Can perform the task at full speed without supervision.

 Can train or supervise others performing the task.

Has contributed to process improvements.

In this case, a formal certification is necessary for an operator to move from one level to the next.

In the US, managerial ability is only evaluated for professional staff, and periodic performance reviews are considered a key management function. The biyearly reviews of shop-floor operators are more elaborate as to performance, training and career planning than those given to engineers and managers in many Fortune 500 companies. American shop floor operators are not used to this kind of attention, nor to subjective assessment by their supervisors. Even in Japan, manufacturing companies have found that it could not be done by the direct supervisor alone. Each operator is evaluated by a panel of reviewers, and, in some cases, has the opportunity to select one of them to be his advocate.

Figure 1. Sample skills matrix

---

Wage components

Each individual's salary is the sum of the following components:

• Basic wage. The basic wage unrelated to performance. It is intended to provide a minimum living wage, and is based on government statistics of the local cost of living by age group.
• Job wage. This portion is related to the grade, to seniority within the grade, and to the type of assignment ó that is, it incorporates shift differentials and additional payments for particularly tiring or unpleasant jobs.
• Ability wage. The ability wage is based on the technical ability shown in the skills matrix and the managerial ability as evaluated in the review process. It provides an alternative way to have higher wages for those who, for whatever reason, cannot be promoted to higher grades.

The breakdown among the three components varies between companies. Both the job and the ability wages reward flexibility. The company values operators who can take one any one of a dozen different tasks more than one who can only do one, and it pays accordingly. Within these parameters, the operators are paid for time, but there are productivity incentives.

The company pays operators to be available during work hours, regardless of whether the company makes effective use of that time. The operators then are paid by the hour for overtime, and penalized by the hour for tardiness. Where operators are organized in cell teams , the tardiness penalties are more than proportional to time lost, because an individual who arrives late wastes time for the whole team.

There is one aspect of the Japanese wage system that we will not discuss here: it is the standard bonus system through which employees receive the equivalent of several monthsí paychecks twice a year. It is part of the country's business culture, not of JIT/lean production. There is no particular reason to emulate it. On the other hand, the notion of a productivity improvement bonus awarded to a team of operators requires careful consideration.

American managers today either do not believe in incentives, or believe in them to excess. Most companies have no way to reward outstanding operator performance. Others offer cars and cruises as prizes for individuals. In reality, incentives work best when they are just large enough to convey a message, and when their psychological impact is strengthened by symbolic gestures such as plaques, trophies, award ceremonies, or newsletter articles, tailored to the specific social and cultural environment of the plant.

A 3% bonus over base pay is enough to get the operator's attention when he is awarded it, while missing it will not jeopardize his livelihood; a 30% bonus -- with a correspondingly lower base pay -- will turn the same operator into a bounty hunter. Because operators are not driven exclusively by greed, the combination of small bonuses with symbolic rewards can durably influence their attitudes. When incentives represent a substantial part of operator incomes, (1) the behavior they induce is not always that intended, and (2) the effects of the incentives stop as soon as they are withdrawn. Team bonuses induce competition, but this competition should remain friendly. As in a poker game, some financial stake is necessary to keep the players focused, but the friendly atmosphere can only be sustained if the amounts in play are small with respect to each player's livelihood.

The key points of this type of team bonus are as follows:

1. The productivity improvement bonus rewards a team of operators for having increased a line's productivity through changes in methods and processes. It is not a reward for working harder. Once a productivity improvement bonus is paid, the standard is updated to the new level. What was 120% of standard in the prior period becomes 100% and no longer rates a bonus. A team that would achieve 120% in one period by extra effort would commit itself to this new level of effort indefinitely for no extra pay. On the other hand, if it achieves 120% by linking machines, reducing changeover times, or implementing mistake-proofing mechanisms, it can get another bonus in the following period by coming up with new ideas. The accumulated productivity improvements of the past are built into the wages.
2. Whereas the wages are determined separately for each individual, the productivity improvement bonus is the same for all team members. This is to communicate the value the company places on teamwork. It encourages the team to discipline itself and to assist its weaker members.
3. The bonus should be a "staircase function" of the productivity improvement. In Figure 2a., the bonus is simply proportional to performance over 100%, which is not as stimulating as a small number of discrete productivity goals, triggering different levels of bonus, as shown in Figure 2b.



Figure 2. Proportional versus staircase function

4. Productivity is calculated as follows:

• The quantities produced for different products are measured against a common yardstick, to take into account the varying time requirements of the different products at each operation. Tons per operator per month is a common productivity measure in the casting industry. In machining, each part is compared to a reference standard part based on cycle times, and productivity is expressed in "equivalent standard parts" per operator per month or per shift.
• Management should not use a measure of productivity that encourages production regardless of need. Productivity for the group of operators should be calculated from the time required to produce the planned quantity:



where the standard time is calculated from the demand and the process times of the various operations, which must be measured.

• Once quick changeovers are implemented, they are carried out by operators, and will be considered part of the operators' productive time. The productivity of changeover work will be calculated as the ratio of the standard time to the measured, actual time.
• The records necessary for productivity measurements will be kept by the team leaders. While they may be tempted to underreport process times in order to increase productivity improvement bonuses, several factors hold them back:
• Management will attempt to make use of the time reported as "available." Therefore, if a group underreports the amount of time used for production, this may be used as a reason to reduce its number.
• Tower lights and other means of making the shop floor visible, allow management to see when equipment is being used.
• The early completion of a shiftÕs work load can be reported to the foreman in the green corner to enhance visibility.
• As the productivity improvement bonus is modest with respect to total compensation, it is not a strong incentive to cheat.






Figure 3. Year-end changes

1. In a line organization that is broken down into several teams, the productivity improvement bonus can be broken down into two components, with 40% of the bonus may be based on line performance, and 60% on team performance.

Both the salary and the performance basis for productivity bonuses are revised once per year, according to the company's formal review schedule. To maintain employee morale, the company must make sure that employees do not see their income drop. At the change of period, what constituted excellent performance in the previous year becomes the new standard, and no longer justifies a productivity bonus. This must be taken into consideration when deciding the salary raise and the new standard, as shown in Figure 3.

In Figure 2, at year's end, the average productivity of the previous year is made the new standard. Because it is the average for the whole year, the operators have a good chance of exceeding it right away. Furthermore, the assumed 4% raise on the basis of an 18 $/hour salary brings it up to 18.72 $/hour, and the individual sees his income go up, even if he does not receive a productivity bonus right away.

Conclusions

Applying to shop floor operators the human resource management techniques and wage systems that have been traditionally reserved for engineers and managers is part of a global effort to professionalize the manufacturing work force. Early this century, the trend was towards simpler worker tasks, requiring less and less skills. This trend is now reversed, and the future we can now envision has factories with few, highly skilled employees, whose primary tasks are programming and maintaining automatic, computer-controlled equipment.

Clearly, elaborate reviews of 2,000 shop floor operators twice a year are a perceptible added workload for plant management, but one that companies like Honda or Toyota have found worth taking on. US plants average 1 personnel employee for each 74 employees. At Toyota, the ratio is 1 personnel employee for 47 employees, which is 60% higher, and, in addition, managers in other departments have to allocate time to sit on review panels and participate in award ceremonies.

The field of personnel and wage administration in Japan is also unusual in that it is a career specialty. While other technical and managerial positions are occupied by generalists who rotate among functional areas within the company, personnel administrators always remain in that area, and share ideas with colleagues at other companies.

The implementation of a JIT/lean production wage system requires time and experimentation on a case-by-case basis. What is certain, is that it is an issue that must be addressed along with the more technical aspects of conversion to JIT/lean production.