ity of a fresh approach embodying this idea to the problem of job evaluation, test- ing it out in a wide variety of industrial, commercial, and public enterprises in ...
T. T. PATERSON
Professor of Industrial Administration University of Strathclyde, Glasgow T. M. HUSBAND
Lecturer, M.B.A. Programme University of Strathclyde, Glasgow
Decision-Making Responsibility: Yardstick for Job Evaluation The
job evaluation is to provide an equitable scale of payment for their value to the organization, and it should, in theory, give a jobs company at least four assists: (i) setting equitable wage differentials within its own wage structure; (2) setting equitable wage differentials between its own wage structure and those of other companies in the community; (3) setting equitable wage rates for new jobs in the company; and (4) responding rapidly in pay in relation
practices to technological change. In one degree or another, the four conventional methods of job evaluationranking, classification, points, and factor comparison-fail to satisfy these demands ; they rely for their evaluation to a great extent on purely subjective measures of essentially personal factors (e.g. education, experience, initiative) and ignore real assessment of the characteristics of the job itself. All these methods, despite their sophistication and earnest management efforts to minimize the subjective elements, still have acknowledged limitations, and management welcomes any new method that might help to overcome these deficiencies. The greatest shortcoming of the conventional methods is that they cannot compare unlike jobs. With them, it is impossible to compare the job of plumber in a company with, say, a sales clerk, not to mention middle and senior management 21
because characteristics such
working conditions, skill,
job families. Companies using the conventional methods invariably plan for the blue-collar workers and quite a separate plan for their white-collar employees. Actually, there should be no difference between wage and salary; there should be only pay. There is today a need for a plan to determine the equitable pay for aIl employees, from shop floor to laboratory, from front office to board room. In recent years, at the University of Strathclyde, we have been establishing the validity of a fresh approach embodying this idea to the problem of job evaluation, testing it out in a wide variety of industrial, commercial, and public enterprises in the United Kingdom. It has also been independently tested and substantiated by consultants in India and by the National Institute of Personnel Research in South Africa and it is now being adopted by an: increasing number of firms in Britain, where the National Prices and Incomes Board has ordered its adoption in the hospital service. European firms, too, are showing a growing interest in this method, for it is easy to learn and apply, inexpensive to establish, and has many side advantages, such as improvements in the basis for bargaining; organization planning ; merit-rating, promotion, and training plans; and incentive plans. vary
This Decision Levels
is called the Paterson Methrecognition that what is common
technique is the
all jobs is decision making and that all jobs are differentiated according to the kinds of decision made. Six levels of decision are differentiated, and jobs are analyzed and graded in terms of these decision levels: The &dquo;higher&dquo; the level, the greater the value of the job to the enterprise and therefore the greater the reward. The six basic kinds of decision that define the whole range of jobs in an enterprise fall into six &dquo;bands;&dquo; (see Figure i). Starting with the highest, they are: to
. Band E. Policy-making decisions. These decisions are made at the very top of the management hierarchy, at board level. The word director suggests the function of directing or guiding the enterprise, and the board does this by making decisions on company policy. The limits set on the choice of decision tend to be wide and, in many cases, are specified only by laws. Band D. Programming decisions. This kind of decision is typically made by senior managers such as divisional managers, plant managers, and chief engineers, and is made within the limits set by the policy decision of Band E. It is a plan or program for execution of the policy, but in Band D no execution of decisions takes .
place. Band C. Interpretive decisions. These are made by middle managers-e.g., departmental heads, such as a machine shop manager or foundry superintendentand are made within the limits set in the program at Band D. At this level there is still no execution of decisions. Band B. Routine decisions. Skilled craftsmen make these decisions, which are concerned with the execution of the interpreted policy decided at Band C. The craftsman or tradesman does not decide why something must be done; he is concerned with the way it is carried out, the process. Band A. Automatic decisions. Here, the semiskilled worker again does not .
have to know why he performs his operation; he need concern himself only with how to carry out the instructions about what to do and where to do it the operations within the process. Band O. Defined decisions. Decisions made by an unskilled worker are typically defined; he has only the most limited discretion, such as speed of working. .
The concept of decision bands can be further refined in the organization structure. Each decision band, except Band 0, is divided into two levels or grades, upper and lower. The upper-level job holder in any decision band coordinates the work of the men in the lower level of the same decision band; he is the line manager of the lower-level worker in the same band. Thus, for example, the upper level of Band B is the decision level of the shop foreman. His function is to coordinate the work of his subordinates, and he can therefore be seen as a manager in the accepted sense. On the other hand, the foreman is not widely accepted as a manager, simply because he does not decide why certain action should be taken. In Band 0, the lowest kind of decision made obviously requires only one level. Band 0 plus two levels in each of Band A through Band E show jobs in ii grades to be of increasing status or importance in the hierarchy in relation to their value to the company. In a manufacturing company, for example, typical grading might be as follows: A shop laborer and a copy typist are usually Grade 0; a sales clerk and radial driller are Grade i-the lower level of Band A; a shop craftsman and senior clerical sales staffer are Grade 3; the shop foreman and, perhaps, the senior sales clerk or section leader, are Grade 4; a machine-shop superintendent is Grade 5; several superintendents are perhaps coordinated by a departmental manager, who is Grade 6, and reports to the factory manager, who is Grade 7. District sales managers, Grade 5, may be coordinated by a regional sales manager, who is Grade 6 and responsible to the sales manager, Grade 7. The sales and plant managers and their status equivalents report to the general manager, who is Grade 8. The president is Grade io, coordinating the sales and manufacturing vice-presidents, both of whom are Grade 9. Grade 11 is represented by board members. This grading by decision responsibility is relevant to staff functions as well as to the line functions in the above Sta ff Functions illustration. For example, the quality-control chemist may decide that a certain temperature increase at a specific stage of a die-casting process will make it more efficient. Obviously, he is backed up by his specialist’s knowledge of the theory underlying the process, so the foundry manager, who has structural authority over the plant, feels obliged to follow his advice. If something then goes amiss, because of an error of calculation by the qualitycontrol chemist, the foundry manager can be held only partly responsible. He can24
be expected to question the expert conclusions of a specialist in such a highly technical area, so the quality-control chemist is also held responsible at this decision level, and his job can be compared with the foundry manager’s by &dquo;decision band&dquo; since he based his advice on his anticipation of a situation in which they were both involved. Every job is divisible into tasks that can be subdivided further into processes, operations, and elements. An employee carries out tasks that require decisions of the band to which he belongs and of the bands below, but never of the bands above. His tasks are analyzed and graded in a process that is very rapid, threequarters to half an hour for each job, sometimes less, and since there can be no mistaking the kind of decision, the placing in band and grade is completely objec-
subjectivity enters-and inevitably it does-is in differentiation within any one grade. The subdivision may be made by gauging the relative difficulty or complexity of the tasks-say, by the number of individual decisions of the highest kind in each task, or by the number of factors involved (requiring decision in-selection), or by the number of tasks themselves. In Grades o to 4 the point system can be adapted to the Paterson Method and the ranking system to Grades 5 to 8. It is impossible to be more objective than to specify grade, and the subjectivity of intragrade differentiation permits expression of the need for personal Where
opinion and so of negotiation. The Paterson Method claims only to provide a framework grades within each of which differentials can be adInappropriate justed by whatever method is suitable to the enterprisedecision difficulty for managers and office people, point system on the shop floor, and the like. Grading by this method is not designed to take into account factors such as skill, years of service, or working conditions. Basing job evaluation on measures of machining accuracy or knowledge of tools and equipment is usually misleading, because the most accurate finish may be expected of workers who happen to operate the most accurate machinery, and this tells little about the judgment or skill of the operator. Similarly, this method does not consider working conditions, because bad working conditions that are unavoidable can easily be compensated for by contingency payment above the basic rate and can be the subject of regular bargaining, especially in cases where the worker is not exposed to the hazard on a daily basis. For example, it is a common but illogical practice to grant maintenance personnel, as part of their basic hourly wage rate, 40 to 50 cents to compensate for working conditions that are only occasionally unpleasant. As for seniority, length of service is a matter that applies entirely to the individual doing a job and tells us nothing of the value of the job to the enterprise. Factors of
of educational background. In fact, the use of the conventional methods may introduce labor relations problems that previously did not exist. For example, a company often defines the degree of education required for various jobs with such terminology as &dquo;degree i-requires professional status; degree 2-requires technician status,&dquo; et cetera. When it comes to evaluating the jobs of a draftsman, who decides if he is a professional or a technician? If he is defined as a technician, there will be many dissatisfied draftsmen who had always thought of themselves as professionals, whereas if the draftsman is ranked as a professional, the laboratory technicians will, in many cases, insist that they, too, be classified as professionals. The point is simply that the conventional approaches lend themselves to terminology regarding status that in turn may lead to disparaging comparisons between the various work groups. The Paterson Method avoids this pitfall by concentrating on the only characteristic that matters, the type of decision required and the responsibility that automatically goes with it.
The method is now being applied in part of the British Steel Corporation. Earlier, the board had decided to grade posts from skilled craftsmen upward mainly on the basis of decision making, but there was no theoretical foundation, and factors of experience and education were taken into consideration. The final grading was carried out by a committee, and decision making was by far the most significant in grading, though the definitions of kinds of decision were vague and liable to subjective interpretation. The grading given by the Paterson Method not only gave a comparable result but provided a theoretical basis, thus eliminating the ambiguities of the empirical method. It was any easy step to modify and correct the existing method and so reduce the sense of inequity that seemed to arise from the earlier method. It took about ten days to analyze and grade a representative sample of jobs to demonstrate the applicability of the Paterson Method.
The Method in Use
with which the method
among its many advantages. The personnel officers of a large group of companies learned the technique in five days well to sufficiently put it into practice themselves. The salary scale of a whole service public covering every department-professional, nonprofessional, techniare
cal, and financial-has been based wholly on this method, and the officers applying it learned the theory and technique in two weeks. A standard questioning procedure can be used for analysis by interview, and grading does not require elaborate committee meetings to discuss degree definitions and the weightings of mental skill against physical skill, as is usual with other techniques. There is no need 26
staff salary plan separate from an hourly paid wage plan and a top executive payment plan. The single criterion of kind of decision is the basis for one to
comprehensive pay system. than 70 firms with nearly 60,000 employees and in practically have been examined for the distribution of their payroll in terms every industry of the Paterson Method, and from the managers of these firms we have learned the most fascinating aspect of the whole theory of decision systems: There is an apparent logarithmic relationship of existing wage rates paid to the different decision grades as defined by this method. The pay differential between each decision grade increases exponentially; that is to say, when the current pay rates are plotted on a log scale against the eleven job grades, the scatter points on the graph do not form a perfect straight line, but normally fall sufficiently close to indicate clearly that the line of best fit is straight. It appears that there is an unconscious tendency by companies to reward decision makers according to this specific All told,
logarithmic relationship. Figure z illustrates this relationship at its best. (The figures were obtained by students and others after only two weeks’ training.) The subdivision within grades was based on a simple tripartite estimate, subjective but near enough to be accurate. Not all firms showed this nearly pure logarithmic line, however; Figure 3
Figure 2. 27
be &dquo;erratic.&dquo; The interesting point, and the
compared to the others, these firms suffered from poor industrial repected, lations and staff problems. It seems, therefore, that if pay differentials conform to the theoretical logarithmic relationship, there is a greater sense of -&dquo;fairness&dquo; and so better morale, motivation, and industrial relations. It is also likely-though this has yet to be fully proved-that firms with the theoretical distribution tend to have an &dquo;organic&dquo; structure, more dynamic than the static, &dquo;mechanistic&dquo; structure, to use terms coined by Bums and Stalker. It should be noted that some departures from the straight line are due not to is that
pressures of the labor market. The type of skill required to do a particular grade of job may be in short supply, and companies often make the mistake of building this market element into the pay, producing an &dquo;unfair&dquo; differential with its attendant problems. Payments for short-supply labor should be contingency payments, not incorporated in the basic rate, so that when the market changes and labor reaches its proper worth, the contingency payment can be smoothed out. Or, as pay rises generally with cost of liv-
organization and industrial relations but
other factors, the contingency payment can be absorbed as corresponding increment. This labor market element of earnings lends itself naturally to the bar-
Figure 3. 28
of Increase The logarithmic
pay scale can be put to many uses. Given Figure 4, which shows the distribution of the payroll of a firm, the gradient, r, that is, the rate of increase of pay between grades, can be calculated:
Say the minimum adult wage is $3,50o and the pay of the general
manager,. Grade 8,
is $22,000, then
That is, there is
26 percent increase between
Figure 4. 29
gradient and the minimum adult pay, it is possible to calculate the pay for any other grade. If the minimum pay is $3,50o and the gradient o.26, then the Y4 = $ 13,500 x (1.26)14 = $8-825. pay for Grade 4 (a foreman) will be
3. It is
calculate the total
the number of
equitable payroll P, from
in any Grade i.
If the board of a firm decided what was to be expended on the total payroll, they could establish the pay value for each grade; they would only have to fix the minimum pay, say by collective bargaining. It is then a simple matter to draw the pay curve at the appropriate gradient from the minimum pay point on the Y axis. It has been found that the pay curve varies between industries and regions in the United Kingdom, probably because of differences in local labor markets and the degree of profitability attained by the various industries, and there also seems to be a relation between turnover and gradient. The gradient in large firms in the United States is greater that that in comparable firms in the United Kingdom. the advantages of an equitable pay system based Economy of Grading on decision band analysis, the Paterson Method can reveal the existence of jobs too highly paid because decisions of the are not made, or, to put it another way, human resources are not proper grade utilized to the full. being Reorganization to eradicate this failure-called the economy of grading-leads to job enlargement and increased motivation as well as labor saving. In one case, some senior and top managers, believing themselves to be overworked, sought to enlist high-level assistance, but analysis of their activities showed they were doing a large amount of work in lower decision bands. Economy of grading was applied, and instead of adding highly paid assistance, some clerks were brought in at the bottom of the structure and the load on the senior men was conIn addition
sequently lightened. Analysis also establishes
the key decision points, or nodes, as they are called, flow, and so leads to organization of the necessary information and advice for optimum decision making. Since many of these decision nodes are meetings, the precise definition of the required kind of decision permits of the use of &dquo;decision dynamics,&dquo; again the optimum use of people in decision making. Training techniques take on a new aspect. Training for promotion becomes in the work
difficult decisions in
higher grade. (Even training in manual skills can be improved for, by work study allied to decision analysis, the skills can be broken down into simpler complexes of decisions and taught on that basis.) Similarly, performance can be measured by the rate of learning and use of more complex and higher-grade decisions and by a superior’s procedure in training subordinates for promotion, which necessarily leads to greater departmental efficiency. In an age of rapid technological development, it becomes increasingly clear that people should be trained not for particular jobs that may become obsolescent, but for careers in increasing complexity of decision making, which will always be with us. A continuing business and industrial problem today is motivation. For the research indicates that the greatest gain is made when (a) jobs are &dquo;enlarged&dquo; by economy of grading, (b) development of decision- dynamics accompanies this enlargement, and (c) a systematic salary differentiation based on decision making is set up. On the shop floor, the blue-collar worker is usually offered (illogically) an incentive based on a piece rate that is not the foundation of his basic rate. But if his basic rate is founded on decision making, the incentive can be also related to decision making by measuring the increase in number (or complexity) of decisions per unit time. white-collar worker,
It has been said that
successful firm requires stability, Social justice high productivity, and social justice. The stability comes with the efficiency and profitability of higher productivity which, in turn, stems from motivation, and one of the most important motivating factors is the sense of social justice, of fair reward and punishment. Reward, to be fair, must be equitably related to the value of the job, which is the contribution a member makes to his firm. As Professor Jaques pointed out some years ago, people sense what is a fair differential; he found that the length of time that expired before a decision was scrutinized (a rough description of his time span of discretion) was frequently related to what was felt to be a fair pay. This is what is to be expected from the theory of decision bands. In general, the &dquo;higher&dquo; the decision, the longer the time before it is executed in full. It follows that differentials based on the Paterson Method will tend to be regarded as fair and acceptable. Even more important in developing a sense of social justice, this method is applicable equally to all members of the firm, whatever their positions, so all are seen to be equal even though different. a