Challenges to Craft Control
Engineers or machinists completing their apprenticeships around 1890 would live to see profound changes in their trade. Over the next thirty years craft control would be undermined and management prerogatives emphatically asserted and exercised in new ways. This transformation did not occur all at once, but in fits and starts, varying in character and extent from one shop or sector to another. The trend was clear, however. Indeed, because changes at work were the subjects of wide debate and contention, the trend was perhaps as clear to managers and workers at the time as it is to the historian today. Its principal features were deskilling, the spread of piecework, and more exacting and detailed management supervision of workers. From the craftsmen's point of view, manual skills were parceled out to specialists and transferred to new machines, and knowledge and control of production were centralized in management hands and embodied in specific instructions and finely calculated piece rates.
The most far-reaching challenge to craftsmen's position was deskilling (or "dilution"), three aspects of which are particularly important: the transfer of knowledge and control to management, the transfer of skills to machines, and the fragmentation of craft work into specialized tasks.
Management, not craftsmen, gradually took charge of production. Instead of handing skilled workers rough drawings and leaving the rest to them, management planned out the job before it ever reached the shop floor. The order in which work should proceed, the routing of materials and parts from one machine or department to the next, and even how each individual operation
should be done (which machine to use, the requisite speeds, feeds, and cutting angles) were increasingly determined in the office by management rather than on the floor by craftsmen and foremen. In the well-managed shop of the 1910s and 1920s, materials, tools, and instructions would be provided to machinists or engineers as required; they could then devote their full attention to productive work.[13]
Skill was transferred to machines as well. Even before the 1890s, specialized machine tools joined other, more versatile machines as part of the standard machine shop equipment. Accomplishing varied tasks on a general-purpose lathe demanded highly skilled workers. Many of the new tools, designed for specific operations (e.g., drilling machines and screw or gear cutters), did not. Others, such as millers, slotters, or grinding machines, performed operations previously done by hand. With single-purpose machines came specialists, operators skilled in a narrow line of work. In their own line they turned out work equal in quality to that of all-around craftsmen, and did so for lower wages.[14]
Newer machine tools in the 1890s and after took this process further. Where large quantities of a standard piece could be sold, employers increasingly adopted automatic machinery. Once set up, turret or capstan lathes, screw machines, and some milling or grinding machines could conduct a series of operations without human intervention. Operatives, who only had to place the piece in the machine, start it up, and remove the finished product, needed little manual skill or technical knowledge. Moreover, with machine tools able to turn out standard pieces to close tolerances, skilled fitting of parts gave way to less skilled assembly. Eventually, this trend led to assembly lines. And with both specialized manual machines and automatic ones, management claimed the right to assign operators as it saw fit, regardless of union rules or Customs.[15]
New technology combined with reorganization to change the functions and reduce the numbers of fully skilled workers. If employers were to pay for craftsmen, they sought to get their money's worth. A variety of tasks customarily performed by skilled men (e.g., getting and grinding their own tools, fixing belts, and chasing after work materials) required neither their skills nor their time. Some of these tasks management assigned to specialists in tool grinding or belt repair. Others were facilitated by installing cranes
and trolleys and by making more rigorous provision for delivering the appropriate materials and tools to workers at their machines. Skilled men thus found themselves gradually confined to their posts, working with little interruption for the full day and often at a single machine—or even several machines of the same type.
The goal of many managers was to shift craftsmen out of production altogether. Production departments would be staffed by less skilled specialists and operatives, turning out large quantities of standardized parts. Skilled workers would be responsible for setting up machines, for repairs, and especially for the varied tasks and limited output involved in making tools, jigs, fixtures, and gauges.[16] Their numbers declined accordingly. In 1910, "nearly three-fourths of all jobs in the auto shops were classified as skilled work. By 1924, skilled workers were estimated as five to ten percent of the work force of the auto shops."[17] An extreme case—the Ford Motor Company—shows the proportion of skilled mechanics falling from 28 percent in 1910 to 2.4 percent in 1917.[18] Among firms belonging to the British Engineering Employers' Federation, 60 percent of workers were classified as skilled in 1914, 34 percent in 1928.[19]
While the numbers of fully skilled workers diminished, the numbers of specialists and semiskilled machine tenders increased. In British engineering firms the proportion of the work force classified as semiskilled stood at 20 percent in 1914 and 53 percent in 1928.[20] At Ford 26 percent of employees were specialists in 1910, 55.3 percent in 1917.[21] Between 1910 and 1920, the number of journeymen and apprentice machinists in the United States grew 82.8 percent and then declined by 22 percent over the next decade. By contrast, machine operatives in auto and farm equipment increased 473.1 percent during the years 1910–1920, and their numbers continued to expand at a slower pace over the following ten years.[22] Semiskilled workers formed a new stratum in the workshops, blurring the traditional distinction between skilled men who operated machines and labors who fetched and carried. The latter also declined in numbers as management improved work flow coordination and installed mechanical conveyances (e.g., cranes, trolleys, and moving belts) to help move materials.[23]
One corollary of dilution was the decline of union apprenticeship regulations. The value of apprenticeship for employers dimin-
ished as the proportion of semiskilled workers increased. Managers who planned manufacturing tasks, divided jobs, and adopted specialized machinery had less need for broadly trained workers. The ASE and the IAM sought to preserve a system under which a limited number of youths acquired all-around skills and to place only workers thus trained on the machines. With increasing success, employers instead hired apprentices and learners in excessive numbers, kept them in one line of work indefinitely, and assigned them to machines claimed by craftsmen.[24] In the United States apprenticeship declined to a point where employers complained that they could not find the few all-around machinists still required. By 1910, efforts were under way to increase the supply of skilled labor through company training programs and technical education. Unfettered by union regulations, these schemes promised employers as many workers as needed, with training tailored to the demands of modern manufacture. If, through government-supported technical schools, this cost employers little and removed boys "during the formative period, from the union atmosphere," so much the better.[25]
Dilution circumscribed the tasks skilled workers performed. Yet it also transferred from craftsmen to management control over assignments to machines and methods for doing the work. On the surface, payment by results seemed merely to base earnings on output rather than on time worked. Here, too, however, a great deal more was involved. Through payment by results employers sought to wrest from skilled workers and foremen their customary control over the pace, character, and costs of machine shop production.
In 1886, 5 percent of men and youths in British engineering and boilermaking were paid by the piece; by 1906, 27.5 percent of them were paid in this fashion; and in 1927, the figure had grown to 49.5 percent. Among specialists, women, and workers in industries characterized by repetitive production, piecework was even more popular with employers. In 1906, 47 percent of "machine men" (specialists), 52 percent of women workers, 52.8 percent of cycle makers, and 67.6 percent of railway carriage and wagon builders were paid by results.[26] Comparable figures do not exist for the United States; but reviewing the period from July 1, 1907, to June 30, 1909, IAM President O'Connell estimated that more than
50 percent of the strikes involving union members resulted from attempts to introduce piecework and that the IAM had been largely unable to prevent the spread of the system. Certainly, piecework would have been even more prevalent in nonunion shops, and contemporary observers found a greater resort to payment by results in the United States than in Britain.[27] By 1922, of twenty-eight metal trades plants studied by the Federated American Engineering Societies, sixteen had well-developed wage incentive plans, and only three had no such plan.[28]
The earliest, simplest, and most widely used system of payment by results was straight piecework. Instead of being paid so much per hour, workers received a certain sum for each piece produced. Particularly after 1900, more exotic systems gained attention and popularity. These differed in operation and name but shared a basic principle. A certain time would be set to complete a job or produce a certain number of pieces. The time saved or the amount by which the quota was exceeded formed the basis for calculating a bonus. Exactly how that bonus was figured varied and could be quite mystifying. In perhaps the most straightforward system (the "Halsey" plan), a job time might be set at ten hours. Workers who completed the job in eight hours earned their hourly time rate for those eight hours plus a bonus of 50 percent of the time saved—that is, they received nine hours' pay for eight hours' work. Other schemes were considerably more complicated.[29]
Whatever the system, the tendency was to relieve foremen of responsibility for fixing rates, this duty falling instead to specialized personnel ("rate fixers" and, eventually, time and motion experts). If the work resembled that done in the past, rates might be fixed from the office on the basis of records of previous times and estimates of how long a machine "should" take to complete the job, given the relevant variables (the materials being used; the amount of metal being removed; the appropriate speeds, feeds, and cutting angles; and the desired finish). Particularly with new work, the tools of scientific management—the stopwatch and the motion study—were deployed to determine how quickly a job could be done. Depending especially on the strength of union organization in the shop, rates might be subject to negotiation or imposed unilaterally.
Payment by results presupposed a degree of standardization
and repetition in machine shop operations. Where output consisted of only a few pieces of a kind, as in much tool room and repair work, it did not pay to invest the time and money required to estimate rates and calculate bonuses. Elsewhere, however, incentive schemes had considerable advantages for employers. Payment by results helped centralize control over incentives and work pace. Craftsmen would no longer be allowed to define "a job well done" or "a fair day's work."[30] Increased control over effort levels, in turn, allowed management to lower the costs of supervision and to estimate more accurately costs and delivery dates for prospective customers.[31]
A final component of the employers' challenge involved efforts to substitute tighter management control over workshop tasks for the discretion and autonomy enjoyed by skilled workers. Instead of relying on their workers' seasoned judgment, managers increasingly issued detailed instructions to their hands, specifying how machine tools were to be set up, the speeds and feeds at which each job should be run, and the specific dimensions of the finished piece. Such orders were necessary above all for less skilled employees, for whom clear directions substituted for personal knowledge. Cheap workers required costly supervision; deskilling thus forced managers to assume new responsibilities. But the detailed direction of tasks applied to skilled men as well. Increasingly, work was given out with job cards attached, indicating precisely how the job should be done and the piece rates or bonus times allowed.[32] Management sought to dispense with craftsmen's judgment in favor of central planning and direction of work performance.
Once instructions were issued, of course, they had to be enforced. Employees of all skill grades faced closer supervision and scrutiny to ensure obedience to rules, adherence to instructions, and constant application to their work. Time discipline became more relentless and exacting as the Bundy time recorder gained favor, and metalworkers faced the prospect of losing from fifteen minutes' to a half day's pay if they arrived moments past starting time. Once at work they might have to clock in and out of each job to ensure compliance with production standards, provide employers with accurate information on times and costs, and prevent worker "fiddling" with premium bonus times. Later, even this check proved too rough and ready, and time discipline narrowed
its focus to individual motions, measured in hundredths of a second.[33] All increased supervision was not so impersonal, however. Particularly as firms grew in size and management assumed new responsibilities, engineers and machinists found themselves "watched and dogged by a whole army of non-producers."[34] At Ford in 1914, one foreman supervised an average of fifty-three workers; in 1917, there was a foreman for every fifteen.[35] An array of new supervisory personnel—rate fixers, speed and feed men, progress chasers, inspectors—confronted workers at every turn, leaving few aspects of machine shop life to the discretion of employees.
At first new supervisory tasks, together with responsibilities for recruitment, job assignment, and piece rates, were added to foremen's duties. Such power could corrupt, and foremen were often found tyrannical and abusive by their subordinates and unreliable by their employers. Gradually, the foreman's functions were parceled out to specialists and centralized in planning and personnel departments.[36] The foreman's responsibility slowly narrowed to ensuring that standards set elsewhere were met by those under him. In this way skilled workers, who might have compensated for their loss of autonomy by informal influence with foremen, found even this form of control circumscribed.
Dilution, incentive pay, and tighter authority were familiar to engineers and machinists employed in large or progressive plants and were widely advocated in trade journals by the mid-1890s. The movement for "scientific management," which developed after 1900, was not in practice the revolutionary innovation heralded by its promoters. Scientific management extended trends already under way: deskilling, payment by results, and more exacting methods of supervising work and workers.[37] The new management systems that gained favor in the United States and to a lesser degree in Britain may be distinguished from earlier developments in three respects.
First, these changes were achieved less through mechanical innovations than through administrative reform. Scientific management did not aim for technical change; it sought instead more efficient means of organizing and disciplining the work force. This reflected a common managerial sentiment of the period: "The greatest problem before us today is not so much the further im-
provement of machinery, but the development of an increased efficiency in men."[38] Thus under scientific management employers pursued deskilling through organizational and supervisory techniques that transferred skill in the planning and execution of work from operatives to new staff experts. Two such techniques occupied a prominent place in scientific management. One was job analysis, sometimes involving time and motion study and sometimes not. Through a careful scrutiny of machine shop operations, management identified less skilled tasks that could be split off and assigned to less skilled workers. Scientific management also demanded systematic attention to the routing of work, the sequence of operations, and the provision of all equipment necessary for each task. Rather than relying on the operative to determine what tools and equipment were needed and to secure and modify those materials as required by the job at hand, employers would undertake those responsibilities. Once such provisions had been made, the narrowed task of running each machine operation could be left to less skilled men or women.[39]
Second, changes after 1900 represented a more self-conscious and systematic approach to shop management and were more likely to be introduced by self-proclaimed "experts" brought in for the purpose of improving efficiency throughout the firm. Despite their wary reception of Taylorism, employers increasingly accepted the view that organizational changes could not be made on a piecemeal basis. Unlike earlier workshop innovations, scientific management entailed a more thoroughgoing, interconnecting reorganization of accounting procedures, layout, work flow, and planning, production, and supervisory tasks.[40] Scientific managers approach to incentive pay, for example, involved more centralized and systematic setting of production standards and premiums. Establishing output quotas and corresponding incentives, in turn, would increase production only if steps were taken to eliminate delays in the movement of materials. And all these tasks required experts.[41]
Third, and most important, scientific management was intended to be a coherent system of employer control to replace the controls exercised by craftsmen.[42] Managers sought to decrease their reliance on the manual and mental skills of machinists and to develop reliable mechanisms through which they, rather than their employees, could control the workplace. The first goal necessitated
the second. As employers replaced fully qualified machinists with less experienced and less versatile workers, they had to assume responsibility for planning and organizing production—often in considerable detail. Scientific management was designed to handle these tasks,[43] and here more exacting methods of supervision played a major role. Closer monitoring of work enabled management to discover the less skilled components of tasks. Combined with more extensive use of time clocks and stopwatches, job study permitted employers to fix production norms on which premium bonus systems then would be based. By issuing detailed instruction cards, tightly supervising workers, and providing systematically for the proper equipment at work stations and the flow of materials, managers hoped to ensure that quotas would be met and tasks performed adequately by relatively inexperienced operatives. Such new functions brought new personnel—rate fixers, progress chasers, speed and feed men, inspectors, tool room clerks, production planners—organized along bureaucratic lines into specialized departments and offices.[44] Scientific management developed for employers a set of standards and mechanisms of control to replace those of craftsmen.
In the United States, fully developed scientific management systems existed more in theory than in practice. Employers typically picked from among its techniques those that suited their business and seemed relatively easy and inexpensive to install. Scientific management also represented novel approaches to traditional goals—dilution, incentive pay, and closer supervision and control. With these same provisos scientific management appeared in British engineering well before World War I.[45] Particularly in new sectors of the industry, employers devoted greater attention to organizing and administering work and hired new personnel to devise and implement more systematic production planning. As in the United States, the goal was to preempt craft control in the shops. But in Britain changes in management practices were less systematic and less successful in imposing alternatives to craft control.
With or without the trappings of scientific management, dilution reduced craftsmen's control of production tasks and diminished their hold on the supply and training of skilled labor, piecework eroded their control over work pace, and these together with new
authority relations diminished skilled workers' discretion and autonomy. Responsibility for shop practices increasingly lay in the hands of managers. What accounts for this transformation at work?