Numerically controlled machines require programers of considerable technical education. Such programers, few in each organization, will have university-level educations in engineering or mathematics. However, in the final development of automation machinery computers will be preparing programs. As at present, the operator of automatic equipment will be unaffected by the skills and technical competence of programers and setup specialists. Although it is not clear at present, the findings of job design research indicate that increased productivity and quality result from the inclusion of setup, inspection, and control among the tasks of operators. For some automated systems, then, there is a distinct possibility that the jobs of operators, or direct workers, will be combined with those of indirect workers.

CONCLUSIONS

1. Under automation, work and jobs at the operator level are so different from more conventional production jobs that there are no means for making direct comparison of content or skills. Some serious questions should be raised about the feasibility of public policies directed toward raising and changing skill levels when so little knowledge is available.

2. Skills for operator jobs appear to be nontransferable and consist largely of monitoring or information handling, adjusting, and making judgments and taking actions which will minimize downtime. These skills have been acquired by workers drawn from backgrounds equivalent to those of semiskilled operators in present mass-production industries. The fears expressed in some quarters concerning a technological lockout of workers seeking to enter automated industries appear to be unwarranted.

3. The pessimistic views expressed by sociologists regarding the effects of jobs in automated industries may be interpreted as reflecting their disappointment over the failure of the predictions that automation would raise skill levels, remove drudgery, and generally provide a new world of jobs for workers. Automation represents a continuation of the rationalization and mechanization of the production of goods and services which began about 150 years ago. The significant differences between jobs under automation and those in mass-production systems are that automation increases the operator's responsibility, his overview of the process, his technical information, his sense of consequences, and his informational integration with technicians and engineers.

4. In the area of union-management agreements, new approaches will have to be developed for providing compensation above base levels and for the evaluation of jobs. Present job evaluation schemes with their built-in inflexibility will prove to be inhibiting to the appli

cation of automation.

5. The ultimate development of operator jobs under automation will probably result in a breakdown of the boundaries between machinefront and machine-back work. These jobs will combine monitoring, regulating, adjusting, and spot maintenance. The prediction is based upon the trends toward job enlargement stemming from job design research, the development of relative self-reliance, and the increased responsibility characteristic of automated jobs.

6. For maintenance work there will be an increase in skills and content. For the same job-enlargement reasons, our equivalent to the

polyvalent craftsman will probably be developed. Control system technicians will also be developed and may well become a new type of maintenance technician.

7. Considering industry and business as a whole, there will be an upgrading of jobs as a consequence of automation, in that the relative number of engineers, scientists, and analysts will continue to increase.

THE SKILL IMPACT OF AUTOMATION

(By Paul Sultan, Claremont Graduate School, and Paul Prasow, Institute of Industrial Relations, University of California, Los Angeles)

Much of the present diagnosis of current unemployment has tended to polarize on two explanations: Unemployment is said to reflect either the inadequacy of demand or the inadequacy of labor force adjustments to "adequate" demand. That distinction, in effect, attempts to attribute unemployment to dislocations on either the demand or the supply side of the labor market, or to establish the labor market's more important cutting edge. While that dichotomy draws attention to the complexities of the employment problem to be found on both sides of the market, and is obviously a crucial distinction in any discussion of remedial action, it does more to encompass than transcend traditional explanations for unemployment. As a consequence, it has somewhat inhibited the full consideration of the wisdom found in existing employment theories, simply because in much of traditional analysis, demand and supply are not usually regarded as competing or mutually exclusive determinants of employment. The purpose of this paper is to review some aspects of employment analysis rooted in supply considerations, and more particularly the range of opinion regarding the impact of automation when consideration is given, not to the amount of labor demanded, but to the quality of labor that must be supplied.

Unemployment that is not explained by a lack of demand is labeled structural unemployment, and since this category is pressed into service to explain everything not explained by a lack of demand, it is not surprising that confusion and uncertainty surround its meaning and cause. It is said to be unemployment not quite synonymous with technological, frictional, or even hard-core forms. Indeed, much of the conjecture on structural unemployment suggests that it involves something more than a reclassification of conventional doctrines: Rather, it categorizes a form of unemployment that is quite distinctive in substance from anything we have faced in the past.1

1 Concepts of unemployment may be developed within the framework of a particular economic theory or, as a pragmatic alternative, to categorize available unemployment data. The "structural" designation has both causual and descriptive pretensions. It is said to be caused by pronounced shifts in the direction of consumer expenditure, technical changes that involve a radical alteration in the form and location of production functions, "administered" wage and price adjustments, and impediments or delays in labor adjustment to labor-market requirements. Since structural unemployment reflects the lag of supply adjustments to demand, it is "identified" by its very persistence. It is further isolated by the heavy incidence of unemployment in those industries, occupations, and regions most affected by change, and by the increasing incidence of unemployment for those components of the labor force with particular age, skill, educational, race, and sex characteristics least willing or able to accommodate themselves to change. It is not surprising that a "theory" of structural unemployment that would encompass the complex of forces affecting the quality and quantity of labor supply, as well as the quality and location of labor demand, has yet to be developed. For discussion, see "Unemployment: Terminology, Measurement, and Analysis." Subcommittee on Economic Statistics of the Joint Economic Committee. 87th Cong., 1st sess. (1961).

There are several reasons for the current attention given to the supply side of the labor market, or, more specifically, to the structural unemployment explained by these supply considerations. First, the collective bargaining process has been increasingly involved with adjustments to automation, or the adjustments of labor supply to the "uncontrollable" shifts in labor demand. But while the levels of aggregate, industry, or corporate demand usually appear well beyond the influence of the parties to the bargaining process, contract terms designed to cushion labor from the impact of technology are not; these are now emerging as the "gut" issues of collective bargaining. There is growing evidence of employer willingness to collaborate with unions on setting up early warning systems, to implement programs for labor relocation and retraining, to phase the introduction of technology in a manner designed to minimize human hardship. A consensus has emerged that every effort must now be extended to increase the volume and improve the distribution of economic intelligence on job vacancies, to encourage the mobility of labor, to undertake the task of upgrading labor skills. And where corporate, union, and personal resources appear insufficient to meet the cost of the adjustment process, there seems to be less resistance to the use of public funds designed to speed the reeducation, retraining, and relocation process.2

The hope of reducing unemployment by adjustments in the quality and location of labor supply has been further nurtured by the growing volume of statistical evidence pointing up the high correlation between education, job skill, income, and job security. Such evidence, together with the eloquence of those with a vested interest in extending our investment in human capital, gives further authority to the hypothesis that we can expand employment by upgrading labor skills. America's balance-of-payments crisis is also relevant to the structural unemployment issue. The efforts of fiscal and monetary authorities to sustain the buoyancy of an economy constantly overloaded with upward cost-price adjustments now seem to be weakening. As we appear to be running aground on the reef of foreign competition, few feel it is prudent to raise the keel of exchange rate stability and even fewer are willing to see the Nation jettison additional amounts of domestic employment. It is not surprising, then, that hope has developed that we can contrive those structural alterations in the ship of state that would allow for both stability and buoyancy. Ironically, the confidence we now have in the capacity of the Government to provide full employment emerges at a time when these fiscal instruments to stimulate demand and employment must be employed cautiously. Barred (to employ the Keynesian analogy) from moving the piano stool to the piano, we are now involved in the tedious task of moving

Unions, of course, have strongly favored such programs. There was striking evidence of agreement at the American Assembly Conference on Automation, Palm Springs, Nov. 8-11, 1962, when both union and management representatives gave full support to the retraining principle. A dispute developed, however, when two educators took issue over the relative importance of the liberal versus the applied programs offered by the schools they represented. A unionist quipped: "Why don't these fellows settle their jurisdictional dispute elsewhere?" Emerson P. Schmidt of the U.S. Chamber of Commerce testified to a congressional committee: "We suspect that a dollar spent on retraining has many times the impact of a dollar spent on loans and grants for industrial or community facilities. We would not be surprised if the $5 million appropriated for retraining in the depressed areas measure approved by the House does more good than the rest of the appropriations combined." "Impact of Automation on Employment," hearings before the Subcommittee on Unemployment and the Impact of Automation of the House Committee on Education and Labor, 87th Cong., 1st sess. (1961), p. 472.

the piano itself. Or, to put the case more directly, the constraints represented by our balance-of-payments require that we give attention to adjustments of labor supply rather than to problem adjustments of labor demand.

Further attention to structural unemployment has developed by those who see cost-push inflation as the cause of cost-push unemployment. In this view, the increase of labor costs has compelled some labor displacement in the short run, and as it encourages the substitution of capital for labor in the long run, it compels even further labor displacement then. The resulting unemployment is said to be structural inasmuch as it reflects the pressure of structural blocs pushing for gains that cannot be supported by industry.

The designation of unemployment arising from this source as "structural" has not inhibited union use of the same term. But in the union view, structural unemployment reflects structural flaws in the pricing mechanism, and more specifically the insensitivity of prices to expanding capacity and expanding labor productivity.* Price rigidity denies that stimulus to market demand necessary to absorb expanding supply. Structural unemployment emerges because of malfunctions in the pricing mechanism, because of the structure of product market concentration, or even because of the capitalintensive innovations that provide dramatic economies in the use of labor.

Although the usefulness of this category of unemployment is weakened when it becomes a term meaning all things to all people, or when a single label is used to explain conflicting or contradictory causes of unemployment, that confusion does not allow us to retreat from the full examination of remedies for unemployment that can be located in the supply side of the market.

In approaching the employment problem from the supply side, one would expect to be confronted immediately with proposals to reduce wages, or to alter the location of the labor supply schedule so that it might intersect the labor demand schedule where higher levels of employment can be obtained. While there is today general support for the proposition that wage adjustments must somehow be

*L. E. Gallaway finds only limited statistical support for the structural unemployment argument, but he is unwilling to surrender to the lack-of-demand alternative. "Rather the findings at the microeconomie level offer the possibility that the impact of union market power may be disrupting the labor market allocative process at the interfactor level; 1.e.. that the relative price of labor in general has been increased by the actions of labor unions to a level that generates widespread unemployment." "Labor Mobility and Structural Unemployment." American Economie Review, vol. L-III. No. 4. September 1963, pp. 714-715. Harold Demsetz concludes his statistical analysis: "Minimum-wage laws and union wage rates make it impossible or dieult for a growing component of our labor force to offer Its services at wage rates setently low to be employed." "Structural Unemployment: A Recons deration of the Evidence and the Theory," Journal of Law and Economics, vol. 4. October 1931, n. 90.

•Solomon Barkin contends that price setters in administered-price industries are confident that governmental monetary and decal action will support their decisions to raise prices: they are subgot to no compelling pressure to lower prices. In his view, this explains the tehrorte imbalance between profzettre expäelty and cenemmption" Impact of Automation on Keplerment en dt. p 195 In more technical analysts. Srlos-Labini attributes pro radite and its Arendiet of eeroomie stication to technological discontinuities wirdt welierte Industries, conterdig that the alternadre to prices that allow large *** dixital past is sheer chres. Cigopoly and Technical Progress" (Cambridge: Narvard University Press, 1962),

kept within the limits of labor productivity, the hypothesis that wage reductions can encourage employment, or even inhibit the pace of capital-intensive innovation, has not yet received much support. But the attention usually reserved for the cost of labor is now given over to the quality of labor.

Even though wage policy and skill upgrading programs are in no sense indentical, both can be analyzed in terms of their influence on the labor supply, or, to put the case more directly, in terms of the employability of labor. Policies of wage reduction or skill upgrading are somewhat analogous in their employment generating consequences, for both are presumed to lower unit labor costs. In the first case, the reduction of hourly wage rates, with labor productivity unchanged, obviously reduces unit labor costs; in the second case, increases of labor productivity, with hourly wage rates unchanged, have a similar effect. And so long as we admit some measure of elasticity in the demand for labor, the reduction of unit labor costs by either device should encourage the employment of labor. In the thirties, Keynes offered a convincing rationale for reducing the "real" cost of labor by increases in prices rather than through the painful struggle to reduce money wages. In the sixties, we still hope to bypass the hostile reception that would follow any exhortation that money wages be cut, by programs to increase labor's productivity or skill. Inhibited as we are and for good reasons-from tinkering with money wage levels, and inhibited further by the discipline of foreign competition from adopting an overt inflationary policy, it is not surprising that we should turn hopefully to the remaining expedient: devices that would increase the employability of labor through the upgrading of labor's skill. There is, therefore, an analogy between the hope of the thirties that everyone might be reemployed if wages (or prices) were set at levels that would allow for their profitable employment and the hope widespread in this decade that everyone might be employed if skill levels were so upgraded that employers could not resist the utilization of such talent.

What, then, are the prospects for minimizing unemployment by speeding labor's accommodation to the requirements of industry, by upgrading labor's skill levels? The range of responses to this question is indicated in charts I, II, and III.

In chart I, frequency distribution A represents a hypothetical skill mix possessed by the existing labor force. In order to preserve the distinction between lack-of-demand and structural arguments, we shall arbitrarily assume that automation does nothing to reduce the number of job opportunities in industry. In other words, to focus on the skill issue alone we shall isolate our analysis from conjecture that automation reduces the total number of job opportunities (viz, reduces employment to the shaded distribution C in chart I). So long as parity is preserved between the total of job opportunities and the number in the labor force, unemployment reflects the problem of matching the specific qualities of labor demand with the specific attributes of labor supply.