a. Short run: From the end of World War II until the 80s many manufacturing industries such as automobiles and steel were negligent in promoting manufacturing innovation. Consequently, manufacturing technology in these industries was inferior to the world state of the art determined by Japan. Since about 1985 there has been a massive restructuring of US manufacturing industries to advance innovation and regain their position as world leaders in technology. US industry is playing catchup to the Japanese. While we are ahead in programming, they are ahead in FMS, robots and quality control. These changes are causing profound changes in manufacturing. First, in these industries blue collar labor will gradually decrease with job destruction occurring at a much faster rate than job creation. New industries such as biological manufacturing products using bacteria are likely to be highly automated from the start. Therefore, the restructuring will displace large numbers of blue collar workers from manufacturing altogether.
In addition, the number of white collar workers in manufacturing will greatly decrease. Staff and middle management presently perform an information function. With integrated corporate data bases, much of this work can be done by assistants reporting directly to the top management group. Chrysler already has laid off numerous white collar workers. GM is considering laying off 10%of its white collar workforce. The move to electronics instead of paper in the corporation and the ability to analyze data through programs means that corporations can be more profitable with fewer white collar workers.
Long run: Once the restructuring is complete in the next few years, the pace of displacement of workers from manufacturing will definitely slow down. In order to discuss how fast workers will be displaced from manufacturing over the long run, we need to consider the automation frontier. Automation proceeds from the structured to the unstructured environment. It is a major advance to get machines to deal with contingencies, for example, what the robot does if it drops a part. A programing approach must consider every possible contingency. In trying to program robots to do even simple tasks like moving blocks, the number of contingencies is large. Genuine progress will be slow. One of the ways is to redesign the production process in a streamlined fashion with fewer random elements. For example, painting and welding by robots were accomplished not by giving robots sight, but by expending a great effort to have the auto exactly positioned for the task. An example of a market where automation will proceed slowly is housing and construction. When housing is automated, the type of house produced will probably be very different from today.
The displacement of employment by automation will probably be gradual. The speed of the automation frontier in displacing humans is a function of technological advance, investment, and the problem of adapting automation technology to similar industries. The knowledge of how to automate a particular operation will be acquired by trial and error on each new plant. A new advance does not suddenly sweep over all industry instantaneously. Automation requires major investments in new facilities which businessmen are generally cautious about until a new technology has been demonstrated a definite winner. Once an operation becomes demonstrably cheaper to do by automation, a period of ten to thirty year exists before all human labor will be displaced from that operation. Another factor which requires time is that while each industry is similar, they do have their own distinct features. Effort is required to adapt successful technology in one industry to a similar related industry. I believe the decrease in employment in manufacturing will proceed, like the decrease in employment in agriculture, gradually over time.