A common argument in favor of promoting manufacturing is that jobs in that sector are among the most valuable to American economic health—certainly more valuable than jobs in the service sector because they are associated with more jobs elsewhere. For example, Peter Navarro, head of the Trump administration’s National Trade Council, argues that manufacturing has a high employment multiplier—i.e., manufacturing generates a large amount of ancillary employment in other fields. Reports from the Manufacturing Institute indicate that each dollar’s worth of manufactured goods creates another $1.34 of value elsewhere in the economy. This is the largest “multiplier” of any sector. The impression sometimes given by such claims is that a dollar spent on manufacturing will “create” more employment opportunities than a dollar spent on services.
It is certainly the case that when manufactured goods finally reach consumers, they are likely to contain value added from other sectors of the economy. For example, autos contain not only the value that was added by the firms that undertake auto assembly but also value added that can be ascribed to materials (iron ore, glass, rubber, etc.), transportation, and services such as banking, accounting, advertising, and wholesale and retail distribution. And it may well be the case, by contrast, that a much higher share of the overall value of certain services, for example haircuts, is added in the haircut industry, and therefore haircuts have a smaller multiplier in other sectors of the economy. But before we jump to conclusions about how value added affects employment, caution is in order.
First, the total employment involved in any production ultimately depends on the value of that product divided by average labor productivity. Take a simple example: Assume that output per hour is $40 in every industry in the economy. How many hours of work would be created by the final sale of haircuts of $10,000? Answer: 250 hours of work—almost all in the haircut industry. How many hours of work for a final sale of autos worth $10,000? Again, the total would be 250 hours, equating to some jobs in assembly but many others in a variety of sectors. If we assume that all the value added for haircuts was added by hairdressers, the haircut multiplier would be 1. If we assume that of the 250 hours it took to make the autos, 100 occurred in auto assembly and the rest in other parts of the value chain, the auto multiplier would be 2.5. Thus, even though haircuts have a “multiplier” of 1 and manufacturing a multiplier of 2.5, the $10,000 spending on each would give rise to exactly the same number of jobs.
A second issue relates to the location of the jobs. The multiplier estimates that the Trump administration refers to are based on US input-output tables that assume all production in the value chain takes place in the United States. But in reality the raw materials, transportation, and other intermediate inputs used in manufacturing need not be produced within the same country. Thus the multiplier does not necessarily refer to domestic jobs. Actually, since services generally rely far less on imports, a final dollar’s worth of spending on services is likely to spur more domestic employment than the same spending on manufacturing of equal value.
A third issue relates to the arbitrary allocation of activities and value addition to manufacturing. Research and development, for example, can be performed in universities and laboratories and thus represent services, or within manufacturing firms and thus be counted as manufacturing value added. The same is true for many other services: janitorial, data processing, accounting, marketing, advertising, legal, etc.
Finally, precisely because much of the final value of manufactured jobs reflects value added in other sectors, it is inappropriate to assume that a deficit of a million dollars in manufacturing trade means that the economy is producing a million dollars less value added in the manufacturing sector. Indeed, if the manufacturing multiplier is 2, closing a trade deficit through increased domestic production would only create half of the additional jobs generated in the manufacturing sector itself, while the other half of the additional employment would be added in other sectors.
All told, we draw conclusions about employment using these “multipliers” at our peril. Certainly, these multipliers should not be given the causal significance often attributed to them.