Introduction
The American economy has been transformed in the last three quarters of a century—and no change has been more profound than the shift from manufacturing to services. In 1940, roughly one in three non-farm jobs were in manufacturing; by 2016, it was fewer than one in ten.
Even in absolute numbers the changes are staggering. In 1940, 10.1 million people worked in manufacturing, a number that grew, over time, to 19.4 million in 1979. Since then, manufacturing employment has declined, and by 2016, it was just 12.4 million. In fact, the American economy has added 112 million jobs since 1940; of those, only 2.2 million were added in manufacturing. Manufacturing durable goods to be exact—machines, vehicles, aircraft, and the like; employment in manufacturing non-durable goods—food processing, beverages, paper, plastics—is actually lower than it was in 1940, even in absolute terms.
| United States: Manufacturing employment, 1940 vs. 2016 | ||||||
|---|---|---|---|---|---|---|
| 1940 | 2016 | Change | ||||
| Employment (000s) | Jobs | % Total | Jobs | % Total | Jobs | % Total |
| Total non farm | 32,407 | 100% | 144,306 | 100% | 111,899 | |
| Manufacturing | 10,100 | 31.2% | 12,348 | 8.6% | 2,248 | +2.0% |
| Durable goods | 5,261 | 16.2% | 7,719 | 5.3% | 2,458 | +2.2% |
| Non-durable goods | 4,839 | 14.9% | 4,629 | 3.2% | -210 | -0.2% |
Source: Bureau of Labor Statistics, Current Employment Statistics
This decline in manufacturing employment has various political, economic and social repercussions. I am not going to explore these here. Rather, I am interested in a narrower question: where do Americans work instead? If not manufacturing, what? And why does it matter?
Employment by Sector: The Big Picture (1940-2016)
How we answer that question depends on what data we use. Are we interested in employment only or also unemployment and people who are no longer looking for jobs? Are we interested in total jobs or in changes over time? How detailed do we want to be? Different approaches yield different answers.
Let's focus on actual employment and begin with the big picture: the relative decline in manufacturing has been totally offset by the rise in three sectors: professional and business services; education and health services; and leisure and hospitality. (These groupings only capture private-sector jobs; they exclude jobs at state and local schools, universities and hospitals—these are grouped under "Government.") The share of manufacturing jobs went from 31% in 1940 to 9% in 2016, a decline of 23 percentage points; by contrast, these three sectors jumped from 17% of total employment in 1940 to 40% in 2016, also a rise of 23 percentage points. In effect, the American economy has moved from manufacturing goods to education and health, professional and business services, and leisure and hospitality.
Employment by industry (as percent of total)
Source: Bureau of Labor Statistics, Current Employment Statistics
Employment by Sector: The Details (1990-2016)
But what exactly do these sectors mean? Is it education or health that is driving change? Leisure or hospitality? And what exactly are professional and business services? It is harder to answer this question—we don't have detailed sub-sector information back to 1940. But we have detailed data since 1990—and that data tell a clear story.
For context, the American economy added 34.8 million jobs between 1990 and 2016. Manufacturing lost 5.3 million jobs in that period, so overall, all other sectors added 40.4 million jobs. These three sub-sectors contributed 27.2 million of those jobs—so roughly two-thirds of all new jobs came from education and health, professional and business services, and leisure and hospitality.
Let's dig deeper into these sectors. The groups below are somewhat arbitrary—they don't follow a strict statistical convention and they show data at different aggregation levels. I have avoided groupings that are too vague to be meaningful ("professional services") as well as groupings so narrow that one misses the big picture ("offices of chiropractors"). These groupings are a balance between those extremes.
Employment by select sub-sectors (in millions)
Source: Bureau of Labor Statistics, Current Employment Statistics
These eight sub-sectors alone added 20.1 million jobs since 1990—which means that almost half the jobs created in America in 1990–2016 came from these eight groupings.
The largest single grouping is "food services and drinking," which added 4.9 million jobs—and that's most of the 6.3 million new jobs that fall under "leisure and hospitality."
Then comes ambulatory health care (+4.2 million), which means doctors' and dentists' offices, outpatient medical care, and so on. Social assistance (+2.5 million) includes caring for individuals, families, children or the elderly. Hospitals and nursing and residential care make up the remainder of health care. In all, health care added 9.7 million jobs—meaning that "education and health" is mostly health (+9.7 million), not education (+1.8 million)—at least not private education (more on public education below). It also means that a quarter of the ~40 million jobs that the American economy has added since 1990 have come from health care.
The other two categories listed above are employment services (largely temporary help services) and computer systems design (effectively, IT). Together, these amounted to 3.6 million of the 9.3 million new jobs categorized under "professional and business services." The other major sub-categories (not shown) include management consulting (+830 thousand) and architecture and engineering (+470 thousand). The other sub-categories are smaller and include legal services, accounting, bookkeeping, etc.
What about other sectors?
Government employment rose by 3.8 million—the growth coming mostly from local government (+3.4 million) and, less so, by state government (+784 thousand). In both cases, most of the jobs are in education (+2.7 million). Given that private education jobs rose by 1.8 million, this means that total education employment is up 4.5 million. The federal government, by contrast, shrank since 1990—with job losses across the board (except in federal hospitals).
Outside government, retail trade added 2.7 million across various sectors (motor vehicle dealerships, super markets, warehouse stores, drug stores, etc.). Financial activities added about 1.7 million jobs since 1990, with roughly equal contributions from investment products (+440 thousand), insurance (+544 thousand) and real estate (+450 thousand). And transportation and warehousing added another 1.5 million jobs, coming from trucking (+331 thousand); support activities for transportation (+296 thousand), largely related to air transportation; couriers and messengers (+267 thousand); and warehousing and storage (+509 thousand).
Interestingly, information has not shown much job growth. There are growing sub-sectors, to be sure: software publishers (+257 thousand); motion picture and sound recording (+165 thousand); data processing, hosting and related services (+88 thousand); and internet publishing and broadcasting and web search portals (+171 thousand). But the job losses in information have been large—especially, newspaper, book and directory publishers (-398 thousand) and telecommunications (-214 thousand). On balance, the information industry has created only 84 thousand direct jobs since 1990.
Conclusion
The move away from manufacturing has meant a transition towards services—chiefly health, education, professional services and leisure.
In terms of specific sub-sectors, of the 40.4 million jobs created since 1990, the main growth areas have been:
- Providing health care and social assistance (+9.7 million jobs).
- Eating and drinking out (+4.9 million).
- Private education (+1.8 million), mostly at colleges and universities; and public education (+2.7 million), mostly at the local level.
- Selling stuff to consumers (retail trade, at +2.7 million).
- Helping companies find (largely temporary) employees (+2 million).
- Managing money, selling insurance, and real estate services (+1.7 million).
- Designing and managing computer systems (+1.6 million).
- Moving around goods and, less so, people (+1.5 million).
- Management consulting (+830 thousand).
This list is not exhaustive, but, in effect, we have four major groupings: leisure and consumption, health care, intermediation (a loose way to group finance, insurance and transportation), and efficiency (a loose way to group education, IT, employment services and management consulting).
What does this mean? First: only a small share of the job growth has been in leisure and hospitality, essentially jobs geared to help us enjoy life more. That seems low to me, although one can take a broader view and include consumption (retail trade). Second: health care is a major employer. In fact, it is hard to understand the health care debate in the United States without appreciating that health care is, increasingly, what Americans do-how they earn their living and provide for themselves and their families. Third: there is a sizable segment of employment focused on making things run better—seen positively, this could be understood in the context of improving efficiency; or, more realistically, this is a response to sub-standard services and inefficient business practices—inefficiencies that should, ideally, disappear.
Such an economy seems fragile. What if the United States were to dramatically reduce the cost of health care? What if companies were managed better? Or computer systems were easier to design and operate? Or technology allowed employers and employees to find each other more easily? What if more and more purchasing goes online? What would Americans do then?
Probably they would move to the next thing that needed fixing. Yet one is left wondering: do people flock to the thing that needs fixing; or do people turning en masse to a new segment of the economy produce the kind of complexity that then requires more people to solve? Put differently: Does health care cost so much because so many people work in health care, or do people work in health care because that's where the money and the jobs are? Probably a bit of both. But does America have better health outcomes as a result of this massive influx of money and people? Probably not.
The same can be said for services such as IT and consulting, which exist to address complexity and/or inefficiency (and often fail to do so). Perhaps the same is partly true for education, where there is huge cost inflation, presumably linked to the number of people who now work in that field. Or what about finance, where there is a growing focus on creating more complex products that are, in theory, more tailored and sophisticated but, in practice, end up adding little to what already exists.
In my mind, these numbers underscore the kind of stress that the traditional social contract will come under as technology improves. What are the employment consequences of reducing health care or education costs? What if we decided that the financial sector is too big and too risky and had to be shrunk? What happens when technology makes even more tasks obsolete? What about better run organizations and computer systems—can we even afford to improve efficiency and thus wipe out the jobs that people have turned to after manufacturing? Can we afford to reduce complexity in the overall economy?
Of course, the economy will evolve. People will adapt. But look at the big picture and this adaptation has costs—the costs of complexity and often sub-optimal outcomes that come hand-in-hand with lots of people seeking employment in a few sectors. So much of the economy is premised on inefficiency and even redundancy. Is there another way? I don't know. But if this is what adjustment to a post-manufacturing world looks like, we have a long way to go to transition to a new economy that can generate full and meaningful employment for all.
Notes
For more information on the data sources and approach, please visit my github repository.