Why We Can’t Shun Manufacturing for the Service Sector


There’s been a lot of talk lately about the shift in the US economy away from production and increasingly into services. Consider the employment data from the US: In 1950, 30% of all US jobs were in manufacturing while 63% were in services. In 2011, 9% of total employment remains in manufacturing, 86% in services.

So does this signify a shift in consumers’ tastes from manufactured goods to services? The short answer is no; if anything, we consume more “things.” The difference is that things are manufactured with far less labor, and they are increasingly made somewhere else. The manufacturing industries still remaining in the US have seen tremendous improvements in productivity. Less-skilled work continues to flow out of the US, but the work that remains is higher-skilled, and more productive. Accordingly, the manufacturing jobs that remain in the US pay well.

Some look to the loss of US manufacturing jobs without concern: the future (they argue) is in service industries. As jobs disappear in manufacturing, others open in services like health care and retail. The problem is that as more manufacturing jobs leave, more productivity leaves as well.

Consider this: Classical economists saw productivity as the key in determining relative wages — the more productive the laborer, the higher his/her wages. Unlike manufacturing, service-sector jobs have strict limits in terms of productivity. For example, a live performance of Beethoven’s 5th requires the same amount of performers/employees as when it was performed early in the 19th century. Compare that with the production of almost anything manufactured — the number of workers now required to produce a bolt of fabric, for example.

So how is it that workers in service sectors, where productivity has relatively little growth, maintain wages competitive with workers in manufacturing, where productivity has done nothing but increase?

At least part of the answer lies in what modern economists have dubbed the “Baumol Effect,” after influential economist William Baumol. The Baumol Effect states that lower productivity notwithstanding, service industries have to pay wages comparable to manufacturing in order to get the workers it needs: it’s a simple matter of labor market competition.

So let’s put a little data behind this. The following table lists the 2010 national sales and employment numbers for 2-digit NAICS industry sectors, ranked in terms of total sales.

Sales (Millions)
Employment Rank
Manufacturing $4,444,349 12,116,153
Government $3,055,594 23,931,184
Finance and Insurance $2,335,933 9,276,170
Health Care and Social Assistance $1,671,158 18,983,244
Professional, Scientific, and Technical Services $1,482,841 11,711,344
Real Estate and Rental and Leasing $1,391,188 7,374,135
Retail Trade $1,194,951 17,369,914
Information $1,135,475 3,252,198
Construction $1,123,601 8,886,854
Wholesale Trade $993,673 6,071,136
Transportation and Warehousing $770,350 6,084,630
Accommodation and Food Services $691,475 11,872,079
Administrative and Support and Waste Management and Remediation Services $601,900 10,138,827
Other Services (except Public Administration) $502,463 8,872,041
Utilities $377,695 595,031
Management of Companies and Enterprises $376,055 1,935,179
Agriculture, Forestry, Fishing and Hunting $360,521 3,456,096
Mining, Quarrying, and Oil and Gas Extraction $355,246 1,410,588
Educational Services $260,555 4,080,407
Arts, Entertainment, and Recreation $208,984 3,780,900
Total $23,334,007 171,198,110
Source: EMSI Complete Employment, 4th Quarter 2010

When considering what industry sectors to prioritize for workforce and economic development efforts it is important to look beyond basic employment numbers. This is because, while a sector might have a lot of jobs, it might not actually be producing a lot of income for the region, which is also very important for overall economic health and vitality.

Sectors that generate more income per worker tend to have much bigger ripple effects, which means that a lot more people are impacted as a result of direct and indirect spending. The following table is organized by sales per worker, derived by dividing the total sales for an industry by total employment for a particular year.

Industry Sector
Sales Per Worker
Finance and Insurance
Mining, Quarrying, and Oil and Gas Extraction
Real Estate and Rental and Leasing
Management of Companies and Enterprises
Wholesale Trade
Professional, Scientific, and Technical Services
Transportation and Warehousing
Agriculture, Forestry, Fishing and Hunting
Health Care and Social Assistance
Retail Trade
Accommodation and Food Services
Administrative and Support and Waste Management and Remediation Services
Other Services (except Public Administration)
Educational Services
Arts, Entertainment, and Recreation
Source: EMSI Complete Employment, 4th Quarter 2010

Here’s our take on manufacturing and a few other basic observations that help to illustrate the difference between production and service sectors.

When it Comes to Income Manufacturing is Still King

At $4.4 trillion in total sales, manufacturing is by far the biggest income generator in our nation, despite a fairly rapid decline in employment (manufacturing has slipped to fourth in overall employment). Despite these trends, manufacturing still manages to far outperform all other industries in terms of pure income creation. Each individual that works in manufacturing generates roughly $370,000 per year. This is a very important fact to consider in a day and age when many folks advocate for improving the service sectors. 

Again, here’s the thing to note: sectors like manufacturing that generate more income per worker have much bigger ripple effects, creating much more impact in a region while helping to raise wages in lower-productivity service sectors. 

Government Services: High on Employment but Low on Productivity

The government sector is twice the size of the manufacturing sector (in terms of employment) but only produces $3 trillion in earnings or $130K in income per worker. Government is a bit trickier to analyze using the sales per worker criteria because the government is essentially capturing tax dollars and spending them on various services (education, military, infrastructure). Government can provide a lot of stability to regional economies, but it’s not really a growth industry (unless you’re in DC!).

Utilities and Finance – Low Employment but High Sales/Job Ratios

The utility and finance sectors have lower employment (ranked 8th and 21st, respectively) but rather large sales to job ratios (250K per worker and 650K per worker, respectively). Keep in mind, the utility sector has a lot of overhead and equipment that factor into the equation. There is a huge amount of capital in play in this sector that requires a relatively small workforce. Finance and insurance can generate very large amounts of capital, and they have much less overhead.

Health Care is Not a ‘Growth Industry’

Health care, the ultimate service sector, has become the second-largest employment sector in the country, yet it produces only $90K in sales per worker, which is pretty low compared to manufacturing, information, or finance. Basically, the health care sector is important for obvious reasons and it can be a source of good jobs for a local region, but it’s not really an “economic driver” that is going to propel our nation into greater prosperity.

Retail Trade vs. Information

The retail trade and information industry sectors have similar income generation ($1.19 trillion and $1.13 trillion, respectively), however, retail trade is five times the size of information in terms of employment. This is why every economic developer is looking for “the next Facebook” and not “the next Napa Auto Parts.” Retail trade only generates $70K per worker while information generates $350K per worker.

So what’s wrong with a service-based economy? It shrinks manufacturing employment as well as the manufacturing sector’s ability to prop up wages. A labor market that loses wage pressures of high-productivity manufacturing industries will settle at wage rates lower than markets where this wage-boosting effect is present. Economic development policy makers should be careful about shunning manufacturing or other production sectors in favor of service sectors.

Dr. Robison is EMSI’s co-founder and senior economist with 30 years of international and domestic experience. He is recognized for theoretical work blending regional input-output and spatial trade theory and for development of community-level input-output modeling. Dr. Robison specializes in economic impact analysis, regional data development, and custom crafted community and broader area input-output models. Contact Rob Sentz with questions about this analysis.

Illustration by Mark Beauchamp

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Land prices matter very much

Good article.

Another important factor to consider, is what land use policies do to urban land prices.

The ratio of GDP to total land value, like the median multiple house price, falls into historic norms. If the urban land values and house prices are inflated completely independently of incomes and GDP, the results for the economic competitiveness of that city or region, are strongly negative.

The actual cost to businesses, of land rent, is higher. The workforce cost of living is higher, with pressures on wage demands. Local discretionary spending on non-housing related items will be reduced. "Churn” of land to more efficient uses, is reduced. The rate of new business startups, a major source of employment growth, is reduced. Increased incumbency advantage leads to oligopoly effects and loss of potential efficiency gains. Capital is diverted from productive uses, to land price bubbles. The actual location decisions made by households and businesses are forced to become LESS efficient (in contradiction of the intentions of the urban planners). Participation in agglomeration efficiencies is reduced; potential participants are locked out by high land prices. The ability to regenerate economic activity after major negative events, is reduced.

In the long term, we can expect inter-city competition to be "decided" heavily on favour of those cities that maintain low land prices through relaxed limits in development, and allowing mixed land uses. I expect even developing nations like China and India to ultimately suffer a serious lack of competitiveness compared to the "Houston" and "Atlanta" style city economies of the USA, if they do not address their serious distortions of land values that occur through heavy handed regulation and corruption.

This is "good news" for the regions of the USA that have not succumbed to the "Smart Growth" suicide. Low urban land prices and auto-mobility ARE a major secret to US economic strength; the "Smart Growth" cities are choosing to throw away this advantage. Responsible politicians who CLAIM to care about loss of industry offshore, and rising pressures on low income groups, are the worst kind of hypocrites if they also support regulations that inflate the price of land in their city.

By the way, the clue to look for, is "discontinuity" in the price of raw land at the urban fringe. This price should NOT "step up" by a factor of 10 or more from "rural" to "urban"; it should have a smooth slope when graphed.

Are you sure that you are

Are you sure that you are properly accounting for the employment base for manufacturing? You properly identify that utilities have a high overhead, which of course includes the purchase of raw materials, but you don't seem to make that same connection for manufacturing.

If I'm a manufacturer and I have 9 employees (so 10 workers, total, including myself) and I sell $5 million worth of widgets, I've generated $500,000 in sales per employee. But I have to then buy steel and electronics and automation equipment from some other supplier - I, like most manufacturers, am not completely vertically integrated from raw materials to the customer. So, really, I also have indirect "employees" involved in making the components I use. And, as often as not, those "employees" are overseas and don't get included in my productivity estimates, or in my sales-per-employee numbers.

Finally, what is really most important, isn't sales per employee, but net profit per employee. As many manufacturers in the Rust Belt discovered, it doesn't do anyone any good to have enormous sales per employee if you aren't capturing any of that as profit. Profit is much less assured when, in addition to employee costs, you have enormous capital requirements and enormous fixed costs, as most manufacturers do. That's the primary difference between manufacturing and service jobs - the profit relative to the fixed capital involved. In the case of financial services, much of the capital involved is cash, entirely liquid cash, which further complicates a comparison of the often purpose-built machinery or commodity capital required in manufacturing.

All that said, I agree that it's important not to marginalize manufacturing. I just don't happen to think that their sales per employee is very relevant - at least not as relevant as you're making it out to be. And if it were relevant, you'd have to capture the real employee base to actually make a fair comparison to services.