EDITION 2 - June, 2006
Bruce Sundquist

Copyright: None.

Reference citation format, e.g. (98C2), cites a document published in 1998 by a lead author whose last name begins with "C". The final integer (2 in this example) is a running index.
Previous Editions: Ed.1 July 2004 // 

(8) -- SUMMARY

Go (Return) to "Globalization: The Convergence Issue" - Go (Return) to the home page of this web site


Public concerns have been growing about the "outsourcing" of jobs, mainly to developing nations where labor costs are on the order of 90% less. These concerns have recently been countered by arguments (using US Department of Commerce documents) that the "insourcing" of jobs, over 90% from developed nations, has created more US jobs than outsourcing has eliminated. Insourcing numbers are simply taken as the number of US employees of US non-bank "affiliates" that are at least 10% owned by foreign entities, e.g. 6.5 million in 2000. These Department of Commerce documents have been examined in detail. It was found that only a negligible percentage of these employees could legitimately be counted in estimates of insourcing. There are two main reasons for the reductions in the insourcing count. First, much insourcing is associated with de facto outsourcing. Second, the significantly expanded US trade deficit associated with US affiliates of foreign entities equates to a significant increase in de facto outsourcing. Neither of these effects has been taken into account in the original values of insourcing.

In the past few years a lot of data has come out on the number and types of US jobs that are being "outsourced" to countries where labor costs are typically 10% of those in the US. Some of these data are quoted in a companion paper (06S1) (Section (2-A)). Data on the apparent effects of outsourcing in terms of pressures on wages, benefits, layoffs, severance, welfare benefits, part-time and temporary employment, savings, debt, bankruptcies, stress, poverty, economic polarization, workforce dropout rates and other indicators of convergence are also given in that paper (06S1) (Section (3-A)). There is widespread agreement that rates of outsourcing, particularly of higher-paying, higher-skilled jobs are bound to increase markedly, even during just the next few years. Below is a summary of outsourcing data.

Beginning around early 2004, articles began appearing in the mass media contending that "insourcing" - allegedly moving jobs from outside the US into the US - is larger than "outsourcing". This may appear hard to believe, but there is a basis for believing it - a deceptive basis as will be seen below. It has become clear that the effects of insourcing on jobs in the developed world need to be examined more carefully. A major barrier to doing this is that insourcing and outsourcing are typically defined and computed in entirely different ways, and reflect different phenomena, making the two quantities difficult to compare directly and meaningfully. Here we take care to define and compute insourcing and outsourcing carefully and in such a way that the two can be compared meaningfully. When this is done, as will be shown below, outsourcing significantly exceeds insourcing - quite the opposite of what one commonly hears today. One example that serves to show how the new analysis obtained the grossly different result is the following. In the comparisons of insourcing and outsourcing one typically reads about in today's media, when a foreign company buys an American company, the US employees of that company are computed to be "insourced" when in reality virtually nothing has changed, job-wise. The analysis below does not make that error, or several other key errors that the more traditional insourcing-outsourcing comparisons make.

Bear in mind, also, that insourcing and outsourcing are just two of many aspects of the overall effect of globalization on the jobs/ wages/ benefits/ working-conditions/ human-capital-formation/ retirement-savings picture in the developed world. Much of the trade deficit is effectively another component of outsourcing that the commonly cited data on outsourcing do not count. Nor do those touting the benefits of insourcing count it. The US trade deficit in 2003 was $490 billion. The labor content of US exports is about 59%, indicating a net outflow of wages of at least $290 billion/ year - and far more than that if the factor-of-ten difference in wage rates between the developed- and developing worlds is considered. For an estimated labor cost (wages + benefits) of $50,000 per year per US job, this trade deficit suggests an outsourcing of 5.8 million US jobs. Also, even minor expansions of the pools of unemployed and under-employed wage earners in the developed world puts downward pressures on jobs, wages, benefits, working conditions, human capital formation and retirement savings. The effects of these pressures ripple throughout the entire job market and the entire economy, even affecting people who do not compete head-to-head with developing world wage-earners (Section (2-A)[3] of Ref. (06S1)).


Table 1 -- Foreign-based, non-bank companies' economic activity in the US (from graph) (Wall St. Journal, 4/04/03) (Source: U.S. Department of Commerce, Bureau of Economic Analysis) (Total assets are in $trillions; total jobs are in millions)





































Table 1 provides the basis of arguments for the benefits of insourcing found in numerous articles in the mass media and advocacy documents by lobbying organizations, including:

The number of jobs insourced into the US is taken simply as the number of jobs in US affiliates of foreign entities (US affiliates that are majority-owned by foreign entities or by foreign entities with as little as 10% foreign ownership) (98U1). Calculations of US-located assets in this category and the associated US jobs are summarized in Table 1 above. The overwhelming bulk (90+%) of these foreign-based entities are headquartered in the developed world (98U1). To understand the deception inherent in these numbers of insourced jobs, note that these US affiliates have four main effects on the US-jobs-and-trade picture:

  1. They produce goods and services for local (US) consumption in the former US-owned facilities that the foreign entities bought out. Basically all that has happened is buyouts that result only in moving former US headquarters offshore. The employees of these US affiliates are now counted as "insourced" even though the number of jobs in the US remains unchanged. In order to avoid deception one ought to set up an account labeled something like "de facto outsourcing" with the same number of people in it as the number of people that have been labeled "insourced" as a result of moving headquarters around.
  2. They produce goods and services for local (US) consumption but merely by displacing US-owned companies. For example, when German, Japanese and Korean automakers establish assembly plants in the southern half of the US, US-owned automakers in the northeast are forced to shut down or move offshore. The establishment of such operations invests additional capital in the US, but sales are limited by the number of customers, not the availability of financial capital since financial capital is readily available in the developed world. Since the added capital does not increase purchases by customers, neither does it produce a net increase in number of jobs. However many employees of this category of US affiliates are labeled as "insourced" require about the same number of people to be added to the account labeled "de facto outsourcing".
  3. They produce goods and services for local (US) consumption but decrease the foreign entity's imports by the same amount as now produced locally in the US. For example, suppose Honda had a Japanese facility that produces and exports automobiles to the US. If Honda then decided to shut this facility down and build a facility in the US producing the same number of automobiles, the new US employees could be counted as "insourced". There would be no need to add a corresponding number to the "de facto outsourced" account because Americans would be buying the same number of Hondas locally as they previously imported. If Honda decided to expand its operations by building a plant in the US, these employees would add to both the "insourced" and the "de facto outsourced" accounts since total US automobile sales would remain approximately unchanged.
  4. They change the US trade deficit by altering the amount of importing and exporting they do relative to the amounts done by US-owned facilities. For example, foreign manufacturers may normally set up assembly plants in the US and import parts for it, whereas US companies of the same type might normally manufacture parts and also do the assembly locally - or vice-versa. Any resulting change in the US trade deficit would change the number of jobs associated with this change.

Those touting insourcing as effectively a reduction in the number of jobs outsourced simply use the numbers in Table 1 above. They invariably ignore the "de facto outsourcing" associated with Effects (1) and (2), and they also ignore Effect (4). Thus their figures should probably not be taken seriously. Our task now is to attempt to estimate better numbers for insourcing, taking into account all four of the above-mentioned effects.

(4) -- ANALYSIS - Effects (1), (2) and (3)
This paper does not correct all the insourced jobs figures in Table 1. Instead it analyzes only 1997 data. The magnitude of the corrections to be applied to the remaining data will then become clear. The year 1997 was chosen because of a huge compilation (200 pages of tables) of US Department of Commerce data available for 1997 (98U1). Below are some key data obtained from or derived from this reference.

Assets of non-bank US affiliates of foreign entities*

$3070 billion #

Employees associated the above assets

5.20 million #


$1726 billion


$ 141 billion ( 8.2% of sales)

US Sales (by difference)

$1585 billion (91.8% of total)

Employees associated with exports (8.2% of total)

0.43 million

Employees associated with US sales (91.8% of total)

4.77 million


$ 265 billion

# Agrees with Table 1

* Either majority-owned by foreign entities or with at least 10% owned by foreign entities.

The data above show that 91.8% of the operations of these US affiliates of foreign owners are associated with selling to customers within the US. The 4.77 million employees related to these sales are thus to be associated with Effects (1), (2) and (3) above. The change in total US employment from Effects (1) and (2) is essentially zero. What fraction of the 4.77 million employees is to be associated with Effect (3) is unknown, and the data needed to estimate this fraction are apparently not available. Effects (1) and (2) are clearly not minor, and would seem to be larger than Effect (3). If that were the case, Effect (3) would have to be less than 1.6 million, i.e. 31% of the employees of US affiliates of foreign entities.

The remaining task is to estimate the magnitude of Effect (4). This issue is taken up below.

(5) -- TRADE BALANCE EFFECTS - Effect (4)
Effect (4) is impossible to calculate directly. The best that can be done is to compare the import/ export data of these US affiliates with the corresponding averages for all US corporations. The assumption inherent in this calculation is that the average US affiliate of a foreign entity has the same characteristics as the average US corporation of the same size. This is not quite true, especially in that US affiliates are somewhat more involved in manufacturing (98U1). The table below gives the Department of Commerce data (98U1) needed to carry out the analysis for 1997.

Total Exports from the US (1997)

$ 910 billion

Total Imports into the US (1997)

$1020 billion

US labor force (est.)

136 million

Exports from the US per member of the US labor force


Imports into the US per member of the US labor force


Exports associated with 5.2 million average US employees

$34.8 billion

Imports associated with 5.2 million average US employees

$39.0 billion

Trade deficit associated with 5.2 million average US employees

$ 4.2 billion

Exports associated with US affiliates of foreign entities

$141 billion *

Imports associated with US affiliates of foreign entities

$265 billion *

Trade deficit associated with US affiliates of " "

$124 billion

Increase in US trade deficit due to US affiliates (124-4.2)

$120 billion

Job loss associated with this trade deficit increase #

2.0 million

Job loss as a percent of the employees of US affiliates


* From the table above.
# Assuming $50,000/ year in annual trade deficit = one US job and assuming that 0.4 million employees of US affiliates are associated with export-related operations (See basis for this in previous table).

Effect (4) appears to have a significant negative effect on US employment. This result needs to be understood. What is clearly happening, based on the data and analyses given above, is that foreign owners are moving their assembly operations into the US mainly to be closer to a major customer base. (Note the very high percentage (91.8%) of sales to US customers.) This makes sense. It is a lot cheaper to import parts for automobiles into the US and assemble autos for US consumption in the US than it is to import far bulkier assembled autos into the US. This is especially true for countries like Germany and Japan where wages are higher than in the southern US and where land prices and corporate tax rates are higher than in the US. (Germany and Japan probably farm out some parts manufacture to developing world nations.) The very high import values for US affiliates (relative to the US corporate average) noted above could be explained by postulating a huge value of all those parts being imported. The US companies that have been "displaced" by foreign-owned companies would normally make parts and also do the assembly in the US. Note that the 1997 trade deficit associated with US affiliates of foreign owners ($124 billion) is actually somewhat larger than the entire 1997 US trade deficit ($110 billion). This suggests that if, somehow, the trade deficit due to US affiliates of foreign entities were to be eliminated, the US would have run a small trade surplus in 1997.

Conclusions: The contribution of Effect (3) is impossible to calculate with any accuracy. The net influences of Effects (1) and (2) on insourcing are zero as noted above. The surmise (above) that the Effect (3) is not likely to be more than a third of the total number of US employees of US affiliates of foreign entities involved in sales within the US suggests that Effect (3) could be no more than 31% of all employees of US affiliates of foreign entities. On the other hand, Effect (4) is found to be a net job loss of on the order of 38% of the total number of employees of US affiliates of foreign entities. This would suggest that the real insourcing number for 1997 is about -7% (31% minus 38%) of the total number of employees of US affiliates of foreign entities. A large uncertainty is to be associated with this figure. However it seems apparent that the estimates of insourcing touted in the mass media in recent months are extremely deceptive. Since total outsourcing in the US is on the order of 8 million (See Summary below), an insourcing figure of a small fraction of 5.2 million (or 6.4 million in 2000) clearly shows that outsourcing remains a serious problem in the US, and that the media attention to insourcing only gives us a false and dangerous sense of security. Some additional serious problems associated with insourcing are examined below.

There is also a conceptual problem in comparing the insourcing numbers of Table 1 with outsourcing numbers such as those given in the Introduction. Table 1 pertains to financial capital and jobs sloshing about mainly (90+%) within the developed world. This is "Type A" globalization as defined in Section (1-D) of the companion paper (06S1). Were Type A globalization to run to complete convergence (as it probably will) in the absence of Type B globalization (involving commerce between developing- and developed nations), the developed world would be little changed. US labor costs are lower than those in some developed world countries and higher than in others. As these cost gradients are reduced in the convergence process, the US could expect small changes in wages, benefits and working conditions. The situation is totally different from that for Type B globalization, which is what most outsourcing concerns are all about. Type B globalization also produces convergence of the economies of the developed and developing worlds. (See Chapter 2 of Ref. (06S1)) This convergence process, however, entails developed world labor costs falling by 80-90% and huge social, economic and political instabilities (See Chapters 4, 5 and 6 of Ref. (06S1)). So any contention or implication that insourcing from Type A globalization will somehow counterbalance the outsourcing due to Type B globalization cannot be credible.

Let us say that, despite the conceptual problems, some fraction of the insourcing jobs listed in Table 1 might effectively counterbalance a like number of outsourced jobs noted in the Introduction and in the Summary below. One then needs to ponder the futures of both insourced jobs and the residual (non-counterbalanced) outsource numbers. Table 1 shows only a modest insourced "jobs" growth rate. On the other hand, the trade deficits of the US, and probably the remainder of the developed world, double every 4-5 years. The rates of real growth of exports during 1985-1996 were as follows: Developing economies: 217%, World: 94.2%, Industrialized economies: 69.6% (DRI/McGraw Hill data, Wall Street Journal 2/24/97). These variations in export growth rates translate directly into growing trade deficits, and trade deficits translate directly into outsourcing. Outsourcing produces unemployment, and unemployment puts pressures on jobs, wages, benefits, working conditions, human capital formation and retirement savings -- even of people who do not compete head-to-head with wage earners in the developing world. Reduced rates of human capital formation and retirement saving create further rounds of unemployment and the entire above list of pressures on those remaining employed. Increasing needs to prop up sagging economies puts downward pressures on interest rates. Falling interest rates shrink the retirement earnings from social security, IRAs and 401Ks and also the pre-retirement size and growth rates of these accounts.

(8) -- SUMMARY
Below is a tally of the various effects of globalization on US job losses based on the above.

Trade-deficit-based job losses (@$50,000/job) #

5.8 million

Outsourcing *

2.5 million

Total (out of 136 million US jobs)

8.3 million

* Textiles 700,000, Information technology and related services 400,000, Manufacturing 1,000,000 (since just 2001), White collar 440,000.
# Includes the 2 million trade-deficit-related jobs lost via Effect (4).

Keep in mind that:

U.S. Dept. of Commerce, Bureau of Economic Analysis, "FDIUS: 1997 Benchmark Survey", (1997) 205 pp., 3.321 MB, http://www.bea.gov/di/FDIUS97tables.pdf (visited 7/21/04).
02P2 Blair Pethel, "Bush pitches elimination of trade tariffs", Bloomberg News, in Pittsburgh Post Gazette, 11/26/02.
03D2 Bob Davis, "With Software Jobs Migrating to India, Think Long Term", Wall Street Journal, 10/6/03.
03S3 Michael Schroeder, Timothy Aeppel, "Skilled Workers Mount Opposition to Free Trade, Swaying Politicians", Wall Street Journal, 10/10/03.
03U1 (Unknown) "North America: Jobs Move Offshore as Firms Continue to Economize", New Haven Register, 4/14/03.
03V1 Gonzalo Vina, Tom Mudd, "Call Centers Migrate to India, and North of England Loses Jobs", Wall Street Journal, 11/5/03.
04H2 Jon E. Hilsenrath, "Behind Outsourcing Debate: Surprisingly Few Hard Numbers", Wall Street Journal, 4/12/04, p. A1.
04M2 Kris Maher, "Next on the Outsourcing List", Wall Street Journal, 3/23/04, p. B1.
04S1 Michael Schroeder, Joseph Rebello, "U.S. Survey Finds Few Jobs Moving to Offshore Homes", Wall Street Journal, 6/11/04, p. A2.
06S1 Bruce Sundquist, "Globalization: The Convergence Issue", Ed. 13, March 2006, 640 KB. Available as WORD 2000 files upon request.