Cover Image

PAPERBACK
$54.00



View/Hide Left Panel

TABLE 3-1 Data on Six Industries, 2002 (except where indicated)

 

Computer Systems Design and Related Services NAICS: 5415

Software NAICS: 5112

Semiconductors NAICS: 3344

Automobiles NAICS: 3361-3363

Construction Engineering/ Services NAICS: 23, 3413

Pharmaceuticals NAICS: 3254

PC Manufacturing NAICS: 3341

Total for 6 Industriesa

U.S. Total

Value-added ($ billions)

173.5

103.5

110.4

469.8

1,358.4

140.6

73.7

2,429.9

10,469.6

Employment

1,107,613

356,708

437,906

1,078,271

8,459,885

248,947

150,751

11,840,081

114,135,000b

R&D performed ($ billions)

11.9

12.9

11.9

16 (est.)c

10.7

10.1d

3 76.5

193.9

 

R&D scientists and engineerse

90,800

80,800

73,000

83,200

n/a

51,800

15,100

394,700

1,066,100

aThe software industry is represented by two NAICS codes, 5414 and 51112, which clearly do not map exactly onto the industry sectors covered in the commissioned papers, particularly for software (figures here understate the revenue, employment, and R&D of interest) and PC manufacturing (figures here overstate the revenue, employment, and R&D of interest).

bTotal private sector employment.

cIn recent years, the auto industry R&D total has not been reported by NSF because it would disclose the total for an individual firm. $16 billion is a rough estimate obtained by subtracting the R&D performed by the aerospace industry from the total R&D for the transportation equipment sector.

d2001.

eR&D scientists and engineers is not an ideal proxy for the population we are interested in, but this data is collected by NAICS code and allows an apples to apples comparison. Note that Moavenzadeh (this volume) gives an estimate of 189,000 engineers in the auto industry for the relevant NAICS codes. Sources: Bureau of the Census, 2004 (for value-added and employment); NSB, 2006 (for R&D performed); and Hecker, 2005 (for R&D scientists and engineers).

began to come down. For example, Broadcom, a software-intensive semiconductor company, reports that its team in Bangalore is now as productive as its teams in San Jose and Irvine, with costs in India running about one-third of those in the United States.

As the institutional infrastructure in India has improved, offshoring has become part of the normal way of doing business in the software industry. The diaspora of U.S.-educated Indian entrepreneurs has helped fuel the growth of the Indian tech sector, which is developing in a way that complements Silicon Valley (Saxenian, 2006). One example cited by Dossani and Kenney is Netscaler, a company that turned to offshoring when it was facing a funding crunch. The tactic enabled the firm not only to survive, but also to grow (both in India and the United States).

Aspray et al. (2006) observe that offshoring has become essential to the globalization of the software industry and will undoubtedly continue and increase. In Dossani’s workshop presentation, he reported that today, in Indore, which is not a large IT center like Bangalore or Mumbai, wages for engineers who work 12 hours a day, six days a week are about $200 a month. However, in larger centers like Bangalore, salaries for experienced engineers are rising rapidly. For example, in a 2006 survey, “State of the Engineer,” published in EE Times, the mean salary for Indian respondents was $38,500. However, as the history of Silicon Valley shows, higher costs are not necessarily a barrier to innovation-fueled growth (Saxenian, 2006). Despite very high costs for skilled labor, Silicon Valley has remained a prime location for innovative start-ups.

In a workshop presentation, Alfred Spector, a consultant and NAE member, outlined three possible scenarios for the future of software-development offshoring (Spector, this volume). In the first scenario, offshoring frees up U.S. talent and money, which can then be focused on higher value-added activities, such as testing, which then becomes much more efficient. In the second scenario, the rise of India and other offshoring destinations in certain sub-disciplines leads to a loss of U.S. jobs in those sub-disciplines, but, again, frees up talent and other resources for the creation of new sub-disciplines or super-disciplines that keep U.S. software innovation strong overall. In the third scenario, when U.S. students learn that certain activities are being moved offshore, they conclude that opportunities for software innovation in the United States are drying up and decide not to pursue careers in those areas; this leads to atrophy in the U.S. talent and skills base.1

The three scenarios are not mutually exclusive—the United States might maintain its leadership position in some aspects of software but lose it in others. Spector says that the

1

One reviewer of this report suggested tracking metrics related to software innovation over time to determine which of these scenarios is being realized.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement