The GAO study showed that U.S. imports from India of business, professional, and technical (BPT) services as published by BEA are substantially lower than India’s data on exports of BPT services to the U.S. (chart 1, left panel).1 However, when adjusted to a similar conceptual basis using information from the GAO report, the difference is actually quite small (chart 1, center panel).

The large gap between the U.S. and Indian data mainly reflects differences in how BEA and India define BPT services. BEA data are consistent with international standards for balance of payments accounting; India’s data, which are based on data from an Indian trade association, do not conform to international standards.  In fact, a 2005 study published by the Reserve Bank of India (RBI) showed that computer services exports (a large component of BPT services) to the U.S. based on international standards are much lower than India’s published data (chart 1, right panel).2 In addition, a 2004 report by the OECD found that 97 percent of India’s exports of computer services to large OECD member countries were unaccounted for in those countries’ data on imports.3

Depending upon how one adjusts for important definitional differences, the gap between the U.S. and Indian estimates either entirely disappears or is substantially reduced.

Chart 1: U.S. and Indian Data on Trade in BPT Services, 2002

Chart 1: U.S. and Indian Data on Trade in BPT Services, 2002

Source: GAO; calculations by BEA

1 Government Accountability Office, “U.S. and India Data on Offshoring Show Significant Differences,” October 2005.

2 Reserve Bank of India, “Computer Services Exports from India: 2002-03,” Reserve Bank of India Bulletin, September 2005.

3 Organization for Economic Co-operation and Development, Information Technology Outlook 2004, October 2004.



 

What are the reasons for differences between U.S. and Indian data on trade in business, professional, and technical services?

The U.S. and Indian data on trade in business, professional, and technical (BPT) services are not directly comparable because of substantial definitional differences.  When the U.S. and Indian data are adjusted for definitional differences, the difference in estimates either entirely disappears or is substantially reduced.  Some major definitional differences are:

      Indian workers in the United States.  India’s data on trade in BPT services include services provided by Indian nationals who reside in the United States.  BEA follows international standards for balance-of-payments accounting by excluding the compensation paid by U.S. firms to U.S. residents.  Foreign workers who are in the United States for less than one year are considered to be foreign residents, and typically their earnings are included as compensation of employees (under “income” in the balance of payments accounts).  Workers who are in the United States for more than one year are considered to be U.S. residents, and so their earnings are excluded from the balance of payments accounts.  According to the GAO study, Indian officials acknowledged that temporary Indian workers in the U.S. have accounted for about 40 to 50 percent of their data on exports of BPT services.

 

b)      Sales through affiliated companies.  India’s data on services exports to North America include sales of services to affiliates of U.S. companies located in India or another foreign country, as well as sales by affiliates of Indian companies located in the United States to other U.S. residents.  According to international standards, BEA excludes these sales from U.S. trade in services because the transactions did not occur between a U.S. resident and a non-resident.  A U.S. company’s foreign affiliate that is located in India is an Indian resident, and so its transactions with other Indian residents should not be included in the balance of payments.  Similarly, an Indian company’s affiliate in the United States is a U.S. resident, and so its transactions with other U.S. residents should not be included.  According to the GAO study, an Indian official stated that inclusion of sales to affiliates of U.S. companies is “likely a significant factor” accounting for differences between U.S. and Indian data.

 

c)      Sales of goods.  India’s data on trade in BPT services include some sales of goods, such as prepackaged software and software embedded on computer hardware.  The U.S. data on trade in these products are included in the goods trade data, not in the services trade data.  According to the GAO study, Indian officials stated that embedded and prepackaged software account for about 10-15 percent of India’s estimate of exports of BPT services to the U.S.

 

d)      Sales of technology-enabled services.  India’s data on trade in BPT services include some technology-enabled services (such as some financial services).  BEA includes these services in other services categories.

 

e)      Intrafirm trade.  Through 2006, U.S. data for trade in services are collected separately for cross-border trade between unaffiliated companies and for intrafirm (or affiliated) trade.  The surveys that BEA uses to collect data on unaffiliated trade are detailed enough to allow BEA to identify trade in BPT services vis-à-vis India.  Affiliated trade, however, is collected on separate surveys, and data for individual foreign countries that separately identify BPT services are unavailable.  Therefore, reported BEA data for BPT trade with India cannot be directly compared with the Indian data, because BEA’s data for BPT services include only unaffiliated trade and India’s data on BPT services include both affiliated and unaffiliated trade.



 

How did BEA calculate the adjusted data shown in the chart above?

BEA adjusted both its own estimates and the Indian estimates to eliminate definitional differences between U.S. and Indian data.

BEA adjusted its own data to include an estimate of affiliated transactions, which are collected on surveys that do not allow for BPT services to be separately identified by individual foreign country. In order to estimate affiliated imports of BPT services from India, BEA used a ratio calculated from global affiliated and unaffiliated imports of BPT services. BEA used the same procedure to estimate affiliated imports of computer services (table 1).


Table 1: Adjustments to BEA’s data, 2002
[Millions of dollars]
 

    BPT Services Computer Services
       
Published BEA estimates (based on reported data)    
       
a. Global imports 33,488 4,315
b.    Affiliated imports 23,940 2,800
c.    Unaffiliated imports 9,548 1,515
d. Ratio [b/c] 2.51 1.85
       
e. Unaffiliated imports from India  288 201
       
Implied estimates (derived from global ratios)    
       
f. Affiliated imports from India [d*e] 722 371
       
g. Total imports from India [e+f] 1,010 572


Source: BEA.


Data from India on BPT exports to the U.S. were adjusted to remove the definitional differences with BEA data and international standards as described above. BEA based the adjustments on information provided in the GAO study. Not all differences were quantified in the study, so some differences remain (table 2).


Table 2: Adjustments to India’s BPT services data, 2002
[Millions of dollars]
 

    High estimate Low estimate
       
I. Exports of BPT services (published by India; chart 1, left panel) 6,464 6,464
       
II. Adjustments for definitional differences (derived from GAO report):    
  a. Indian workers in the U.S. 40% 50%
  b. Sales through affiliated companies 20% 30%
  c. Sales of goods 10% 15%
       
III. Total [a+b+c] 70% 95%
       
IV. Adjusted Indian estimate of exports of BPT services [I*(1-III)] (chart 1, center panel) 1,939 323


Source: GAO; calculations by BEA.


The ranges for adjustment factors (a) and (c) were provided by Indian officials and published in the GAO study. The effect of factor (b) was estimated by BEA at 20-30 percent, because the GAO study said that Indian officials thought the effect of this factor was “significant” but that the effect of factor (c) was “insignificant” at 10-15 percent.

The September 2005 RBI study on Indian sales of computer services provided data that corrected for some definitional differences but the estimates still were not directly comparable to U.S. estimates because the Indian estimates were presented based on trade negotiation categories (i.e., GATS rules) that included sales by Indian-owned companies in the United States to U.S. residents, sales by Indian workers in the United States, etc. BEA used data from the RBI study to calculate a ratio of Indian exports of computer services using balance of payments definitions to total (broadly defined) deliveries of Indian computer services using GATS rules at the global level; this ratio was 39%. In addition, the study provided data on the delivery of Indian computer services (broadly defined) to North America. BEA used the global ratio to estimate balance-of-payments basis exports of computer services to North America. BEA then estimated the portion of exports to North America attributable to the U.S. using information from an Indian software trade association (table 3).


Table 3: Adjustments to RBI’s Computer Services Data, 2002
[Millions of dollars]
 

    India's exports of computer services
a. Global delivery of services - broad definition 31,133
b. Global exports - balance-of-payments basis 12,077
c. Ratio [b/a] 0.39
     
d. Delivery of services to North America - broad definition 4,046
e. Exports to North America - balance-of-payments basis [c*d] 1,569
     
f. Ratio (U.S./North America) 0.82
     
g. Exports to United States [e*f] (chart 1, right panel) 1,287


Source: RBI; NASSCOM; calculations by BEA.

 

Published