This year’s Budget has received a negative response from the IT industry which was expecting an early assurance that the substantial tax benefits it has been receiving for the last decade would be extended for a further period. Those benefits have been huge indeed, with this year’s Budget estimating the “effective tax rate” on profits of IT-enabled services providers (ITeS) and business process outsourcing (BPO) firms at 7.36 per cent and software d evelopment agencies at 6.38 per cent. This compares with the statutory tax rate of 33.66 per cent that would have been paid if the concessions had not been provided.

The argument of the industry is that given the dramatic increases in output, exports and employment it has delivered this benefit should continue. While the rapid growth of the industry cannot be denied, there is increasing evidence that the size of the industry’s revenues and contribution to GDP is exaggerated, because there is no proper effort being made to independently evaluate the industry’s performance.

The two sources normally quoted when defining the industry’s performance are Nasscom and Dataquest, both private agencies which can hardly be described as disinterested parties. In its strategic review for 2008 released recently, Nasscom estimates IT industry revenues at $48 billion in 2006-07 and $64 billion in 2007-08. Dataquest, on the other hand, has estimated the industry’s revenues at $50.4 billion in 2006-07, which is not too far removed from the Nasscom estimate for that year. But this estimate is arrived at by converting estimated rupee revenues (Rs 2268.79 billion) to dollars at a conversion rate of Rs 45.05 per US dollar.

However, the official Economic Survey for 2007-08 estimates the average exchange rate for 2006-07 at Rs 42.25. If this rate is used the dollar revenue works out to $53.7 billion. This makes the Dataquest estimate as much as 12 per cent higher than the Nasscom estimate.

The only official source of data is the figures on exports of computer services that were estimated by the Reserve Bank of India for 2002-03 (Reserve Bank of India Bulletin, September 2005). The RBI’s estimates of exports of software services are influenced by those from Nasscom. Noting that “although RBI collects data on software exports through Software Exports (SOFTEX) forms, it uses Nasscom data as a controlling total for gross receipts from software exports”, the National Statistical Commission (2001) advised the RBI “to re-examine the current methodology on collection of software exports data”.

Responding to that request the RBI constituted a Technical Group for that purpose, which undertook a rather comprehensive survey (in the nature of a census) to estimate the value of exports in 2002-03 of “Computer Services” as defined in the GATS Manual of Statistics on International Trade in Services (2002).

The “Computer Services” category includes database services, such as development, storage, and access to online time series, data processing (including tabulation, provision of processing services on a timeshare or specific (hourly) basis, and management of facilities of others on a continuing basis); hardware consultancy; software implementation (including design, development, and programming of customised systems); and maintenance and repair of computers and peripheral equipment.

The survey did not include ITeS and BPO services (such as call centres, medical transcription, back office operations, revenue accounting and other ancillary operation, insurance claims processing, legal databases, content development/animation, payroll logistics management), which are otherwise included in IT-related services because firms providing these two kinds of services are often the same. These exports are also covered in the RBI’s figures on exports of “Software services”.

The RBI received responses to its questionnaire from 1,036 companies, which included all the top IT companies. Of these, as many as 912 firms were involved solely or partly in computer services provision. It must be noted that in March 2003 there were around 840 IT software and services companies that were members of Nasscom, which, according to Nasscom, accounted for around 95 per cent of the revenues of the software industry in India. Going by this, the coverage of the RBI survey must have been near total.

According to the RBI’s survey, exports of computer services in 2002-03 amounted to Rs 30,880 crore, of which, Rs 26,020 crore was to companies and units other than subsidiary(s)/associate(s) abroad and Rs 4,860 crore was the billing by Indian exporters to their subsidiary(s)/associate(s) abroad.

In that year, Nasscom reported exports of IT services and software to be $7.1 billion, which when converted at the average annual exchange rate of the rupee vis-À-vis the dollar (Rs 48.395), amounts to Rs 34,360 crore. Thus, despite its smaller coverage in terms of number of firms, the Nasscom figure exceeds that of the RBI by as much as 11.3 per cent.

This difference does not come through when we consider the aggregate figure for exports of software services as reported in the RBI’s balance of payments statistics that year. Exports of software services are reported as Rs 46,424 crore. This figure compares with the Nasscom total of $9.6 billion of exports of both IT services and software and ITeS-BPO, which amounts to Rs 46,459 crore when converted at the average annual exchange rate quoted above.

This reveals the effect of using the Nasscom figures as the “controlling total”, resulting in the inflation of the RBI’s figures as well. Dataquest’s figure for all IT-related services (including ITeS) exports for that year was Rs 45,867 crore, which too was close to the Nasscom figure, falling short by just 1.3 per cent.

In sum, there is reason to believe that the figures based on which discussions on IT sector growth are conducted are overestimates.

However, there are many who argue that the export figures do not capture in full India’s presence in the global market for IT-enabled services. This argument is based on the fact that software and IT services are delivered in two forms. One is through the presence of the service provider at the point of consumption of the service, requiring the temporary movement of professionals who provide the services concerned onsite (actual short-term migration).

The other is through the remote delivery of services that has gained in share with improvement in the quality and reduction in the costs of digital communication. Of these two kinds of providers, short-term migrants (to the US on H1-B visas for example) earn incomes in the US, and savings out of these incomes that are brought back to the country, it is argued, should be included in the revenues garnered by the country from the export of software and IT-enabled or IT-related services.

It has been suggested that if this component is taken into account it would make a significant difference to the foreign exchange earnings from IT-related services. That argument rests on two sets of facts. The first is that besides services revenues, a major source of foreign exchange inflows recorded on the current account of the Indian balance of payments is private transfers, consisting predominantly of remittance flows.

Remittances vs software services

Initially remittances were far more important than software services earnings. But because of the rapid rise in IT-related earnings, the ratio of software and IT-related services earnings to remittance inflows has risen from 48 per cent in 2000-01 to 57 per cent in 2003-04 and 111 per cent in 2006-07. Yet remittances are still important in absolute terms and till recently were higher than earnings from software and IT-related services.

A second feature is the structural transformation of remittance flows reflected in a change in the sources from which remittances originate. It has been argued that starting from the early 1990s there has been a structural break in the composition of remittances with an increase in the importance of remittances from the US in total inflows of remittances into the country.

The break was reflected in an increase in the share of remittances coming from the dollar area (mainly the US) as opposed to the sterling area (mainly West Asia) (Tables 1 and 2). This has been attributed to the growing number of Indians residing in the US on temporary visas to provide onsite software and computer-related services.

Earlier, migration for work to the US was more in the nature of intended permanent migration, whereas migration to West Asian countries (after the oil boom) was in the nature of short-term or temporary migration. This resulted in the fact that remittances to India came more from Indian workers in the Gulf, who had to finance the consumption needs of their families as well as save for their future in India, rather than from those in the US.

That is seen to have changed because of the short term migration of IT professionals implying an increase in India’s foreign exchange earnings from the export of IT-related services. Thus the provision of labour services through actual migration is seen as having substantially increased earnings from the export of IT-related services, diluting the case that quoted figures are overestimates.

Claims of this kind need to be treated with caution. This is because estimates of revenues provided by at least some sources do possibly incorporate the foreign earnings of onsite workers. This possibility was first noted by the World Trade Organisation , which questioned the veracity of the Indian data. The study was concerned with the rather large discrepancy between the reported volume of India’s IT and IT-enabled services exports to the US as reflected in the Balance of Payments (BoP) statistics of India, on the one hand, and the US, on the other.

The Indian data suggested that software exports to the US in 2003 amounted to be $6.77 billion. On the other hand, IMF statistics indicated that US imports of IT services through unaffiliated agents from India amounted to $330 million in 2003.

The same source reveals that the share of unaffiliated trade in total US IT imports stood at 36.5 per cent. Assuming that the same figure holds for US bilateral trade with India, the (estimated) imports of IT services from India works out to just $900 million.

One way of reconciling this major difference in the numbers ($6.77 billion and $0.9 billion) is to take account of the possible mis-categorisation of other business services as software services. BoP data on “business services” are reported under two heads: ‘computer and information services’ (CIS) and ‘other business services’.

However, national statistics do not reflect a uniform principle with regard to the allocation of services trade between these two categories. In particular, in India a range of IT-enabled services, such as call centres and medical transcription services, are reported along with export of software services.

For example, the Reserve Bank of India’s (RBI) Annual Report 2004 shows that Indian “software” exports worth $12.2 billion in 2003-04 include IT-enabled services amounting to $3.6 billion.

But assuming that the figure on India’s exports of “software” to the US is inflated by the same proportion, we should expect that software exports to the US would have amounted to $5.75 billion. This still is way above the US BoP figure.

To explain this difference, the WTO examined the possibility that what are paid as “salaries” to India IT-workers on short term H1-B visas gets reported as software exports. It is true that the Nasscom reports that a large share of India’s “software exports” are delivered “onsite”. Inasmuch as these onsite services are provided by a foreign affiliate of the “exporting” company, these are treated in US statistics as local sales of these foreign affiliates and not included in the BoP data.

Even if the export is not to an affiliated enterprise, but involves onsite delivery to an unaffiliated enterprise, the inclusion of the salary paid to the service provider in the figures on imports of services depends on how long the provider remains in the host/importing country.

“Onsite” delivery by Indians employed abroad can be treated as an export from India only as long as these employees have not been staying abroad for more than one year.

If they do they should definitionally be considered residents of the host country. The WTO study supports the view that erroneous inclusion of salary payments in the Indian data and their exclusion from the US data is possibly what explains the discrepancy between US and Indian BoP statistics.

The WTO’s conjecture is by no means definitive. But what its analysis does suggest is that the salaries of Indians working on short-term visas in the US are included in the Indian export figures. It would be wrong to consider remittances from such workers as an additional source of “service export revenues” for the country. Adding on the remittances from these incomes to overall exports would amount to double counting.

This conclusion is also corroborated by a study undertaken by the United States Government Accountability Office (GAO), which attempted to understand the variations in the evidence on the extent of off-shoring of business, professional and technical (BPT) services from the US to India.

Thus, for 2002 and 2003, the US reported $240 million and $420 million respectively in unaffiliated imports of BPT services from India, while India reported about $6.5 billion and $8.7 billion respectively in affiliated and unaffiliated exports in similar services categories.

One obvious cause for the difference is that the Indian data included exports to the US by affiliated and non-affiliated firms in the country, while the US data covers only imports from non-affiliated firms. But the GAO assesses that these differences cannot be attributed to coverage alone.

In its view the following additional factors could explain the difference:

Indian data on trade in services include packaged software and software embedded on computer hardware, which the US classifies as trade in goods. This factor accounts for approximately 10-15 per cent of Indian exports.

India includes in its data certain information technology-enabled services, such as some financial services, that are not included in BEA’s definition of BPT services.

India treats sales to US-owned firms located outside of the US as exports to the US, but the US does not count these as imports.

India counts the earnings of temporary Indian workers residing in the US as exports to the US. However, the US only includes temporary foreign workers who have been in the US less than one year and who are not on the payrolls of firms in the US. Indian officials estimate that this factor may account for 40-50 per cent of the difference between US and Indian data.

These assessments suggest that the Indian data do take into account the earnings of Indian workers in the US when computing export figures, so that adding on figures based on remittances from them would amount to double counting.

The problem of a lack of robustness of the Indian figures remains. Given that the industry has been celebrated and pampered for long, the least the Government could do is get down to measuring how big Indian IT really is.