Splintering

Discussing determinants of rapid growth in services sector, Bhagwati (1984) pointed out that splintering (the externalisation of some activities) played a significant role in the growth of services sector. Splintering in simple words refers to the outsourcing of various types of services by agricultural and industrial sector. Splintering or outsourcing arises because with the expansion of production in a sector it becomes economical for the sector to contract out some of its activities to other efficient or specialized sector (Rakshit, 2007).

According to Gordon and Gupta (2004) changes in production technique make it econom- ical for the producer to outsource some of its activities such as business operations which were done internally by individual firm. Such contracting out enhances services interme- diation leading to increase in services output. Using input-output coefficients, Gordon and Gupta (2004) show that services sector input in industry increased by about 40 per cent during 1979-80 to 1993-94. Another way to measure the splintering effect is to estimate the change in input-output coefficients of services input in other sector. Applying this ap- proach, the authors estimated that splintering in services had contributed 0.5 per cent point to the annual growth of services during the 1980s whereas contribution of splintering in the annual growth declined to 0.25 per cent during 1990s. Using similar method, Nayyar (2012) estimated effect of splintering in services sector for the period 1979-80 to 2006-07 and found about 0.17 percentage point effect of splintering in determining annual growth in services sector. According to the author, the contribution of splintering in the growth of services is negligible and has declined over time. In addition, the author noted that usage of storage, communications, hotels and restaurants, public administration and defence, and, education and health services in production of other two sectors as input remained more or less constant during the period 1993-94 to 2006-07 whereas moderate increase was recorded in wholesale and retail trade, financial services, real estate and business services and personal services. Central Statistical Office, under the Government of India compiles input-output table which gives information on the input flow from one sector to other sec- tors for Indian economy. In Table 3.8, proportion of output of variour services segments used as input in other sectors is presented for the years for which input-output tables are available. It is clear from the table that the inter-industry usage of services declined in railways, transport by other means, storage, communication, banking and insurance and other services during 1993-94 to 2007-08. In contrast to the decline of splintering in most

services segments, splintering increased in trade and hotel and restaurant.

In the discussion of determinants of services sector, three channels are analysed i.e.

high elastic demand, international trade and splintering. In order to assess the impor- tance of these factors in determining output in services sector, the present study attempts to extend the investigation with the help of econometric techniques. Gordon and Gupta (2004) examined important determinants of services growth in India. They utilised two approaches to identify significant determinants of services sector growth in Indian econ- omy.

In the first approach, they estimated separate regression using time series data to ex- plain growth rates in different services activity such as business, communication, financial, hotels and restaurants, distribution, transport, public administration and defence and com- munity and personal services for the period 1952-2000. In the regression analysis, they considered annual growth rates of each of the services sector output as dependent variable

Table 3.8:Percentage of Inter-Industry use to Total output of Various Services Segment

Sector 1993-94 1998-99 2003-04 2006-07 2007-08

Changes over 1993-94

to 2007-08

Trade 44.80 44.72 51.05 51.58 54.66 9.86

Hotel

and Restaurant

16.20 17.02 21.58 18.86 45.07 28.88

Railways 68.90 70.51 60.24 70.69 56.67 -12.23

Transport by Other Means

50.10 72.91 43.49 45.82 47.32 -2.78

Storage 99.08 98.96 98.27 98.94 98.70 -0.39

Communication 65.45 51.17 69.63 75.49 57.78 -7.67

Banking and Insurance

80.52 68.67 75.65 80.79 71.88 -8.64

Other Service 45.55 62.53 30.97 35.66 38.51 -7.04

Source: Input-Output Table compiled by Central Statistical Office (Various Year); Last column presents changes over 1993-94 to 2007-08.

whereas growth rates of commodity producing sector, external trade of goods, service ex- ports, and two dummy variables for 1990 and 1980 were used as explanatory variables in their model. The dummy variables were used to capture the impact of economic reforms related to liberalisation. In the second approach, they used a panel data model for different services activities including trade, hotels, railways, transport by other means, storage, com- munications, insurance, banking, dwellings business and personal services. They utilised five-year average data ranging 1970-2000 for the estimation of the panel regression. The dependent variable in that model was annual growth rates in services a activity. The in- dependent variables were average growth rates in industry and agriculture; the volume of external trade in goods; the value of exports in services and a dummy variable to capture whether reforms were carried out in each services sub-sector. Results in the first approach show that for components like business, communication, financial, hotels and restaurants and community services the dummy variable for 1990 is positive and highly significant.

The dummy for 1980 is also found significant for business, financial and community ser- vices. Thus they concluded that the sectors which were opened up experienced faster growth. Furthermore, the growth of external trade in goods is found to have important impact on transport services. They found domestic demand as well as foreign demand less important in determining the growth of most of the services segments. Growth in ser- vices export has been found significant for business services only. In the second approach, dummy variable for reforms is found to be significant implying that sectors opened for ex- ternal factors such as FDI, trade or private ownership experienced faster growth. Among other variables, output growth in the industrial sector is found to be significantly correlated with services sector output growth.

Eichengreen and Gupta (2010) also examined important determinants of services sec- tor growth in India. They used growth in value addition of different services as the de- pendent variable in their model. The explanatory variables were per capita income, the square of per capita income, the tradability of the service in question, index for liberalisa- tion, skilled-labour intensity, and whether the activity is correlated with industrial growth.

Another explanatory variable they included is the difference between the share of a ser- vices category sayito the GDP in India and the share of thei^th services category in other countries. According to the authors, this variable captures catch-up; the extent to which an activity with initial small shares is likely to grow. Their study reveals that per capita

income and per capita income squared are not individually significant but they found these two variables jointly significant. Tradable services exhibit significant growth rate. The ex- planatory variable, ‘liberalisation’ also exhibits positive and significant impact on services sector growth. The study found that those services category with a very small share in the initial period registered higher growth rate.

In a more recent study ranging the period from 1990-91 to 2008-09, Nayyar (2012) examines the determinants of the services sector in India considering eleven services sub- sectors i.e. trade, hotels and restaurant, transport, storage, communication, financial ser- vices, real estate and renting services, business services, public administration and defence, social and community services and personal, cultural and recreational services. Using two different models, the author tried to identify the potential determinants of services sector growth in India. In the first approach, Nayyar (2012) regressed share of services sub-sector in GDP on sum of the value of exports and imports by services sub-sector, index for liberal- isation, total enrolment of students in higher education, coefficient to measure the intensity of the use of different services in producing industrial output and price of a service sub- sector output relative to industrial output. Lagged values of the explanatory variables were used in the regression model. In the second model, the author used the following set of explanatory variables; private final consumption expenditure on services, government fi- nal consumption expenditure and export by services sub-sector. The dependent variable was the same as in the first model. The results of the first model depict that the levels of GDP, services trade and liberalisation have a statistically significant and positive impact on the share of services in GDP whereas in the second model, liberalisation and private final consumption expenditure on services and services export are found to be statistically significant.

With this brief review of literature, the present study tries to examine the effect of various factors in the growth of services sector in India.

The present exercise examines the effect of three factors i.e. per capita income, size of export and effect of splintering on the expansion of services sector in India. Thus, the present approach does not differ much and largely remain similar to the existing studies in selecting the explanatory factors of the services sector growth in India. However, the present study differs from the earlier ones on several counts. First, the present study dif- fers from the existing ones in terms of its sample period. The sample period taken for

the present study is 1980-81 to 2012-13, whereas the most recent available study (Nayyar, 2012) covers the period 1990-91 to 2008-09. Moreover, this data period allows us to take into account two major economic events i.e. the economic liberation of Indian economy in 1991 and the other is the global financial crisis of 2007-08 that took place during this pe- riod. Second, the present study departs from the existing literature in terms of its choice of dependent variable. While the earlier studies have taken growth rate of services sector or share of services sector to the GDP as their dependent variable, the present study considers gross value addition (GVA) in the services as the dependent variable. It is assumed that the GVA is a better option for dependent variable as it is independent and not relative to any factor. In contrast, growth rate is relative to the previous period and share of services sector to the GDP is relative to the share of other sectors in the economy. Third, the present study uses VAR modelling in explaining the effect of income and export in GVA of services. Ad- ditionally, in order to examine the role of splintering in services, an econometric approach in the form of a panel model has been used while most of the studies discussed above have made use of mathematical approach for the purpose. The details of the time series analysis and analysis under panel data framework are discussed separately in the following section.