Services In The Indian Growth Process
Regional Dimension of Economic Growth in India
Concluding Observations

IV. SERVICES IN THE INDIAN GROWTH PROCESS

3.92 The phenomenal expansion of services world-wide led to services being regarded as an engine of the growth and even as a necessary concomitant of economic growth. In development economics, seminal works on phases of growth (Fisher, 1935; Clark, 1940; Rostow, 1960; Kuznets, 1971) suggest that development is a three-stage process. The dominance of the services sector in the growth process is usually associated with the third stage of growth. In this context, the ascendancy of services even in developing countries has been regarded as a mutation of growth. During the 1980s and 1990s, services accounted for a share of close to or above 70 per cent of GDP in industrialised countries and about 50 per cent in developing countries. In India, services accounted for 38.6 per cent of the GDP in the 1980s and 44.3 per cent in the 1990s.

The Stylised Evidence

3.93 The growth of services sector has imparted resilience to the economy, particularly in times of adverse agricultural shocks as also during cyclical downturns in industry. The share of services sector has been steadily increasing, with a fairly rapid growth in the 1990s (Table 3.29 and Chart III.43).

Table 3.29 : Sector-wise Average Shares, Growth Rates and Contribution to GDP Growth
										(Per cent)
		Services			Industry#			Agriculture
Period		Share	Growth	Contribution	Share	Growth	Contribution	Share	Growth	Contribution
		in	Rate	to GDP	in	Rate	to GDP	in	Rate	to GDP
		GDP		Growth	GDP		Growth	GDP		Growth
1		2	3	4	5	6	7	8	9	10
1950-51 to 1959-60*		28.2	4.1	32.2	16.0	5.7	25.3	56.0	2.3	42.5
1960-61 to 1969-70		31.4	4.9	38.1	21.1	6.5	32.9	47.8	2.5	29.2
1970-71 to 1979-80		34.4	4.5	52.7	22.8	3.7	28.7	42.8	1.3	18.6
1980-81 to 1989-90		38.6	6.6	43.6	25.0	6.8	28.9	36.4	4.4	27.5
1990-91 to 2000-01		44.3	7.6	57.6	27.1	5.9	27.6	28.6	2.9	14.8
#	Inclusive of construction
*	9 year data for growth and weighted contribution since data for 1949-50 are not available.
	Due to rounding off, sectoral data may not add up to 100.
Source : National Accounts Statistics, CSO.

3.94 The weighted contribution of the services sector to GDP growth has been rising since the 1960s except during the 1980s when industry and agriculture recorded substantial acceleration, which led to erosion in their contribution to GDP growth. However, a resurgence of services in the 1990s with a growth of 7.6 per cent enabled services sector to contribute 57.6 per cent to GDP growth.

3.95 A notable feature of the structural transformation of the services sector has been the growth of skill intensive and high value added sectors, i.e., software, communication and financial services. The rapid growth of services can be attributed, inter alia, to the advent of information technology (IT) and the knowledge economy. This has enhanced the growth of the high productivity segment of the services sector as well as a variety of service activities involving low productivity activities catering to a large mass of people. The phenomenal growth of low skilled service activities has occurred due to reduced opportunities in the manufacturing sector, particularly in the unorganised sectors.

3.96 Trade, hotels, restaurants and transport and communication are the major segments in terms of their share within the services sector; however, their share in the value added in the services sector has remained constant at around 49 per cent during the 1970s and 1980s, before marginally declining in 1990s. The share of finance, insurance, real estate and business services witnessed significant improvement during the 1990s on account of the rapid pace of financial development (Chart III.44 and Table 3.30). The buoyancy in the services sector output is concentrated in the new economy sectors such as computer software and financial and business services. In the national income accounts, the contribution of the software industry in the real GDP is estimated at 0.78 per cent in 1999-2000. If raw materials in this sector are accounted for, contribution of the software sector (net of inputs) to GDP stands at about 1 per cent as per the estimates of the NASSCOM and the CMIE.

Table 3.30 : Average Share of Sub-sectors in Services Value-added (at constant prices)
			(Per cent)
Period	Trade,	Finance,	Social and
	Hotels,	Insurance,	Personal
	Transport and	Real estate and	services
	Communi-	Business
	cation	services
1	2	3	4
1950-51 to 1959-60	44.3	23.0	32.8
1960-61 to 1969-70	48.2	19.5	32.3
1970-71 to 1979-80	49.1	18.4	32.5
1980-81 to 1989-90	48.7	20.4	30.9
1990-91 to 2000-01	46.3	26.1	27.1

3.97 Some of the activities in the services sector are multidimensional, being part of industry as well as services, such as information technology and construction. Service statistics in most countries including India provide information on value-addition of various activities of business services, hotels, trade, financial services, etc. For an empirical analysis, sub-sectors including trade, transport and communication, financing, insurance, real estate and business services can be categorised as producer services with hotels and restaurants and other services as consumer services. Government services comprise public administration and defence services. During 1999-2000, producer services accounted for about 70 per cent of the total services followed by consumer services (17 per cent) and government services (13 per cent). The high share of producer services reflects the strong inter-linkages between services and goods producing sectors of the economy.

Income Elasticity of Demand for Services

3.98 A rising share of services in GDP is regarded as an outcome of higher income elasticity of demand for services. The empirical studies have shown that the income elasticity of demand for services could be greater than or equal to unity (Gemmell, 1982; Summers, 1985; Bergstand, 1991; Falvey and Gemmell, 1991). Income elasticity of demand for services increases with rising income which favours the fulfillment of more sophisticated desires. During the development process, distribution of GDP and employment register sectoral shifts. Such shifts may occur on account of the hierarchy of needs, distinguished into basic needs for food and shelter and needs for other material and non-material goods including services (Maslow, 1970). According to this view, income elasticity of demand depends on per capita income and differs across various sectors.

3.99 The empirical estimates of price and income elasticity for various categories of services in India are summarised in Table 3.31⁸. It is important to mention that the actual behaviour of the services sector in real GDP depends on the relative strength of the coefficients of income and price elasticity.

Table 3.31: Income and Price Elasticities for the Services Sector
Sector	Income Elasticity	Price Elasticity
1	2	3
Services	1.20 *	-0.68 *
1. Producer Services	1.22 *	-0.78 *
2. Consumer Services	1.00 *	-0.10
3. Government Services	1.41 *	-1.05 *
* Statistically significant at 1 per cent.

3.100 The income elasticity of demand is greater than unity and price elasticity is negative and significant for the total services, producer services and government services. In other words, demand for overall services rises with increase in per capita GDP and decreases with increase in prices of services. The higher income elasticity of demand in the case of producer services underscores its forward linkages. This is corroborated by the emergence of producer services comprising advertising, publicity, marketing and other IT-related activities in the recent period as important service industries in India. Therefore, producer services can be regarded as a major source of economic growth. Similarly, public administration, social services, rural extension services and defence sectors which together comprise government services, have a high income elasticity of demand. In the case of consumer services, the income elasticity is almost equal to unity and price responsiveness is not detected. It indicates that demand for consumer services increases in same proportion to change in real per capita income and is price insensitive. In order to check for robustness of the empirical findings, the hypothesis of constant share of services in GDP is tested⁹. The results indicate that the share of overall services responds by 0.34 per cent to a unit change in per capita real GDP while producer and government services respond by 0.41 per cent and 0.27 per cent, respectively. On the other hand, in the case of consumer services, with a unit income elasticity and insignificant price elasticity, its share is found to increase by a marginal 0.08 per cent. However, the results need to be interpreted with caution as the coefficient of determination, i.e. adjusted R², is quite low in the case of consumer services (Table 3.32).

Table 3.32 : Service Share Regression Coefficients
	Coefficient	Adjusted R2	D.W.
1	2	3	4
Services	0.34	0.92	1.55
1. Producer Services	0.41	0.94	1.38
2. Consumer Services	0.08	0.22	1.11
3. Government Services	0.27	0.43	0.48
Note : All the coefficients are significant at 1 per cent .

Services, Employment and Productivity

3.101 The level of employment in services is strongly correlated with the stage of economic development; while agricultural and manufacturing activities account for a major share of employment in developing countries, services activities account for a major portion of employment in most developed countries. Various studies have explored the sources of growth in services employment (Baumol, 1967; Fuchs, 1968). Lagging productivity in the services sector is considered as the main reason behind the rising share of service employment in total employment even though share of services in real GDP remains constant, implying the existence of Baumol's cost disease (Baumol, 1967).

3.102 In India, the services sector accounted for 18.1 per cent of the total employment during 1965-66, going up to 23.5 per cent in 1999-2000. Generally a substantial increase in the share of services in employment in most countries could imply a growth of low productive, low income, informal sector (Bhaduri, 1996). However, the increase in earnings per worker in the services sector compared with the industry could mean that at least part of the increase in employment in the services sector is in the formal, perhaps public sector, or in the new IT-related industries with higher use of capital per worker (Table 3.33).

Table 3.33 : Share of Services Sector in Total Employment
Period	Employment	Total	Share of
	in Services	Employ-	Services in
	Sector	ment	total Employ-
	(in crore)	(in crore)	ment (%)
1	2	3	4
1965-66	3.97	21.93	18.1
1970-71	4.82	24.09	20.0
1980-81	5.71	30.24	18.9
1990-91	8.70	35.68	24.4
1999-2000	10.29	43.81	23.5
Source: National Sample Survey Organisation (various rounds) and Visaria,1996.

3.103 In order to focus upon the differences in growth rate of employment and gross value added in services sector since 1970-71, a difference of means test is employed with the following null hypotheses: (i) there is no difference in the growth rate of employment in services sector and growth rate of gross value added in services sector; (ii) there is no difference between labour productivity growth and employment growth in services sector. Labour productivity is defined as value added in services sector divided by total labour in services sector (Table 3.34).

Table 3.34 : Difference of Means Test for Differences in Growth Rates
Period			Null Hypothesis	Mean	Finding at 1%
				difference	level of significance
	1		2	3	4
i)	1970-71 to 1999-2000		No difference between the growth	-3.5	Null hypothesis Rejected
			rates of Employment and gross
			value added in services sector
ii)	1970-71 to 1999-2000		No difference between the growth rates	0.7	Null hypothesis Accepted
			of labour productivity and employment
			in services sector
	a)	1970-71 to 1979-80	No difference between the growth rates	0.7	Null hypothesis Accepted
			of labour productivity and employment
			in services sector
	b)	1980-81 to 1989-90	No difference between the growth rates	-2.2	Null hypothesis Rejected
			of labour productivity and employment
			in services sector
	c)	1990-91 to 1999-2000	No difference between the growth rates	3.6	Null hypothesis Rejected
			of labour productivity and employment
			in services sector

3. 104 The results show that hypothesis (i) can be rejected, i.e., growth rates of employment and value added in services sector are statistically different from each other during 1970-71 to 1999 -2000. As the mean difference is negative, the value added growth in services sector is, on an average, higher than employment growth in services sector, which confirms the finding of other studies that services sector has undergone a less than proportionate increase in employment in relation to output (Mitra, 1988; Bhattacharya and Mitra, 1990). The second hypothesis about the differences in productivity and employment growth in services has been tested decade-wise. Growth in services productivity for the whole period of 1970-2000 is observed to be higher but not statistically significant. The growth rate of productivity in services sector has been higher during the 1990s as the mean difference is found to be positive and statistically significant. Thus, unlike many other countries, growth of average productivity of labour in India has remained higher than employment growth in services sector during the post-reform period. The empirical findings are supported by the declining capital-output ratio in the services sector which can be interpreted as efficient use of capital by skilled labour and low contribution of services to total employment in contrast to its high contribution to overall GDP. This points towards a growth in total factor productivity in services (Chart III.45).

Terms-of-Trade

3.105 The services sector is not covered in the construction of Wholesale Price Index (WPI), leaving a major portion of economic activities outside the measurement of headline inflation. Therefore, the inter-sectoral terms-of-trade can only be studied with the help of sectoral deflators. However, such deflators have their inherent limitations, as they are available with low frequency.

3.106 There are divergent views on intersectoral terms-of-trade; nevertheless, the trend in sectoral deflators shows that terms-of-trade have remained in favour of agriculture vis-a-vis industry or services all through the 1990s. Thus, the gap between the sectoral deflators has been widening in favour of agriculture at the cost of services and industry. On the other hand, a comparison between services and industry shows that terms-of-trade have remained in favour of services (Chart III.46).

3.107 Within the services sector, the terms-of-trade have remained in favour of government services, closely followed by consumer services. The gap in the deflators between producer services, on the one hand, and consumer and government services on the other, has been widening since 1993-94. This is possibly due to the fact that for government services, which include public administration and defence, the behaviour of the deflator depends largely on government policies. The recent pay hike following the fifth pay commission's recommendation might have contributed to a relatively high deflator (Chart III.47).

3.108 The terms-of-trade have remained in favour of community, social and administration services since 1995-96 while the deflator of financing, insurance and business services is moving upward at a faster rate since 1998-99. Deflators of other two sub-sectors (transport, storage and communications and trade, hotels and restaurants) have been moving in close proximity of each other (Chart III.48). The behaviour of inter-sectoral and intra-sectoral terms-of-trade in services would have a bearing on India's position vis-à-vis the evolving multilateral framework for international trade in services (Box III.4).

Box III.4

WTO and Services

The growing role of international services and their implications have come to be recognised in the General Agreement on Trade in Services (GATS) of the World Trade Organisation (WTO). The WTO rules on services trade, as embodied in the GATS and negotiated in the Uruguay Round, are the first ever set of multilateral, legally enforceable rules covering international trade in services. Like the agreements on goods, GATS operates on three levels: the main text containing general principles and obligations; annexes dealing with rules for specific sectors; and individual countries' specific commitments to provide access to their markets. Unlike in goods, GATS has a fourth special element: lists showing where countries are temporarily not applying the 'most favoured nation' (MFN) principle of non-discrimination. The temporary withdrawals of MFN treatment are also an integral part of GATS. A WTO council on services oversees the operation of the agreement. Under the framework of GATS Article, it covers all internationally traded services and the MFN clause applies to all services except the one-off temporary exemptions. It deals with foreign competition although the negotiations and commitments made can have a bearing on dismantling of domestic monopolies. The definition of what constitutes trade in services is currently a subject of multilateral negotiations.

India falls in the category of developing countries that show relatively strong position in earnings from labour and travel, with negative positions in trade in goods and most of other service categories excluding IT related services. During 2000-01, there was a surplus amounting to US$ 294 million and US $ 135 million on account of travel and insurance, respectively, in India's Balance of Payments while the deficit on merchandise account amounted to US $ 14,370 million.

The advocates who favoured inclusion of services in the Uruguay Round negotiations argued that free trade in goods is insufficient. On the other hand, many of the developed nations have opined that only those sectors (e.g., banking, insurance and telecommunication) in which they enjoy a comparative advantage, should be brought under the purview of GATS. It is generally felt that India should strive for a liberalised deal in the case of professional services in general and software services in particular in which she has a comparative advantage. On the other hand, there is a need to tread cautiously in respect of some services sub-sectors like banking and insurance. Liberalisation in financial services sector could be introduced in a phased manner with some regulations, as unbridled liberalisation may affect the financial stability. However, keeping with its commitment to multilateral agreement on services, India has recently raised the limit for banking licenses in respect of foreign banks from 8 to 12 in the Financial Services Agreement. In addition to this, as a part of unilateral liberalisation, India has opened up its insurance sector and other financial services including financial consultancy for foreign investment with some limit on market access. Thus, India's interest in the ongoing GATS negotiations is uniquely placed, given her comparative advantage in the professional services and her concerns for a well calibrated and cautious liberalisation of financial services.

The recently concluded Ministerial Conference of the WTO at Doha recognised the work already initiated since early 2000 under Article XIX of GATS and many countries have submitted proposals on a wide range of sectors and other issues as well as on movement of natural persons. Among the different modes of services supply, India is most interested in movement of natural persons and has also submitted a proposal at the WTO Council for trade in services. However, other modes, viz., commercial presence, cross border supply and consumption abroad are also important for making the Indian service sector a major player in emerging international scenario.

References

1. Government of India (2001), WTO and India, various Issues, Ministry of Commerce.
2. WTO (2001), Trading into the Future - Agreements -Services.

Sustainability of Services Sector Growth

3.109 Besides its direct contribution to GDP, services sector can be a source of productivity for other sectors and can thereby facilitate expansion in other sectors of the economy. In India, it is the services sector which has kept the GDP growth around 6.0 per cent in the 1990s when industry and agriculture sectors did not perform relatively well. The coefficient of variation of sectoral growth rates, used as a proxy for output variability, is found to be low for services as compared with other sectors (Table 3.35).

Table 3.35: Coefficient of Variation of Sectoral Growth Rates
			(Per cent)
Period	Agriculture	Industry	Services
1	2	3	4
1970-71 to 1979-80	623.0	100.0	33.3
1980-81 to 1989-90	138.6	30.9	19.7
1990-91 to 1999-2000	125.0	58.7	26.3

3.110 As real per capita GDP grows, demand for services increases more than proportionately and this, in turn, reinforces GDP growth itself. Within the services sector, demand for producer and government services, which constitute mainly intermediate consumption, have strong multipliers impacting on real GDP. On the other hand, the demand for consumer services, which can be considered as final consumption, does not increase proportionately with the increase in real GDP. Therefore, producer and government services seem to be more important as future source of growth. Besides, there is now a general consensus on the vast potential in some segments of services sector for yielding increasing returns, particularly IT related and software services even though their shares are currently small. There has been a gradual shift towards use of IT both in the public sector and the private sector, particularly in education, medical services and exports. Accordingly, the Task Force on Human Resource Development in Information Technology set up by the Central Government has recommended government intervention in promotional activities like distributive services, financial services, business services, administration services, entertainment services and personal services which, in turn, would provide the much needed demand push for IT development from domestic sources. The growth of such dynamic service activities, which are intensive users of communication and information technology, will generate employment opportunities on a rising scale. Already, banking and insurance sectors have started synergising with IT.

3.111 The financial services sector, with its significant share in services could turn out to be an influential source of growth. Export of services which, at present, contributes approximately 26.7 per cent of total exports can give a major boost to overall economic growth. With information technology-led global progress, trends are likely to be in favour of a major expansion of the world trade in services, particularly communication, financial services, computer and information services, technology transfers and different business services. India, with its large and expanding knowledge base, can explore the opportunities (Raipuria, 2001). The labour productivity in software is twice the ratio of India's manufacturing and 1.3 times that of the US (Arora and Athreye, 2001). The potential of its high capacity to generate wealth, foreign exchange and employment has been recognised at all levels. The global IT industry offers Indian companies unique opportunities in four broad areas: value-added IT services, software products, IT enabled services and e-business (Unni and Rani, 2000). As per NASSCOM projections, India's software and IT services sector is set to grow at a rate of 30 to 35 per cent during 2001-02. The IT industry is slated for a target revenue of US $ 87 billion by 2008 with a 7.5 per cent contribution to the GDP.

3.112 The process of economic growth has itself led to the emergence and expansion of new services such as advertising, publicity, marketing, etc. These sub-sectors provide essential service inputs to other sectors in the economy, thereby developing strong linkages with the rest of economy. Various studies have examined the expansionary potential of services sector in non-services industries and found strong forward linkages, as 50 per cent of the industries in the economy are found to be directly or indirectly service intensive (Dutta, 1989; Bhowmik, 2000).

3.113 Efficient delivery of services increases the productivity of both labour and capital in the economy as a whole. In general, services sector appears to be highly growth inducing with positive externalities for other sectors, making services a catalytic agent of growth. It needs to be recognised that the services sector in itself is not homogenous. Therefore, the expansion of these sectors needs to be calibrated and integrated simultaneously into the overall growth process to impart an element of sustainability.

V. REGIONAL DIMENSION OF ECONOMIC GROWTH IN INDIA

3.114 The regional dimensions of growth of the Indian economy are assuming increasing relevance in the context of the progressive diffusion of structural reforms at the sub-national level. The quality of growth is getting increasingly assessed in terms of durable improvement in the regional growth profiles in which the interface between public policies for accelerating development and standards of living is the greatest. Moreover, regional patterns of growth provide a gauge of the quality of public policies themselves and their impact on macroeconomic welfare. The various facets of the growth experience of States in India are critical for developing an understanding of the sources of demand generation as well as changes in productivity and growth. The growth performance in the States is often the outcome of institutional and non-economic factors interacting with the initial conditions which encompass various aspects of human capital development.

3.115 Varying degrees of reform in different States have yielded wide variations in growth performance. In this context, it has been pointed out that the popular characterisation of backward States such as Bihar, Madhya Pradesh (M.P.), Rajasthan, and Uttar Pradesh (U.P.) as a homogenous group of poor performers does not hold good in terms of recent economic performance (Ahluwalia, 2000). It has also been argued that reforms have unshackled a number of States like Andhra Pradesh (A.P.), Gujarat, Karnataka, Maharashtra and Tamil Nadu (T.N.) who could achieve their true economic potential in recent years (Bajpai and Sachs, 1999). From a completely different paradigm, the faster growth in select States has been interpreted as the operation of some kind of unbalanced growth with differing rates of catching up (Chaudhuri, 2000). Three other issues are raised in the context of Indian regional development, viz., fiscal, infrastructure and human resources development (World Bank, 2000).

Regional Growth Profiles¹⁰

3.116 The State-wise profile of per capita State Domestic Product (SDP) drawn up for 15 major States (representing nearly 90 per cent of Indian population) exhibits significant variation. In 1980-81, there were only four States viz., Maharashtra, Punjab, Gujarat and Haryana whose per capita real SDPs (at 1980-81 prices) were higher than the all-India per capita real GDP (Table 3.36). The trend remained more or less similar in 1990-91. In the 1990s, this group expanded to include Tamil Nadu. Relative rankings in terms of absolute SDP need to be evaluated against a comparison of growth rates. In terms of trend growth rate, Karnataka, West Bengal, Maharashtra, Tamil Nadu, Andhra Pradesh have done well in the 1990s (Chart III.49).

Table 3.36 : Trend Growth Rate of Per Capita SDP
						(Per cent)
				1980-81 = 100		1993-94 =100
States				1981-82 to	1981-82 to	1993-94 to
				1990-91	1993-94	1999-2000
	1			2	3	4
1	Karnataka			4.0	4.5	4.5
2	Gujarat			3.3	3.1	2.4
3	West Bengal			2.7	2.9	5.6
4	Kerala			4.3	3.7	3.5
5	Tamil Nadu			3.9	4.2	5.1
6	Orissa			3.2	3.7	5.8
7	Maharashtra			2.2	3.1	5.1
8	Andhra Pradesh			2.6	2.7	5.7
9	Punjab			3.4	3.2	3.7
10	Madhya Pradesh			2.0	2.0	4.6
11	Uttar Pradesh			4.0	3.6	2.5
12	Haryana			2.7	2.3	3.2
13	Bihar			2.4	1.7	2.0
14	Rajasthan			2.5	0.9	2.7
15	India			3 .4	3.2	4.8

Notes : 1.		Due to non-availability of data the average growth rates reported under column 4 for Madhya Pradesh, Gujarat, and Kerala are calculated over the period 1993-94 through 1998-99.
2.		The trend growth rates are calculated from semi-logarithmic function.

Source : Directorates of Economics and Statistics of respective State Governments.

3.117 The regional growth experience of the 1990s suggests that States pursuing reforms seemed to have experienced higher growth rates in recent years with some tendency towards convergence. At the same time, the regional growth experience is indicative of some kind of unbalanced and divergent growth (Box III.5).

Regional Dimension of Select Infrastructural Indicators

3.118 The convergence literature in India highlights the role of infrastructure-related facilities in fostering growth. A critical determinant of the divergence in growth profiles among regions is the inter-state differences in infrastructural facilities. Two select indicators of infrastructural conditions, i.e., electricity and transport are examined below.

Box III.5

Convergence of Economic Growth

One of the basic predictions of neo-classical or endogenous growth theory is that economies with lower capital per person tend to grow faster in per capita terms. Thus, there will be convergence of growth rates. The notion of convergence has attracted recent attention in the endogenous growth paradigm. It has been interpreted in two distinct senses. The hypothesis that poorer economies tend to grow faster than the richer ones is often referred to as the notion of absolute convergence. The empirical evidence in favour of absolute convergence from cross-country growth regressions has received mixed support. In fact, in some sample of countries the growth trajectories have been found to be quite independent of the initial conditions. This, however, could be due to the presence of heterogeneity in the sample of countries. It is in this context, the notion of conditional convergence has been proposed whereby an economy tends to grow faster, the further it is from its own steady-state value. In other words, each economy is conceived to be converging to its own steady state value, and the speed of convergence varies inversely to the distance from the steady state.

How far are the predictions on convergence valid in the Indian regional set-up? There are divergent views in this regard. While a number of studies found out that the pattern of growth of per capita SDP has followed a divergent tendency in absolute terms, controlling for internal migration and centre-state grants, there is evidence in favour of convergence among Indian states. The speed of convergence has been, however, found to be quite low with estimate of the required time for a typical Indian state to close one-half of the gap between its initial per capita income and the steady-state per capita income at 45 years!

What explains the pattern and extent of divergence among the growth profiles of Indian states, if any? Incorporation of omitted variables in the growth regression across Indian states has attracted attention in recent years. In the determination of regional steady-state level of income and ensuring conditional convergence, a number of variables have been highlighted, viz., literacy and per capita investment. In fact, it has been hypothesised that states with higher levels of income can undertake more ambitious infrastructure investment programmes, since they enjoy more fiscal revenues to pay from it. A contrary view regarding diverging tendency of per capita SDP growth is found as well. A number of reasons for such diverging trends has been cited, viz., (a) differences in private investment and public expenditure, (b) differing distribution of state Government expenditure, (c) regressivity of implicit inter-governmental transfers, (d) lack of horizontal equity in inter-governmental transfers, and (e) regional policies.

In the ultimate analysis, while much of the evidence of growth differentials among Indian states hints at some operation of conditional convergence, the determination of steady-state growth rate for a State is by itself a serious policy question. Thus, any explanation for convergence of (or its lack) growth across Indian States needs to be sought in the interplay of all these control variables, perhaps along with the institutional set-up prevalent in the region.

References:

1. Aiyar, Shekhar, 2000, 'Growth Theory and Convergence across Indian States: A Panel Study', in Callen, Tim, Patricia Reynolds, and Christopher Towe (eds.). India at the Crossroads: Sustaining Growth and Reducing Poverty, International Monetary Fund, Washington DC.
2. Barro, R.J. and X. Sala-i-Martin,1992, 'Convergence', Journal of Political Economy, Vol. 100.
3. Cashin, P. and R. Sahay, 1996, 'Internal Migration, Centre-State Grants and Economic Growth in the States of India', IMF Staff Papers, Vol. 43.
4. Marjit, S. and S. Mitra, 1996, 'Convergence in Regional Growth Rates: Indian Research Agenda', Economic and Political Weekly, August 17.
5. Nagaraj, R., A. Varoudakis, and M. A. Veganzones, 2000, 'Long-Run Growth Trends and Convergence across Indian States', Journal of International Development, Vol. 12.
6. Rao, M Govinda, R.T Shand, and K.P.Kalirajan, 1999, 'Convergence of Income across Indian States: A Divergent View', Economic and Political Weekly, March 27.

Power Supply

3.119 Electricity occupies a critical role in infrastructure. Various factors determine the demand and supply of this crucial input. The level of industrialisation, urbanisation, user cost and tariff structure, presence of subsidies, all could play a significant role in the determination of demand; on the other hand, technological and institutional arrangements are the crucial factors influencing supply. The electricity sector in various States suffer from endemic problems - unviable State Electricity Boards (SEBs), transmission losses and distortive subsidy structure. The relationship between the power situation and the regional growth in the 1990s reveals that there has been a general improvement in the power supply position in all the States in terms of deficit (i.e., requirement less availability as a percentage of requirement). In fact, two States viz., Orissa and West Bengal have managed to transform their power position from deficit to surplus (Table 3.37).

3.120 The existence of surplus power is, however, no indication of well being in the electricity sector. Out of the five States with deficits higher than the all-India average, three States, viz., Karnataka, Gujarat, and Maharashtra were high growth States in the 1990s.

Table 3.37 : State-wise Power Supply Position
							(in MUs)
	State	1991-92			2000-01
		Requirement	Availability	Deficit (%)	Requirement	Availability	Deficit (%)
	1	2	3	4	5	6	7
1	Orissa	8065	7499	7.0	11710	12070	-3.1
2	West Bengal	11140	10140	9.0	18787	18958	-0.9
3	Punjab	17238	16177	6.2	27670	26923	2.7
4	Haryana	10326	10123	2.0	17275	16793	2.8
5	Rajasthan	13220	13030	1.4	25080	24178	3.6
6	Kerala	7440	7197	3.3	13564	12670	6.6
7	Bihar	7415	5215	29.7	9208	8563	7.0
8	Tamil Nadu	23210	22086	4.8	42702	39462	7.6
9	Andhra Pradesh	24015	22415	6.7	47792	44055	7.8
10	Karnataka	20350	15550	23.6	30242	27490	9.1
11	Gujarat	25505	24417	4.3	53038	47877	9.7
12	Maharashtra	42070	40166	4.5	79527	71184	10.5
13	Madhya Pradesh	21115	19942	5.6	39644	34747	12.4
14	Uttar Pradesh	31540	28280	10.3	46295	39556	14.6
15	All India	288974	266432	7.8	507213	467401	7.8
Note	: MU = Million Unit, Minus sign indicates surplus
Source : Ministry of Planning and Programme Implementation.

In terms of availability of power, Maharashtra, Uttar Pradesh (U.P.), Gujarat, Andhra Pradesh (A.P.), and Tamil Nadu were leaders in 1990-91; in 2000-01, the front runners in terms of availability were Maharashtra, Gujarat, A.P., U.P., and Tamil Nadu. Excepting U.P, all other States are clearly the high growth states of the 1990s. This indicates that high growth is closely associated with demand- induced expansion in productive capacity. As the limits of capacity expansion are approached in the absence of significant technology and institutional reform, the power sector could emerge as a binding constraint on growth.

Roadways

3.121 Another aspect of infrastructure is the mode of transport. For the purpose of analysis, two distinct transport indicators have been taken into consideration, viz., roadways and railways as proxy for State-wise availability of transport infrastructure. In terms of the share of roads of a particular state in the all-India total, a number of States like Maharashtra, Tamil Nadu, M.P, Orissa experienced a rise in their shares, relative to their positions in 1981 (Table 3.38).

3.122 The only State that crosses the ten per cent mark in this context is Maharashtra. Again a close association is discernible between shares in all-India roads and the growth record of industrial States in the 1990s. In terms of surfaced roads, taking the all-India average of 56.5 per cent as a benchmark, States like, Haryana, Gujarat, Punjab, Maharashtra, Karnataka, Tamil Nadu, Rajasthan, Andhra Pradesh and Uttar Pradesh are found to be above the benchmark. By association, this list highlights the importance of infrastructure for growth.

Railways

3.123 In the case of railways, however, the distribution is explained in terms of geographical location as well as size of the State. Uttar Pradesh (undivided) accounted for the highest proportion of all-India railway routes, followed by Madhya Pradesh, Gujarat, Bihar and Andhra Pradesh (Table 3.39).

Table 3.38 : State-wise Roadways Development
State		Percentage Distribution of			Intra-state share of
		Total Roads			Surfaced Roads
		1981	1991	1997	1981	1991	1997
	1	2	3	5	6	7	8
1	Haryana	1.5	1.3	1.1	86.2	91.2	90.7
2	Gujarat	3.9	5.1	3.7	69.9	87.5	87.3
3	Punjab	3.1	2.7	2.6	76.9	78.1	81.5
4	Maharashtra	11.9	11.0	14.7	46.0	70.6	75.1
5	Karnataka	7.4	6.5	5.8	58.4	65.0	69.0
6	Tamil Nadu	8.2	9.7	8.4	76.5	15.6	68.0
7	Rajasthan	4.5	6.1	4.9	52.9	50.9	63.1
8	Andhra Pradesh	8.0	7.4	7.2	48.2	53.4	61.6
9	Uttar Pradesh	9.9	9.9	10.4	46.8	52.0	58.1
10	West Bengal	3.8	3.1	3.1	44.6	47.1	56.4
11	Madhya Pradesh	7.1	6.9	8.1	50.6	54.7	44.3
12	Bihar	5.6	4.2	3.6	34.3	36.1	37.3
13	Orissa	8.1	9.7	10.7	13.8	9.7	33.1
14	Kerala	7.0	6.7	5.9	22.4	27.8	31.1
15	All India	100.0	100.0	100.0	46.0	51.8	56.5
Source: Transport Research Wing, Ministry of Surface Transport, Government of India.

Table 3.39 : Distribution of Railway Routes in Major States
				Per cent
	State	1980-81	1990-91	1995-96	1999-
					2000
	1	2	3	4	5
1.	Andhra Pradesh	7.7	8.1	8.0	8.1
2.	Assam	3.6	4.0	3.9	3.8
3.	Bihar	8.8	8.5	8.4	8.4
4.	Gujarat	9.2	8.5	8.5	8.5
5.	Haryana	2.4	2.4	2.3	2.5
6.	Karnataka	4.9	4.9	5.0	6.3
7.	Kerala	1.5	1.6	1.7	1.7
8.	Madhya Pradesh	9.4	9.4	9.5	9.4
9.	Maharashtra	8.7	8.7	8.7	8.6
10.	Orissa	3.2	3.2	3.5	3.7
11.	Punjab	3.5	3.5	3.4	3.3
12.	Rajasthan	9.2	9.3	9.4	9.4
13.	Tamil Nadu	6.4	6.4	6.4	6.7
14.	Uttar Pradesh	14.5	14.3	14.2	14.2
15.	West Bengal	6.1	6.1	6.1	5.9
Source: Ministry of Railways, Government of India.

Select Social Dimensions of Regional Development

3.124 Apart from infrastructure, human capital and various social indicators play a crucial role in fostering growth. The State-profile of literacy, as an indicator of State-level human capital, along with urbanisation is discussed below.

Literacy

3.125 As various empirical studies show, disparities in knowledge operate analogously with factors of production to determine divergences in growth profiles among regions. A number of stylised facts emerge from the trends of literacy between 1991 and 2001 in eighteen select Indian states (Table 3.40).

3.126 First, all States have experienced an upward secular movement in literacy rates during the 1990s. Considering the fact that the 1990s were also characterised by a higher per capita GDP for all the States (in levels), a higher income has been accompanied by an improvement in the knowledge base. This is indicative of the symbiotic relationship between knowledge and growth. Secondly, the ranking of the States between the 1991 and 2001 remained more or less unaltered, indicating the uniformity in this regard.

Table 3.40 : Literacy Rate in Indian States
				(Per cent)
	State	2001	1991	Changes
		Census	Census	in the
				Literacy
				Rates
				during
				1991-2001
	1	2	3	4
1	Kerala	90.92	89.81	1.11
2	Maharashtra	77.27	64.87	12.39
3	Tamil Nadu	73.47	62.66	10.81
4	Uttaranchal	72.28	57.75	14.53
5	Gujarat	69.97	61.29	8.68
6	Punjab	69.95	58.51	11.45
7	West Bengal	69.22	57.70	11.52
8	Haryana	68.59	55.85	12.74
9	Karnataka	67.04	56.04	11.00
10	Chhatisgarh	65.18	42.91	22.27
11	Madhya Pradesh	64.11	44.67	19.41
12	Orissa	63.61	49.09	14.52
13	Andhra Pradesh	61.11	44.09	17.02
14	Rajasthan	61.03	38.55	22.48
15	Uttar Pradesh	57.36	40.71	16.65
16	Jharkhand	54.13	41.39	12.74
17	Bihar	47.53	37.49	10.04
18	India	65.38	51.63	13.75
Source: Provisional Population Totals: India , Census of India 2001.

3.127 Finally, the States with literacy rates higher than the all-India average are also those with higher growth rate of per capita SDP during the 1990s vis-a-vis the 1980s. The correspondence between the rankings on account of literacy and per capita SDP growth is not one-to-one. Kerala ranked third in terms of per capita SDP growth during 1991-92 through 1999-2000 and it is the first as per literacy standards in India. This close association between literacy and growth is perhaps indicative of the role of human capital in fostering growth.

Urbanisation

3.128 In the tradition of the stages of growth theories, urbanisation is symptomatic of economic development. An urban centre's primary function is to act as a service centre for the hinterland around it. Though urban development need not be always synonymous with economic growth, spatial distribution of growth is linked with the emergence of a formal manufacturing sector.

3.129 Interestingly, most of the States, either with higher per capita SDP growth than the all-India average, or those experiencing a rise in per capita growth rate during the 1990s visa-vis the 1980s are also the States with higher than all-India average urbanisation as per the 2001 census. There are, however, some exceptions like Punjab and Haryana with quite a high degree of urbanisation but which are placed relatively low in the growth ladder in the 1990s. Kerala, despite its high growth performance is somewhat low in terms of the percentage of urban population (Table 3.41).

Table 3.41 : Percentage of Urban Population of the Indian States in 2001

		(Per cent)
	State	Percentage of Urban
		Population
	1	2
1.	Tamil Nadu	43.86
2.	Maharashtra	42.40
3.	Gujarat	37.35
4.	Karnataka	33.98
5.	Punjab	33.95
6.	Haryana	29.00
7.	West Bengal	28.03
8.	India	27.78
9.	Andhra Pradesh	27.08
10.	Madhya Pradesh	26.67
11.	Kerala	25.97
12.	Uttaranchal	25.59
13.	Rajasthan	23.38
14.	Jharkhand	22.25
15.	Uttar Pradesh	20.78
16.	Chhattisgarh	20.08
17.	Orissa	14.97
18.	Bihar	10.47
Source: Provisional Population Totals: India. Census of India 2001.

3.130 The regional growth profile during the 1990s suggests that higher income capabilities did not get translated into higher growth trajectories for all the States. In terms of a 'before-after' comparison, most of the States with higher average per capita SDP growth rate during the 1990s (such as, Karnataka, Gujarat, Tamil Nadu, Maharashtra, or Andhra Pradesh) are those which embarked on reform programmes. The rank correlations of the States as per various indicators of development are given in Table 3.42.

Table 3.42 : Correlation between the Select Relative Rankings of States
	Growth	Electricity	Roads	Literacy	Urbanisation
1	2	3	4	5	6
Growth	1.00	0.22	0.18	0.57	0.48
Electricity		1.00	0.50	0.07	0.69
Roads			1.00	-0.08	0.13
Literacy				1.00	0.47
Urbanisation					1.00
Note:'Growth' refers to improvement in Growth Rate during 1991- 92 through 1999-2000, 'Electricity' refers to availability of electricity in 2000-01, 'Roads' refers to State-wise Roads as % to All India Roads in 1997, 'Literacy' refers to Literacy Rates in 2001, 'Urbanisation' refers to Urbanisation in 2001.

3.131 All the social and infrastructural indicators exhibit a high correlation with growth of per capita SDP reflecting the fact that in the quest for growth, the role of social and physical infrastructure is crucial. Rank correlations between growth and literacy, on the one hand, and growth and urbanisation, on the other, are found to be highly positive and significant. The relationship between ranking as per infrastructural indicators and ranking as per growth, though positive and significant, turns out to be low.

3.132 The growth profile of the Indian economy has many dimensions interacting with the heterogeneous nature of the growth efforts of various States. Given the vastness and diversity of the economy, it would be difficult to expect that all States would grow at the same rate (Ahluwalia, 2000). Reforms appear to be impacting on growth differentials of the States in the 1990s. This has implications for competition among States for scarce resources for the acceleration of growth. Despite an absence of a systematic evidence on the causal relationship between reform and growth, the strength of associations in the behaviour of growth at the level of the States suggests the possibilities of 'growth arbitrage' being created by reforms.

VI. CONCLUDING OBSERVATIONS

3.133 The current slowdown in the Indian economy has become a subject of intense debate. Empirically the current phase appears to represent a loss of speed rather than a halt in growth. Cyclical patterns in activity are detected but these are of limited duration and impact. The Indian economy is characterised by stable and converging cycles which could be modulated with counter cyclical policies. The impulses for growth can be generated and nurtured primarily by releasing the structural constraints, which can shift the potential growth frontier outwards. So far, private consumption has been providing the predominant contribution to aggregate demand relative to investment. Discretionary fiscal stabilisers mainly in the form of government consumption have been holding up aggregate demand over the period of the downturn.

3.134 Capital formation has been slowing down across all three sectors of the economy. The size of accelerators suggests that greater investment needs to be directed towards manufacturing so as to revitalise growth. Manufacturing slowdown is reflecting as much a demand slack as a fall in capacity utilisation and gaps in availability of infrastructure, particularly power. Structural constraint to industrial growth are more dominant than cyclical variations. Productivity changes and technological progress have been identified in the 1990s as important sources of industrial growth.

3.135 Abstracting from measurement issues, capital formation in agriculture has been declining which is a matter of concern. The lack of capital has been a primary impediment to the adoption of new technology. Given the overall resource constraint, a conscious choice between subsidies and investment is being imposed on the conduct of public policies for agriculture. Integration of Indian agriculture with international markets would be necessary in the context of commitments to WTO, complemented by well developed domestic future markets for agricultural commodities.

3.136 The services sector has imparted resilience to the economy, particularly in times of adverse agricultural shocks as also during cyclical downturns in industry. A notable feature of the structural transformation of the services sector has been the growth of skill intensive and high value added sectors. As growth gathers momentum, the demand for these services is expected to increase more than proportionately and this, in turn, would reinforce growth itself.

3.137 Issues relating to regional growth have added a new dimension to the size and quality of the growth process. Differential growth performances in the States is reflective of the varying degree of reform penetration. In some developing economies (Brazil, China), regional reforms have taken the lead. In India, given the federal structure, there are several areas where it is the States which can initiate reforms. In this context, it is necessary to emphasise the role of States in improving the provision and quality of the two key infrastructural services which would determine the sustainability of reforms: education and health. Greater involvement of the private sector in the production and distribution of health and education services in a cost-effective manner can have beneficial externalities for the growth process. These services, which currently account for about 5 per cent of GDP, help to develop the social and distributive infrastructure. They enrich the quality of human capital and expand demand for the output of other sectors. Improvements in delivery of health and education can generate a variety of specialised occupations with potential for synergisation. Investing in these sectors will not only improve quality of human capital but can also become a source of higher productivity and growth.

---

1 (1) GDSR = -25.75 + 6.61 LPINC + 0.08 RDRT + 3.4 IR + 0.74 AR(1) R2 = 0.88, DW = 1.80
(2.78)*		(1.91)*		(2.12)*		(5.90)*
(2) PVSR = -25.65 + 3.66 LRPDI + 0.05 RDRT + 3.84 IR + 0.53 PVSR(-1) R2 = 0.95, Durbin’s h = 1.45
(3.9)*	(3.11)*		(4.06)*		(4.43)*

GDSR: Gross Domestic Saving Rate, LPINC : log of real per capita income (per capita GDP at factor cost at constant prices), RDRT : real deposit rate (i.e., one year deposit rate less inflation rate derived form GDP deflator), IR : Intermediation Ratio, PVSR: Private Saving rate, LRPDI : log of real per capita disposable income. t- statistics are in parenthesis and * indicates significance atleast at 10 per cent level.

2 Investment (Gross Capital Formation-GCF)

(1) GCF = 4807 + 0.20 GDPFC + 0.69 DGDPFC(-1) - 699 rl + 12629D1 + 26938D2														R2 = 0.96, DW=1.87
(14.02)	(3.57)							(-2.20)		(2.95)			(2.18)
(2) GCFa = 4018 + 0.04 D GDPFC(-1) + 0.18 Ips														R2 = 0.66, DW=1.14
(2.0)	(3.0)
(3) GCFm = -19261 + 0.61 D GDPFC(-1) - 811 rl + 2.2 Ips														R2 = 0.76, DW=1.82
(2.4)					(-2.4)			(4.7)
(4) GCFs = -3889 + 0.22 D GDPFC(-1) - 304 rl + 1.41 Ips														R2 = 0.84, DW=1.96
(2.6)		(-2.02)				(6.54)

GDPFC: GDP at factor cost, rl : real bank lending rate, D1: dummy for the 1980s, D2 : dummy for 1993-96, GCFa, GCFm and GCFs are private investments in agriculture, manufacturing and services, respectively, Ips : Public investment in services. All the t-statistics (presented in parentheses) are significant atleast at 10 per cent level.

3 (1) PFCE=108913 + 0.60GDPFC + 0.8877 AR(1)					R2 = 0.999; DW=2.07
	(23.32)*		(9.4)*
(2) GFCE = -9541.66 + 0.14GDPFC + 0.7834 AR(1)					R2 = 0.988; DW=1.21
		(14.2)*		(5.3)*

PFCE : Private Final Consumption Expenditure, GDPFC: GDP at factor cost, GFCE : Government Final Consumption Expenditure. Figures in brackets represent t-statistics and * indicates significance atleast at 10 per cent level.

4 Finance Minister’s Address to Parliament Consultative Committee, November 6, 2001.

5 Illustratively, in case a small farmer takes credit from a large farmer and is subsequently forced to sell his agricultural product at much lower price than post-harvest prices, he is said to be affected adversely by factor-product market inter-linkage. In case the borrowing farmer is forced to work on the creditor’s land at critical period like sowing, then the exploitation is in terms of adverse factor - factor market inter-linkage.

6	LCU = - 1.785 + 0.583 LPFCE - 0.288 LWPIFUEL + 0.481LCU{t-1} - 0.064 DUM91
		(2.97)	(-3.13)	(3.70)	(-4.32)

Where LCU: log of capacity utilisation index, LPFCE: log of real private final consumption expenditure, LWPIFUEL: log of WPI of fuel, power, light and lubricants, DUM91: dummy variable for post-1991 period. Figures in brackets are t-statistics. The coefficients are significant at 1 per cent level.

7 The conversion of nominal value added into the real value added is done either with single deflation or double deflation method. In the case of single deflation method both nominal output and nominal material inputs are deflated by output price index, while under double deflation method, the nominal output is deflated by output price index and nominal material input is deflated by input price index.

8 Log (RGDPs) = a+a₁ Log (RGDP) + a₂ Log (PDEFs / PDEFgdp)

RGDPs = Real GDP of Services Sector as a proxy for expenditure on services,

RGDP = real GDP,

PDEFs = Price Deflator in Services Sector and

PDEFgdp = Price Deflator of overall GDP

9 Log (S) = b + b₁ Log (RPCGDP)

S = Share of services/ sub sectors in real GDP RPCGDP = Real Per Capita GDP

10 There is no uniform series of real SDP over the period under consideration. While the real SDP from 1981-92 through 1993-94 are available with 1980-81 as the base, the data from 1993-94 through 1999-2000 are available with 1993-94 as the base. Hence, growth rates from 1981-82 through 1993-94 are calculated over the SDP numbers with 1980-81 base, while the same from 1994-95 through 1999-2000 are calculated over the SDP series with 1993-94 base.