ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Regional Dimension of Growth and Employment

Regional inequality has emerged as a key issue in recent discussions of development policy. States within India differ greatly in terms of economic growth and employment potential. This paper examines some aspects of this regional employment growth in India during 1983 to 2004-05. The results confirm widening interstate disparities in income in the first quinquennium of the 21st century, a continuation of the trend of the 1990s. Urban employment occurs strongly in initially urbanised states. All states are found to be diversifying, but at a slower pace in low income states. A geographic concentration of skilled labour is observed in financial and business services.

SPECIAL ARTICLE

Regional Dimension of Growth and Employment

K V Ramaswamy

Regional inequality has emerged as a key issue in recent discussions of development policy. States within India differ greatly in terms of economic growth and employment potential. This paper examines some aspects of this regional employment growth in India during 1983 to 2004-05. The results confirm widening interstate disparities in income in the first quinquennium of the 21st century, a continuation of the trend of the 1990s. Urban employment occurs strongly in initially urbanised states. All states are found to be diversifying, but at a slower pace in low income states. A geographic concentration of skilled labour is observed in financial and business services.

This paper is based on a study done for the Institute of South Asian Studies, National University of Singapore, Singapore. I am grateful to an anonymous referee for helpful comments on an earlier draft of this paper. The usual disclaimer applies.

K Ramaswamy (swamy1378@gmail.com) is currently at the Institute of South Asian Studies, National University of Singapore, Singapore.

Economic & Political Weekly december 8, 2007

T
he regional disparities (interstate) in economic well-being are an unmistakable feature of economic growth and change in India. In the years prior to independence, a pattern of ‘agglomerated’ growth emerged, with islands of concentrated growth but having very weak dispersal effects…As late as 1948, the presidency states (Bombay, Madras and Calcutta) accounted for 76.7 per cent of the total industrial workers and 77 per cent of industrial production… the share of mineral rich states of Bihar, Orissa and Madhya Pradesh was 9.6 per cent (industrial production)… The southern region around Madras and Bombay, and especially what later became the state of Gujarat, was better placed and had a better start in terms of agriculture and industry [Krishna Bharadwaj 1982: 609].

Later studies of regional disparities during the period of economic planning in India observed that the impulses of growth are more widely dispersed than before but confirmed the persistence of wide disparities in development levels [Srivastava 1994 among others]. Whether these development disparities have tended to accentuate or diminish in recent years of reforms, trade liberalisation and grater integration with the global economy is an important question with social and political economy implications.

The issue of regional disparities in employment in recent years of openness is important simply because labour markets are the key avenue through which international trade and investment openness affects the domestic economy. The problem of regional income inequalities has attracted much attention in recent years [Sachs et al 2002; Ahluwalia 2001; Shetty 2003; Bhattacharya and Sakthivel 2004a; Bagchi and Kurian 2005 among others]. Most of these have focused on the disparities in per capita incomes and report a tendency for diver gence. However, studies of interstate disparities in employment opportunities or labour market outcomes are very few and deserve equal attention. Among them, studies by Bhattacharya and Sakthivel (2004b) and Ahsan and Pages (2006) constitute the recent key studies with their detailed analysis of interstate differences in employment outcomes. The time period covered in these two studies from 1983 to 1999-2000 corresponding to the then availability of national sample survey (NSS) employment and un employment data. Other studies have focused on the impact of labour regulations and trade liberalisation on manufacturing employment and labour demand [Besley and Burgess 2004; Hasan, Mitra and Ramaswamy 2007]. These two econometric studies mainly utilise state level data on manufacturing industries available in the annual survey of industries (ASI).

In this paper, I investigate the growth and structure of employment in 14 major states of India during 1983 and 2004-05. This will help maintain comparability with two important recent studies of regional income disparities, namely, Ahluwalia (2001) and Sachs et al (2002). Sachs et al (2002) have carried out both the sigma and beta tests of convergence for the 14 major states using per capita gross state domestic product (GSDP) data for the period 1980 to 1998. They found that 14 major Indian states for the period are diverging overtime. Major states in India exhibited a lack of both sigma and beta convergence. Their analysis leads them to suggest that the forces of convergence are weak in India. Ahluwalia (2001) in his comparative evaluation of the economic performance of states observed that the estimated Gini-coefficient

Table 1: Employment by Sector – 1983 to 2004-05 (in millions)

Sector 1983 1993-94 1999-2000 2004-05

Agriculture 207.1 239.5 240.3 258.8

Mining and quarrying 1.8 2.7 2.3 2.5

Manufacturing 32.3 39.8 43.8 55.9

Electricity, water, etc 0.8 1.4 1.0 1.2

Construction 6.8 12.1 17.5 26.0

Trade (retail+ wholesale), hotels and restaurants 19.1 28.4 40.9 49.6

Transport, storage and communications 7.5 10.7 14.6 18.6

Other services such as finance, business, public administration, education, etc 26.7 39.8 38.1 45.4

All sectors 302.3 374.3 398.4 458.0

Source: NSS employment and unemployment surveys adjusted for population censuses. Employment is measured by number of workers by the usual principal and subsidiary status.

Table 2: Employment Growth Rates by Sector – 1983 to 2004-05*

Sector 1993-94 1999-2000 2004-05 over 2004-05 over over 1983 over 1993-94 1999-2000 1993-94

Agriculture 1.4 0.1 1.5 0.7

Mining and quarrying 3.7 -2.8 2.4 -0.4

Manufacturing 2.0 1.6 5.0 3.1

Electricity, water, etc 4.8 -4.8 3.1 -1.2

Construction 5.7 6.4 8.2 7.2

Trade (retail+ whole sale), hotels and restaurants 3.8 6.3 3.9 5.2

Transport, storage and communications 3.4 5.3 4.9 5.1

Other services such as finance, business, public administration, education, etc 3.9 -0.7 3.6 1.2

All sectors 2.1 1.0 2.8 1.9

*Annual compound growth rates. Source: NSS employment and unemployment surveys adjusted for population censuses. Employment is measured by number of workers by the usual principal and subsidary status.

(a key measure of income inequality) has increased from about

0.16 in 1986-87 to 0.23 in 1997-98.

I have estimated the Gini-coefficients for two years, 1993-94 and 2004-05, using per capita GSDP data. It is found to have risen from 0.28 in 1993-94 to 0.36 in 2004-05. These 14 states have large populations and together have a share of more than 93 per cent of India’s population.1 My analysis of employment growth is primarily based on the quinquennial NSS employment and unemployment surveys (EUS) spanning the period 1983 to 2004-05. It covers the longer period and makes possible analysis of crossstate employment growth and structure in the years of ecothe concentration and diversification of the regional employment structure is presented. This section includes a discussion of dualism (organised and unorganised sectors) within sectors across states. Section 3 contains a discussion of interstate differences in labour productivity and educational attainment of population. Section 4 presents a summary of the main findings and con cluding remarks.

1 Growth in Interstate GSDP and Employment

It is useful to summarise the employment changes at the all-India level. The employment growth profile is presented in Table 1 (absolute numbers) and in Table 2 (growth rates). The first fact to be noted is the constancy in the growth rate of aggregate employment in the 1980s (2 per cent during 1983 to 1993-94) and in the period spanning 1993-94 to 2004-05 (1.9 per cent). Second, the declining growth rate of employment in agriculture from 1.4 per cent in the 1980s to 0.7 per cent in the period 1993-94 to 2004-05. Third, the acceleration in employment growth rate in the construction sector. Fourth, the recovery of employment growth rate in the manufacturing sector in the recent period that is, 19992000 to 2004-05. Fifth, relatively higher rates of employment creation in the three service sectors, namely, trade and hotels, transport and communication and other services that include banking and business services. Sixth, one observes a greater change in employment shares of total employment in the postreform period compared to the pre-reform period.3

This aggregate picture serves as the background for my state level analysis.4 The state level estimates of growth in GSDP and per capita GSDP cover the period 1993-94 to 2004-05. This period is further subdivided into two sub-periods, namely, 1993-94 to 1999-2000; 1999-2000 to 2004-05. This corresponds to the NSSEUS years and enables me to systematically relate output and employment growth across states.5 In the analysis of employment differences, I cover the period 1983 to 1994. However, GSDP growth for that period is not presented due to unavailability of a consistent series based on 1993-94 prices. I have divided the 14 states into three groups by ranking each state based on its per capita GSDP for the year 1993-94. The bottom five are states with relatively low income, the middle four are medium income states and the top five are the relatively rich states.6

Figure 1: Annual GSDP Growth Rates 10 (in %)

nomic reform beginning in 1991. I pay greater attention to the problem of sectoral concentration and within sector changes 9 in terms of duality (organised vs unorganised). The employ-8 ment estimates are based on the estimates of all-India population for the four survey years (January-December) 1983, 6 and (July-June) 1993-94, 1999-2000 and 2004-05.2 This is supplemented by other sources for the organised sector 4 like the ASI and the data from the employment information system of the Directorate General of Employment and 2 Training (DGE&T).

Following the introduction, this paper is organised 0 into four sections. Section 1 begins with a brief review of growth and employment trends in India for the period 1993 to 2003-04 and presents the state level analysis of GSDP growth and employment. In Section 2, an analysis of

1993-94 to 1999-2000 1999-2000 to 2004-05 Source: National Accounts Statistics available at www.mospi.nic.in BiharOrissaUttar PradeshRajasthanMadhya PradeshWest BengalAndhra PradeshKarnatakaKeralaTamil NaduGujaratHaryanaMaharashtraPunjab

december 8, 2007 Economic & Political Weekly

SPECIAL ARTICLE

What have been the growth trends in GSDP and per capita GSDP during 1993-94 to 2004-05? The estimates are shown in Table 3. First, the middle four and the top five ranking states (in terms of per capita GSDP) have grown at a faster rate than the bottom four states in the entire period as well as the two sub-periods. Among the bottom five, Rajasthan has performed above average. Andhra Pradesh and Punjab are the under performing states in the medium (middle four) and high income groups (top five). The GSDP

Table 3: Annual Growth of GSDP and Per Capita GSDP – 1993-94 to 2004-05 (in %) *

GSDP at Constant Prices* Per Capita GSDP*
Rank* State 1993-94 to 1999-2000 1993-94 to 1993-94 to 1999-2000 1993-94 to
1999-2000 2004-05 2004-05 1999-2000 to 2004-05 2004-05
14 Bihar 4.4 4.7 4.6 1.7 2.4 2.0
13 Orissa 4.3 5.9 5.0 3.0 4.3 3.6
12 Uttar Pradesh 4.6 4.2 4.4 2.3 1.0 1.7
11 Rajasthan 8.2 4.8 6.6 5.5 2.5 4.1
10 Madhya Pradesh 5.4 2.9 4.3 3.4 0.9 2.2
Mean of bottom five 5.4 4.5 5.0 3.2 2.2 2.7
9 West Bengal 7.1 7.0 7.1 5.4 5.6 5.5
8 Andhra Pradesh 5.5 6.5 5.9 4.1 5.4 4.7
7 Karnataka 7.6 6.1 6.9 6.0 4.7 5.4
6 Kerala 5.6 6.8 6.2 4.5 6.2 5.3
Mean of middle four 6.2 6.2 6.2 5.0 5.5 5.2
5 Tamil Nadu 6.6 4.1 5.5 5.5 3.1 4.4
4 Gujarat 7.8 6.7 7.3 6.1 4.3 5.3
3 Haryana 5.9 6.9 6.3 3.8 4.1 3.9
2 Maharashtra 6.2 5.0 5.7 4.1 3.4 3.8
1 Punjab 4.8 3.9 4.4 1.9 -6.7 -2.1
Mean of top five 6.3 5.3 5.8 4.3 1.6 3.1
All 14 states 6.1 5.3 5.7 4.1 2.5 3.4
Coefficient of variation 23.1 24.5 18.7 36.6 128.2 61.3

* Average annual compound growth rates. The estimate of group mean is unweighted. **Rank based on per capita GSDP 1993-94. Source: Estimates based on NAS available at www.mospi.nic.in and CMIE-NAS, October 2006.

growth rate of West Bengal is similar to Gujarat, the fastest growing state in this period. The interstate disparities in per capita income has widened as suggested by the rising coefficient variation

(CV) in the second sub-period (128 from 36.6). This is partly due to the poor performance of Punjab whose growth rate is negative in the period 1999-2000 to 2004-05. Even if one excludes Punjab, the CV increases from 33 to 45.7. The interstate differences in GSDP growth rates are shown in Figure 1 (p 48). The employment outcome of this differential output growth needs to be analysed. In Table 5 (p 51) and Figure 2, the estimates and graph of growth rates in employment for the same set of states for the two sub-periods are shown respectively.

five states, namely, Uttar Pradesh and Rajasthan. Among others, four states have recorded impressive growth rates in the second period, namely, West Bengal, Karnataka, Haryana and Maharashtra. The relevant question is what has been the nature of this employment growth across states in terms of rural urban divide and formal and informal composition? Which sectors have grown and which have fallen behind? This will determine the quality of employment growth in a broader structural perspective.

A comparison of urban and rural employment growth rates between pre-liberalisation years and post-liberalisation years is presented in Table 5 (p 51). The urban bias in relative growth rates of employment is evident. Across the 14 states, employment has grown faster in urban areas in both the sub-periods of the post-liberalisation period (1993-94 to 1999-2000 and 19992000 to 2004-05). Kerala is the only exception with low growth rates in both urban and rural areas. In the top five states, the average urban employment growth is higher than that in the bottom five states. A significantly positive development has been the recovery of rural employment, on the average, in the second sub-period (1999-00 to 2004-05) across all the states. However, two states have experienced a slow down in rural employment, namely, Gujarat and Tamil Nadu. It is negative growth in Tamil Nadu, a state of high urban employment growth. Is there a positive relationship between the initial level of urbanisation and urban employment growth? We observed a significant positive correlation between initial urbanisation in 1993-94 and total urban employment growth rate for the period 1993-94 and 2004-05 (Figure 3, p 50).

This clearly implies that benefits of growth in terms of employment have largely gone to urbanised states in the years since liberalisation. Perhaps this is the dark side of employment growth in India aggravating interstate inequalities. Figure 4 (p 51) exhibits the interstate differences in rural and urban employment growth rates for the period 1999-2000 to 2004-05.

2 Regional Employment Structure

The traditional Kuznets-Chenery perspective of structural transformation suggests a reallocation of labour from agriculture to manufacturing and services as per capita income rises. The

Figure 2: Annual Employment Growth (in %)

The second sub-period (1999-2000 to 2004-05) is a 6 period of recovery of employment growth in India. Job creation has reappeared in the Indian economy after a period of jobless growth in the 1990s. This is correctly reflected in

4

the state-wise employment growth trends in Table 4 (p 50). In the 14 major states employment grew by 2.8 per cent per annum. This is similar to the all-India growth rate that we referred to earlier (Table 2). This recovery in employment 2 growth is across the 14 states with Kerala as the exception (growth of 1.3 per cent in the first against 1.2 per cent in the second). Recovery is stronger in the higher income states 0

(3.2 per cent from 1.4 per cent). Employment growth in the bottom five states with a share of more than 44 per cent of the workforce, has picked up to grow at the average of 14 states (2.8 per cent) in the second period. We may note the impressive employment performance of two of the bottom evolution of sectoral shares in India is observed to be unusual and may have far-reaching implications for employment growth [Kochhar et al 2006 among others]. India’s share of services in GDP has risen rapidly from 37 per cent to 49 per cent between

Economic & Political Weekly December 8, 2007 BiharOrissaUttar PradeshRajasthanMadhya Pradesh.West BengalAndhra Pradesh.KarnatakaKeralaTamil NaduGujaratHaryanaMaharashtraPunjab 1993-94 to 1999-2000 1999-2000 to 2004-05 Source: NSS employment and unemployment survey results of 50th, 55th and 61st rounds.

Table 4: Annual Growth of Employment – 1993-94 to 2004-05 (in %)

Employment Employment Growth
Rank State Share in 1993-94 to 1999-2000 to 1993-94 to
1993-94 1999-2000 2004-05 2004-05
14 Bihar 9.0 2.0 2.2 2.1
13 Orissa 4.1 0.8 2.5 1.6
12 Uttar Pradesh 15.5 1.1 3.8 2.3
11 Rajasthan 6.3 0.8 3.0 1.8
10 Madhya Pradesh 9.1 1.1 2.7 1.8
Average of bottom five 43.8* 1.1 2.8 1.9
9 West Bengal 7.6 0.8 3.0 1.8
8 Andhra Pradesh 10.3 0.2 1.9 1.0
7 Karnataka 6.3 0.8 3.1 1.8
6 Kerala 3.3 1.1 1.3 1.2
Average of middle four 27.6* 0.8 2.4 1.5
5 Tamil Nadu 8.1 0.0 1.7 0.8
4 Gujarat 5.5 2.3 2.6 2.4
3 Haryana 1.9 1.2 5.6 3.1
2 Maharashtra 10.9 1.0 3.4 2.1
1 Punjab 2.3 2.6 2.8 2.7
Average of top five 28.6* 1.4 3.2 2.2
All 14 states 100.0 1.0 2.8 1.8

*Sum of shares states, average annual compound growth rates (per cent); rank based per capita GSDP 1993-94.

1980 and 2000. However, the rise in employment share – from

18.6 per cent to 22.4 per cent – during the same period is marginal. This implies a rapid increase in labour productivity in the services sector, perhaps due to growth in skill-intensive services [Gordon and Gupta 2004]. This all-India aggregate picture hides many regional variations. The regional variations in per capita income could perhaps be due to the uneven spread of service sector employment both in quantity and quality. Higher income states will have a greater share of productive services, while the low income states may end up with low productivity employment that is actually a spillover of lack of alternative productive employment opportunities. A preliminary look at the evolution of sectoral diversification of state economies is likely to throw some light on this issue.

$ 9,000 (constant 1985 US dollars). India and her constituent states are far below this level of income and are likely to experience increasing diversification. However, it is important to what know the level and speed of change in diversification underlying the present ongoing development process are. This will reveal the inertia or structural backwardness constraining the inter-regional differences.

The first cut would be the employment shares of three important sectors, namely, agriculture, manufacturing and services. This is presented in Table 6 (p 51) for two selected years 1993-94 and 2004-05 – a gap of 11 years; a reasonable period to consider as structural change is a long-run process. The interesting question is whether trade and structural reform years would show-up inter sector labour reallocation or structural inertia. The study by Wacziarg and Wallack (2004) that examined 25 liberalisation episodes could not detect any dramatic or increased structural shifts in employment shares across the nine 2-digit sectors. At the all-India level, I had observed greater changes in employment shares relative to the change during the pre-reform years of 198394 [Ramaswamy 2007]. This is an important finding because lack of change in employment shares would have suggested an absence of resource movements to gain from comparative advantages. In other words, intersectoral flows of workers are found to be greater in the post-reform period. Whether this is translated into welfare gains is another issue that we will take up later.

At the sub-national level, expectedly, the bottom five states (lower income) have higher shares in agriculture than the higher income states (middle and top). It is important to note that Madhya Pradesh is the state with the highest share in agriculture and Kerala is the state with the lowest share in agriculture to begin with in 1993-94. However, in 2004-05, status quo has been maintained by these two states. More interesting are the cases of West Bengal and Tamil Nadu. In 1993-94, West Bengal had an employment share (48.8 per cent) close to that of Kerala but had a higher share in manufacturing (19.9 per cent). In 2004-05, Tamil Nadu

has become the state with

Figure 3: Initial Urbanisation and Urban Employment Growth – 1993-2004*

The pattern of sectoral

diversification along the de

velopment path has been

recently examined by Imbs

and Wacziarg (2003). Their

detailed empirical study

showed a u-shaped pattern

in sectoral concentration

– countries begin their de-highest share in manufac

AP BH GU HR KA KE MP MH OR PU RA TN UP WB 0 5 10 15 20 25 30 35 40 45 0 1 2 3 4 5 Urbanisation 1993-94 (in %)

turing (closely followed by

Gujarat), the second highest

in services and the second

lowest share in agricul

ture (next to Kerala). West

Bengal and Tamil Nadu are

similar in terms of services

sector share. We need to

Average Growth Rate of Urban Employment (in %) velopment journey at a high * The urban growth rate for Bihar is negative but is taken as zero with a view not to clutter the diagram. determine whether these Source: NSS employment and unemployment survey results of 50th, 55th and 61st rounds.

level of concentration (low similarities and differences

income levels) and diversification increases reaches a minimum imply much more in terms of productivity of employment. I will level and then the economic activity structure starts concentrating take up this aspect again when I discuss interstate differences in again. This scheme therefore suggests that there are two stages sectoral labour productivity levels (sectoral NSDP per worker, see of diversification in the development process. The first one is of Table 6). increasing diversification followed by one of increasing concentration. However, the minimum point occurs quite late in the evolu-2.1 Sectoral and Spatial Concentration tion process of sectoral diversification. This is interpreted to sug-I have used the Herfindhal-Hirschman index (HH index) of congest that countries diversify most of their development path. They centration to measure sectoral concentration of employment. estimate it to occur at the per capita income level of approximately The HH index is one of the most commonly used measures of

december 8, 2007 Economic & Political Weekly

SPECIAL ARTICLE

concentration of output and employment in the literature on regional economics dealing with spatial concentration of activity.7 I utilise the nine sector classification of the national industrial classification (NIC-98) followed by the NSS employment surveys. I estimate the HH index for each of the four quinquennial survey years, namely, 1983, 1993-94, 1999-2000 and 2004-05.

The HH index is defined as the sum of the squares of (percentage) employment shares in each state. ∑ Ei2; where Ei = employment share of the ith sector in a state and i= 1…9.

Table 5: Annual Employment Growth by State – Urban vs Rural (in %)

Urban Rural State 1993-94 over 1999-2000 2004-05 over 1993-94 over 1999-2000 2004-05 over 1983-84 over 1993-94 1999-2000 1983-84 over 1993-94 1999-2000

Bihar 0.1 2.4 3.9 1.3 1.9 2

Orissa 3.7 1.5 3.4 1.7 0.7 2.4

Uttar Pradesh 3 2.8 4.6 1.5 0.7 3.7

Rajasthan 3 2 4.3 2.2 0.5 2.7

Madhya Pradesh 3.4 3.1 4.5 1.9 0.7 2.2

Average of bottom five 2.6 2.4 4.1 1.7 0.9 2.6

West Bengal 2.6 1.6 3.2 2.1 0.4 2.9

Andhra Pradesh 3.9 0.1 4.3 2.3 0.2 1.3

Karnataka 2.9 2.5 3.2 2.1 0.2 3

Kerala 4.3 0.7 0.3 0.2 1.3 1.6

Average of middle four 3.4 1.2 2.8 1.7 0.5 2.2

Tamil Nadu 3 3 4.9 1.1 -1.4 -0.2

Gujarat 3.4 2.8 4.4 1.6 2.1 1.8

Haryana 4.2 2.6 5 2.5 0.7 5.8

Maharashtra 3.7 2.4 4.9 1.6 0.4 2.6

Punjab 2.7 3.9 3.9 0.1 2 2.3

Average of top five 3.4 2.9 4.6 1.4 0.8 2.5

Total of 14 states 3.3* 2.3 3.8 1.7* 0.7 2.4

Source: Chadha and Sahu (Table 12, 2002) for 1993-94 over 1983-84 and others are author’s estimates based on NSS employment surveys (EUs).

Table 6: Employment Share by Sector – 1993-94 and 2004-05 (in %)

1993-94 2004-05 State Agriculture Manufacturing Services Agriculture Manufacturing Services

Bihar 76.7 4.9 15.6 68.9 7.2 18.0

Orissa 73.7 7.5 15.0 62.3 11.4 19.1

Uttar Pradesh 68.4 8.7 20.1 60.6 12.3 20.9

Rajasthan 69.2 6.2 15.3 61.3 9.1 18.2

Madhya Pradesh 77.7 5.5 13.4 69.1 7.5 18.2

Average of bottom five 73.1 6.6 15.9 64.4 9.5 18.9

West Bengal 48.8 19.9 27.1 45.7 17.5 31.6

Andhra Pradesh 67.1 9.2 19.6 58.4 11.0 24.8

Karnataka 65.1 10.7 19.7 60.8 10.6 23.8

Kerala 48.3 14.3 29.6 35.5 14.4 37.7

Average of middle four 57.3 13.5 24.0 50.1 13.4 29.5

Tamil Nadu 52.6 18.0 24.8 41.2 21.1 30.9

Gujarat 58.9 15.2 21.4 54.8 17.1 23.1

Haryana 56.9 9.1 27.7 50.0 13.5 27.7

Maharashtra 59.4 11.3 25.1 53.1 12.5 28.7

Punjab 56.4 10.3 28.1 47.4 13.5 29.8

Average of top five 56.8 12.8 25.4 49.3 15.5 28.0

All 14 states 64.5 10.5 20.7 57.0 12.4 24.1

The row sum of sectoral shares does not sum to 100 as mining, construction and electricity have been left out. Source: NSS employment survey 1993-94 and 2004-05.

The HH index reaches a maximum value of 10,000 when only one sector has all the employment (100 per cent) and has a lower bound of 1111, that is all the sectors have an equal share (note that the lower bound varies with the number of sectors), in the case of nine sectors.8 Lower the estimated HH index, more equal the sector shares and more diversified is the economy (states of India in our case). The estimated HH index is shown in Table 7 (p 52). All the 14 states in our sample clearly show the tendency for diversification (the change in HH index is negative across all states). The HH index for the aggregate of 14 states show a decline of 18 percentage points over the period 1994 to 2004. Relative to

Economic & Political Weekly December 8, 2007

this average, only one state in the low income category, namely, Orissa shows a substantial decline (25 per cent). In the middle income group, Kerala’s diversification is higher (a decline of 29 per cent). In the top five group, Tamil Nadu stands out as a state with greater diversification tendency. Andhra Pradesh and Punjab look similar in their diversification trends. The economies of West Bengal and Karnataka show a substantially lower rate of diversification of economic activity relative to the average. It is important to note that low income states have more concentrated structures to begin with and are changing at a much slower pace. How has the level of initial diversification had an impact on employment growth prospects in different states?

More diversified states should have gown faster with the opening of the economy in the 1990s as they would be in a better position to take advantage of trade and growth opportunities. If this is true then a negative relationship between the initial HH index (concentration) and employment growth may be expected. In Figure 5 (p 52), I plot the HH index in the year beginning 1993-94 against the subsequent employment growth rates in the selected 14 states. The figure shows a positive correlation between the initial extent of diversification in a state and employment growth in the subsequent years. Interstate disparities in the level and changes in diversification is obviously the cause of interstate income disparities. Slow diversification of some of the major states such as Bihar, Uttar Pradesh, Madhya Pradesh and Rajasthan is certainly matter of concern for policy. At the same time, there may be concentration within sectors like registered manufacturing as reported by Kochhar et al (2006) that may accentuate the divergence tendencies. The slow growth of employment in low income states is partly due to the slow rate of diversification of economic activity in these economies.

Another important question that is often discussed in this context of regional disparities in developing countries is that of geographic (spatial) concentration of particular sectors (like manufacturing) across locations.9 Due to historical accidents, industrialisation began earlier in certain states. I examine whether the degree of geographic concentration (note the difference between sectoral concentration mentioned above and geographic con centration) of selected sectors has increased or decreased in recent years. Here, I estimate the spatial (or locational) HH index for each sector. The spatial HH index is defined as follows: ∑ (si – xi)2; where si is the employment share of a state in the ith sector and

Figure 4: Annual Employment Growth in Urban and Rural Areas – 1999-2004 (in %) 6

4

2

0

-1 Urban

Rural Source: NSS employment and unemployment survey results of 55th and 61st rounds.

BiharOrissaUttar PradeshRajasthanMadhya PradeshWest BengalAndhra PradeshKarnatakaKeralaTamil NaduGujaratHaryanaMaharashtraPunjab

xi is the state’s share in total employment in the economy (or aggregate of selected states)10 and i=1……..14.

This is estimated for three selected sectors, namely, manufacturing, services and a sub-component of services sector, namely, financial, real estate and business services. The services sector is defined as the aggregate of transport, trade, communication and financial services sectors. The last sector is estimated separately because of its nature as a skill-intensive sector that has come to prominence in recent years. The estimates of spatial HH indices for three selected sectors for four selected years are presented

Table 7: Sectoral Concentration of Employment by State*

State 1983 HH Index 1993-94 1999-2000 2004-05 Change in HH Index HH Index HH Index HH Index 2004 over 1994

Bihar 6061.2 6003.6 5500.9 4945.4 -17.6

Orissa 5479.2 5586.9 5186.8 4184.4 -25.1

Uttar Pradesh 5387.5 4910.9 4332.8 4015.7 -18.2

Rajasthan 5964.6 4960.9 4566.6 4049.0 -18.4

Madhya Pradesh 6291.2 6143.3 5599.3 4974.3 -19.0

West Bengal 3818.4 3042.5 2853.9 2740.5 -9.9

Andhra Pradesh 5045.9 4732.5 4525.0 3759.5 -20.6

Karnataka 4940.8 4508.3 4194.3 4009.1 -11.1

Kerala 3709.2 2866.3 2204.9 2019.9 -29.5

Tamil Nadu 3341.2 3317.7 2826.9 2485.1 -25.1

Gujarat 4681.8 3882.1 3868.6 3494.2 -10.0

Haryana 5408.1 3638.5 3219.8 3001.1 -17.5

Maharashtra 5269.6 3882.4 3555.5 3250.5 -16.3

Punjab 4924.1 3596.2 3274.7 2794.2 -22.3

All 14 states 5057.7 4434.1 4043.1 3630.3 -18.1

Source:* Estimates use employment shares of nine sectors in each state based on NSS employment-unemployment surveys.

Table 8: Geographic Concentration of Sectoral Employment – HH-Spatial Index

Sector/Year 1983 1993-94 1999-2000 2004-05

Manufacturing 75.9 146.0 94.6 74.0

Services 50.9 51.2 56.2 29.9

Finance, real estate and business services* 189.5 216.2 212.6 223.5

* This includes software services defined since 1999-2000. Source: Estimates based on NSS employment-unemployment surveys.

in Table 8. Is there a change in the geographic concentration of sectoral employment?

To begin with, note that the finance and business services sector is more concentrated than the other two, namely, manufacturing and total services in the initial year, that is, 1983-84. Over the next 20 years, concentration first increases and then declines in manufacturing. The level of concentration in 2004-05 is found to be similar to the concentration level in 1983-84. In the services sector, the aggregate of trade, transport and finance, concentration remains flat till 1993-94 and then declines in 2004-05. However, in the sub-group, finance and business services, it rises sharply in 1993-94 and shows some marginal decline in between but raises again in the last year 2004-05. This supports the proposition that skilled labour-intensive activities are becoming geographically concentrated. The flatness of aggregate services perhaps simply reflects the geographical spread of transport, retail and services like telecommunication and public administration with economic development.

2.2 Employment Growth: Organised or Unorganised?

India is well known as a classic case of the Lewisian dual economy with a small organised and a large unorganised sector. Many interesting and provocative questions have been asked about the continuity of this dichotomy in India. Is there an intensification of duality in recent years of trade reform? Is the service sector more dualistic than the manufacturing sector in terms of wage differentials? All these are pertinent questions. As we note the stagnation in agriculture sector jobs in the last 20 years, we see that most of the addition to the labour force is absorbed by the non-agricultural sector. Actually, this absorption mechanism has been driven by the unorganised informal sector.

Are there interstate differences in this changing structure of duality? What are the implications? It may be noted that official definition of the unorganised sector is much broader than the standard concept of informal manufacturing enterprises. In manufacturing, all factories with less than 10 workers or less than 20 (if they are not using power) are considered informal enterprises. However, the official unorganised sector includes all unincorporated household enterprises, partnership enterprises, cooperative enterprises, private and limited companies. These unorganised sector enterprises have created much employment in India across states. They are characterised by low wage and low (labour) productivity activities. The unorganised sector is also known to be the “waiting” sector where migrants from rural areas locate themselves before they can get a job in the urban organised enterprises. The working and labour conditions in this sector are well documented.

What has been the experience of Indian states with respect to unorganised employment? How they are estimated? The standard procedure for estimating employment in the unorganised sectors is the use of the residual method. In this method, the estimates of organised sector employment provided by the Directorate General of Employment and Training (DGET), based on their employment information system, are subtracted from the NSS based estimates of total employment in each sector. Following this method, I have estimated the growth rates of unorganised sector employment, absolute change in the private sector within the organised sector and the share of the organised sector in total employment in each of the 14 states (Table 9, p 53). The unorganised sector employment growth is uniformly positive across the 14 states. More importantly, the private sector within the organised sector has created substantial absolute number of employment in three states, namely, Andhra Pradesh, Karnataka and Gujarat.

In the aggregate, the public sector in India has shed jobs in the 1990s. At the same time, four states at the bottom of the income ladder (Bihar, Orissa, Uttar Pradesh and Madhya Pradesh) and two middle income states (West Bengal, Kerala) and one high income state (Maharashtra) have negative net employment

Figure 5: Initial Diversification and Employment Growth

MH WBKE TN AP KARA HA MH OR UP BH PU MP 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 HH Index 1993-94 0 1 2 3 4 5 6

Average Employment Growth 1993-2004 (in %) Source: NSS employment and unemployment survey results of 50th, 55th and 61st rounds.

december 8, 2007 Economic & Political Weekly

SPECIAL ARTICLE

growth in the private sector segment of the organised sector. The levels over time. Recall that the 14 states were ranked in terms of organised sector share within each state is lower in the bottom their per capita income in the base year 1993-94.

five states. It is lower I begin with the comparison of agriculture and manufacturing Unorganised (in %) than the average of 14 sectors in Table 10.

Table 9: Employment Disparities – Organised vs

Employment Growth Organised Sector Share

in 2004-05 over 1993-94 in Total Employment states (5.5 per cent). Let us focus on a comparison of the beginning year (1993-94)

State Unorganised Private 2005 1994In all the middle four and end year (2004-05) levels with a gap of 11 years. Expectedly,

Sector Sector Growth(%) (‘000)* and top five states, the the manufacturing sector labour productivity level is higher in

Bihar 2.4 -78.1 4.0 5.5 organised sector share all states. Manufacturing productivity in the bottom five states is Orissa 1.8 -2.3 4.5 5.6

is higher than the av-not only lower than the average but it is declining over the years.

Uttar Pradesh 2.7 -47.5 3.4 4.9 Rajasthan 2.1 8.6 4.4 5.6 erage except in And-The top five states have average productivity that is more than

Madhya Pradesh 2.2 -52.9 3.7 5.3 hra Pradesh (5.1 per the average in both agriculture as well as manufacturing. Tamil West Bengal 2.2 -34.1 6.3 8.8

cent). This provides Nadu appears to be an exception with lower than average pro-

Andhra Pradesh 1.1 207.9 5.1 5.2 Karnataka 2.0 305.1 6.9 6.9 us with an idea that ductivity in manufacturing. Recall that Tamil Nadu is a state with

Kerala 1.5 -22.0 8.8 10.5 the low income states a high rate of urban employment growth (in fact employment is Tamil Nadu 1.0 9.8 7.4 8.5

are likely to have pro-entirely urban!).

Gujarat 2.9 132.3 6.8 8.9 Haryana 3.8 32.1 7.1 9.7 portionately more low Next, I compare two service sectors that are presumably mod-

Maharashtra 2.6 -34.0 7.4 9.9 productivity jobs cre-ern and relatively skill intensive, namely, transport and commu-Punjab 3.3 13.9 7.2 10.4

ated in recent years. nication (T&C) and financial and business services (Table 11). The

All India 2.2 438.8 5.5 7 *Absolute change in number of employees. If this conjecture is differences are sharper with the five bottom states have lower Source: Estimates based on DGET data on organised sector

true, then the low in-productivity and declining in financial and business services.

employment as on March 1994 and March 2005. March 1994 data is from www.indiastat.com and March 2005 data is come states should It has three exceptionally high productive states, namely, West

available in quarterly employment review at www.dget.gov.in

have relatively lower Bengal, Andhra Pradesh and Karnataka. T&C fares better due to

labour productivity levels across sectors. This is taken up in the

Table 10: Interstate Differences in Labour Productivity Levels (Rs per worker)

next section. Agriculture Manufacturing State 1993-94 1999-00 2004-05 1993-94 1999-00 2004-05

Bihar 5,936 6,166 7,937 25,177 24,419 21,833

3 Employment, Labour Productivity and Education

Orissa 6,870 6,570 7,509 11,063 14,202 12,677 The structural transformation process of development is sup-Uttar Pradesh 8,437 10,373 9,675 21,248 18,033 16,393

Rajasthan 6,748 8,979 9,840 23,829 40,147 28,889

posed to create greater productive jobs not merely jobs of aver-

Madhya Pradesh 7,620 8,116 7,589 36,344 39,020 31,692

age rural sector productivity. This greater productivity drives

Average of bottom five 7,122 8,041 8,510 23,532 27,164 22,297 output growth and in turn, generates more employment for all. West Bengal 12,962 16,390 16,439 14,311 22,085 24,593

Andhra Pradesh 7,684 8,683 10,675 19,215 29,745 26,919

This process of change demands greater skilled (more educated

Karnataka 9,622 12,111 9,021 27,391 34,446 45,704

or number of years of schooling) labour. This simple stylisation

Kerala 13,858 18,051 15,598 16,570 21,018 18,532 gets complicated in dual economies like India with segmented Average of middle four 11,032 13,809 12,933 19,372 26,824 28,937

Tamil Nadu 8,943 10,870 13,933 25,744 29,714 25,552

labour markets (formal and informal) within sectors, whether it

Gujarat 9,133 9,025 12,292 37,975 66,077 56,802

is manufacturing or services. Greater employment growth in the

Haryana 22,427 25,298 23,186 60,924 69,269 59,572 non-agricultural sector may turn out to be low productivity jobs Maharashtra 9,366 10,865 8,704 57,020 72,803 58,433

Punjab 28,817 30,146 32,866 47,089 51,165 39,367

if it is mostly in low technology-low wage-low labour productivity

Average of top five 15,737 17,241 18,196 45,750 57,806 47,945

segments. As I observed earlier, the low income states suffer from

Total of 14 states 9,085 10,398 10,752 28,273 35,116 32,164 lower rates of diversification measured by HH indices of concentra-Source: NSDP from NAS and employment from NSS.

tion and added to that they have lower shares of organised sector

Table 11: Labour Productivity Level Differences (Rs per worker)

employment within their economies. Both these should constrain Transport and Communication Financial and Business Services State 1993-94 1999-2000 2004-05 1993-94 1999-2000 2004-05

them to lower average labour productivity levels. In order to test

Bihar 24,037 10,247 27,184 1,34,927 1,54,563 1,41,049

this proposition I estimated labour productivity for the selected

Orissa 34,861 50,024 65,545 3,55,060 2,57,256 1,66,289 six sectors for three selected years, namely, 1993-94, 1999-2000 Uttar Pradesh 25,204 28,652 41,718 1,81,407 2,05,846 1,73,369

Rajasthan 24,345 37,320 55,287 1,73,567 2,02,021 1,85,702

and 2004-05. These years correspond to the three quinquennial

Madhya Pradesh 36,847 48,261 76,019 2,34,735 2,93,653 2,02,015

NSSs. The output data at constant 1993-94 prices is available in

Average of bottom five 29,059 34,901 53,151 2,15,939 2,22,668 1,73,685 national accounts statistics.11 I measure sector output by net state West Bengal 26,668 29,208 45,237 1,74,297 3,23,040 3,73,785

Andhra Pradesh 29,199 38,359 54,615 1,99,424 2,39,214 2,31,165

domestic product (NSDP) at constant 1993-94 prices. Sector em-

Karnataka 24,751 43,029 62,422 1,64,741 1,85,655 2,39,700

ployment levels are taken from the estimates based on NSS em-

Kerala 23,250 30,476 71,306 1,32,699 1.41.270 1,30,422 ployment surveys. I exclude the electricity and mining sectors as Average of middle four 25,967 35,268 58,395 1,67,790 2,22,295 2,43,768

Tamil Nadu 29,739 40,906 64,429 1,36,076 1,97,724 1,34,751

they have different structural features in many ways like regula-

Gujarat 35,733 54,128 81,614 3,49,738 3,17,093 3,28,620

tion, natural resource base, etc. The category other services that

Haryana 40,877 74,634 1,04,411 2,40,810 2,52,155 1,75,342 comprises education, public administration and social services is Maharashtra 47,964 60,571 1,13,641 2,81,750 3,24,470 2,29,480

Punjab 23,191 34,232 62,466 2,43,812 2,79,223 2,26,416

also excluded. Growth rates of labour producti vity are not pre-

Average of top five 35,501 52,894 85,312 2,50,437 2,74,133 2,18,922

sented, as the proposition that I focus on is whether initial low

Total of 14 states 31,151 40,013 64,564 2,07,142 2,45,729 2,13,842 income states also have depressed or relatively lower productivity Source: NAS for NSDP and NSS for employment.

Economic & Political Weekly December 8, 2007 53

high productivity in Madhya Pradesh and Orissa. They are the bright spots in an otherwise dismal productivity levels in these states across sectors.

I move on to a comparison of construction and trade (includes hotels and repair services) two well known unskilled

Table 12: Labour Productivity Level Differences – Construction vs Retail Trade and Hotels (Rs per worker)

Construction Trade, Repair Services and Hotels
State 1993-94 1999-00 2004-05 1993-94 1999-00 2004-05
Bihar 29,083 25,283 18,015 23,533 24,746 21,140
Orissa 33,917 23,512 11,616 20,319 21,233 20,877
Uttar Pradesh 27,898 27,181 21,884 28,579 23,868 21,411
Rajasthan 18,017 25,525 25,293 38,561 42,807 42,842
Madhya Pradesh 47,730 61,113 42,180 38,574 29,685 24,713
Average of bottom five 31,329 32,523 23,798 29,913 28,468 26,197
West Bengal 26,421 34,350 28,748 21,372 26,443 35,779
Andhra Pradesh 26,199 29,445 35,609 27,512 33,240 32,610
Karnataka 33,293 49,135 47,509 28,033 32,183 42,012
Kerala 27,025 18,286 22,752 35,518 31,800 48,726
Average of middle four 28,235 32,804 33,655 28,109 30,917 39,782
Tamil Nadu 26,188 36,231 34,341 28,940 35,050 46,460
Gujarat 31,588 44,348 51,285 36,486 35,583 48,734
Haryana 43,253 43,290 31,534 43,923 43,152 71,492
Maharashtra 41,286 38,371 37,899 39,052 41,424 53,536
Punjab 38,818 33,930 36,759 41,395 37,815 39,103
Average of top five 35,579 40,560 38,765 37,100 38,802 55,056
Total of 14 states 29,983 33,062 30,290 30,811 31,874 36,917

Source: NAS and NSS.

labour-intensive sectors (Table 12). Surprisingly, labour productivity is declining in both the sectors in the bottom five states. In the trade and hotels sector, productivity levels have gone up in the middle four and the top five states. Low income states have a lot of catching up to do even in these sectors.

Finally, let us note that labour productivity in aggregate manufacturing per se is not meaningful as it has a large informal component. Therefore, I compare two segments within manufacturing, that is, registered and unregistered. I estimate the ratio of unregistered sector to registered sector labour productivity (Table 13). Within the manufacturing sector, differences emerge rather sharply. The striking fact is the large and widening gap in productivity between the registered and unregistered sectors in the bottom five states. Relative productivity of the unregistered sector is tending towards abysmally low. This suggests increasing divergence of productivity between states. The registered sector is galloping with high labour productivity growth across states. Informalisation of the labour force is driving down productivity in the unregistered sector perhaps more intensively in the low income states. Tamil Nadu is perhaps the only state that has maintained the relative productivity of the unregistered sector over the years. The reasons for this would be worth exploring.

3.1 Education, Skill Supply and Labour Productivity

Bosworth, Collins and Virmani (2007) in their detailed study of sources of growth in India, covering the period 1960-2004, call attention to the low levels of educational attainment of the Indian population and workforce. They point out that India has recently attained an average level of schooling comparable to that achieved in other Asian countries a quarter century earlier [Bosworth, Collins and Virmani 2007, Table 7]. In term of the educational attainment of the workforce, their estimates indicate that nearly 40 per cent of the workforce is found to be illiterate

54 and those who have completed secondary schooling account for 14 per cent of workers, while an additional 6 per cent are estimated to have a university degree (ibid).

The recent NSS survey on education and training (NSS report no 517) points out that in India, among the persons of age of 15 years and above, only 2 per cent had technical degrees or diplomas or certificates. I present in Table 14 (p 55) the interstate differences in educational attainment of persons (rural and urban areas) in India in 2004-05. Literates with general educational level secondary and above including diploma/certificate course have been considered to be educated (NSS report 517, pp 25). Following this definition, the numbers in Table 14 indicate large interstate variation in educational attainment. The bottom five states suffer from a serious shortage of educated persons. In these states, only 17 per cent are found to be educated and more seriously only 12 per cent are found to have secondary education or higher secondary education as against the all India average of 24 per cent and 16 per cent respectively. Expectedly, Karnataka, Tamil Nadu, Gujarat and Maharashtra emerge as educated states. The educational performance of Kerala is well known. Among middle income states, Andhra Pradesh and West Bengal have below average education. In brief, high income states also have better potential supply of educated persons. Notice, in particular, the relative advantage in terms of secondary education attainment in better off states. This will prove to be a great source of comparative advantage for these states in the years to come.

It is fairly well argued that secondary education is crucial for economic growth [Lewin and Caillods 2001]. Modern industry, whether it is manufacturing or services sector like telecommunications, emphasises train-

Table 13: Labour Productivity Ratio – ing and skill acquisition Unregistered to Registered Manufacturingon the job. A workforce Registered

with secondary school at-

Bihar 0.07 0.01 0.01

tainment will turn out to

Orissa 0.09 0.03 0.01

State 1993-94 1999-00 2004-05

be the best bet for such Uttar Pradesh 0.13 0.09 0.07
a job market. What has been the relative position Rajasthan Madhya Pradesh Average of bottom five 0.26 0.25 0.16 0.10 0.14 0.07 0.11 0.07 0.05
of Indian states in this West Bengal 0.18 0.16 0.12
area of education? In Ta- Andhra Pradesh 0.21 0.28 0.14
ble 15 (p 55), data on gross KarnatakaKerala 0.15 0.23 0.20 0.16 0.13 0.13
enrolment ratios for up- Average of middle four 0.19 0.19 0.13
per primary and secondary schooling in India for Tamil Nadu GujaratHaryana 0.16 0.18 0.51 0.17 0.26 0.33 0.16 0.14 0.17
the 14 states for the year Maharashtra 0.21 0.21 0.13
2003-04 is shown. The Punjab 0.49 0.26 0.16
relatively better devel- Average of top five Total of 14 states 0.29 0.17 0.25 0.14 0.15 0.11

opment of education in Source: NAS for NSDP and NSS for employment and unemployment survey.

middle income and high income states emerges clearly. West Bengal lags behind and looks more like bottom five states in this respect.

The constraint of skilled labour (human capital) is likely to be the binding constraint for growth and employment in many states in India. This leads me to ask the following question: Do states with better initial gross enrolment ratios (GER) in secondary education have better labour productivity growth? I investigate this in a preliminary way. The results are shown in Figure

december 8, 2007 Economic & Political Weekly

SPECIAL ARTICLE
6 (p 55), where I plot the secondary education GER in 1990-91 units once the economy growth process is initiated has important
on the growth rates of non-agriculture labour productivity in 14 implications for future growth and well-being. The structure of
states for the period 1993-94 to 2004-05.12 The positive relation employment growth and variation across states in India is a key
ship observed is very encouraging. Labour productivity is the key outcome of this unfolding development process. I have examined
proximate determinant of output and employment. States with some aspects of this regional employment growth in India. My
better supply of secondary school educated workers are likely to analysis is confined to 14 selected major states in India account-
Table 14: Educational Differences by State – 2004-05 State Not Literate Middle Secondary Higher Diploma/ Graduate All ing for 93 per cent of the population. GSDP growth in the high and medium income states grew faster relative to the low income
Literate and Up to Secondary Certificate and Primary Above states over the period 1993-94 to 2004-05. As a consequence, the
Bihar 516 198 118 92 40 2 33 1000 coefficient of variation increased from 36.6 in 1993-94 to 128
Orissa 412 252 177 75 35 5 43 1000 in 1999-2000 to 2004-05 suggesting widening of regional dis-
Uttar Pradesh 478 178 144 82 61 4 52 1000 Rajasthan 524 192 125 62 47 5 44 1000 Madhya Pradesh 456 252 115 61 55 8 54 1000 parities in the reform period. At the national level, employment has picked up in the pe-
Average of bottom five 477 214 136 74 48 5 45 1000 riod 1999-2000 to 2004-05, with manufacturing, construc-
West Bengal 325 321 162 82 49 3 58 1000 Andhra Pradesh 491 193 109 99 45 13 48 1000 tion and services creating substantial number of jobs. At the
Karnataka 382 200 171 124 58 11 55 1000 sub-national level employment, growth is unevenly distributed
Kerala 94 268 305 154 63 58 58 1000 across states.
Average of middle four 323 246 187 115 54 21 55 1000 Tamil Nadu 293 278 160 117 68 18 66 1000 The disquieting feature is the urban bias in the relative growth
Gujarat 318 229 189 124 63 16 61 1000 rates of employment. Urban employment has grown faster in
Haryana 351 218 117 150 78 15 71 1000 states with higher initial levels of urbanisation. Across the 14
Maharashtra 271 199 226 132 69 29 73 1000 Punjab 315 213 126 175 87 19 65 1000 Average of top five 310 227 164 140 73 19 67 1000 states, employment has grown faster in urban areas in both the sub-periods of the post-liberalisation period (1993-94 to 1999-
All India 382 228 160 102 58 12 57 1000 2000 and 1999-2000 to 2004-05). The benefits of growth in
Source: NSS 61st round report no 517, Table 3.8.1 (pp 66) on per 1,000 distribution of persons of 15 years and above by general educational level. terms of employment have gone largely to urbanised states in the
years since liberalisation. This is the dark side of the employment
Figure 6: Initial Education and Non-Agriculture Labour Productivity growth story in recent years of structural reform.
There has been increasing diversification across sectors on
Indian states, though the rate of diversification varies across
states. Employment growth is faster in states that have had
initially more diversified economies. It is important to note that
low income states have more concentrated structures to be
gin with and it is changing at a much slower pace. Interstate
disparities in the level and changes in diversification are obvi
ously the cause of interstate income disparities. Slow diversifica
tion in some of the major states such as Bihar, Uttar Pradesh,
Source: Rani (2007) and author’s estimates based on NAS and NSS survey results . Labour Productivity Growth Rate (%) Madhya Pradesh and Rajasthan is certainly a matter of concern for state policy. The slow rate of diversification of economic
get more investment and jobs coming in their way. It is now well activity is a key factor of the slow growth of employment in low
established that a major chunk of investment, domestic and for income states.
eign direct investment (FDI), has gone into five selected states, Spatial measures of concentration indicate varying changes
namely, Maharashtra, Gujarat, Karnataka, Andhra Pradesh and across sectors. In the subgroup finance and business services,
Tamil Nadu [Bagchi and Kurian 2005]. Incidentally, these are all spatial concentration rises Table 15: Gross Enrolment Ratios – 2003-04
relatively well endowed states with an educated workforce. This finding is troubling, because it also reflects the reality of higher sharply in 1993-94 and shows some marginal decline in be-State Upper Primary Secondary (11-14 years) (14-18 years) Bihar 25.3 16.9
unemployment of secondary educated workers. The higher out tween but increases again in Orissa 54.0 32.7
put growth rate has not sufficiently absorbed the additions to the educated labour force. One might conjecture that this suggests the last year 2004-05. This supports the proposition that Uttar Pradesh 48.6 37.9 Rajasthan 61.5 32.6 Madhya Pradesh 63.3 34.9
a serious mismatch of demand and supply in the labour market skilled labour-intensive ac-West Bengal 64.3 32.6
for trained workers. This interesting area needs to be explored further more deeply than possible in this paper. tivities are getting geographically concentrated over time, Andhra Pradesh 64.9 44.6 Karnataka 76.2 41.7 Kerala 93.6 48.0
which may explain the higher Tamil Nadu 100.4 56.9
4 Concluding Remarks Development is bound to be inegalitarian, as Nobel laureate regional income disparities observed earlier in the paper. Gujarat 70.4 40.0 Haryana 65.5 45.5 Maharashtra 87.6 53.9
Arthur Lewis pointed out long ago, because it does not start Across states, the un-Punjab 60.1 39.0
in every part of the economy at the same time. However, the diffusion of economic and social development across sub-national organised/informal sector has absorbed the additions to the All India 62.4 38.9 Rani (2007, Table A2) based on Selected Educational Statistics 2003-04.
Economic & Political Weekly December 8, 2007 55
AP BH GU HA KA KE MP MH OR PU RA TN UP WB 0 1 2 3 4 5 6 7 8 45 40 35 30 25 20 15 10 5 0Initial Secondary Education levels 1990-91

workforce in the period 1993-94 to 2004-05. More importantly, in secondary and higher secondary education. Non-agricultural the private sector within the organised sector has created sub-labour productivity has grown faster in states with initially highstantial absolute number of employment in three states, namely, er educational attainment. Andhra Pradesh, Karnataka and Gujarat. At the same time, the The structure of interstate disparities is defined by employlow income states show a net contraction of employment in the ment outcomes. The employment inequalities observed in this private sector. This provides us with a clue that the low income paper need to be further investigated along two lines: first, by exstates are likely to have proportionately more low productivity amining the interstate differences in the quality of employment jobs created in recent years. in terms of self-employment, regular employment and casual

Manufacturing labour productivity is higher in the high income labour. Second, the relationship between employment, labour states. However, the unregistered manufacturing productivity productivity and wages across states overtime. States with better falling is relative to registered manufacturing over time. This is supply of secondary school educated workers are likely to attract consistent with the creation of informal low productivity jobs in more investment and jobs coming in their way. The creation of a recent years. labour force, employable and amenable to skill training and up-

Educational attainment differs widely across states, with the grading, is an uphill task. States will have to find ways of meeting

low income states having much lower levels of young individuals this challenge.

Notes

1 For a brief review of studies of regional income disparities see Ramaswamy (2007).

2 I am grateful P P Sahu for providing me the mid-year estimates for the first three time points, namely, 1983, 1993-94 and 1999-2000. They are based on intercensal interpolations based on the 1981, 1991 and the 2001 population censuses. The estimates for 2004-05 are based on population projections for India and states, 2001-2026 prepared by the technical group on population projections constituted by the National Commission on Population, May 2006. This report is available on the internet: www.censusindia.net. The employment estimates were separately calculated for male and female in workers in rural and urban areas in each state and then summed to arrive at employment numbers.

3 The estimates not reported here to save pace, see Ramaswamy (2007)

4 A word about the employment concept used in the NSS surveys would be useful before we move on to the state level analysis of income and employment. The employment data in India is based the quinquennial surveys carried out by the National Sample Survey Organisation (NSSO) of the Ministry of Statistics and Programme Implementation (MOSPI). The estimate of employed (worker) according to the usual principal status and subsidiary status includes the person who (a) either worked for a relatively longer part of the 365 days preceding the date of survey, and (b) also those persons from among the remaining population who had worked at least 30 days during the reference period of 365 days preceding the date of survey.

5 The GSDP estimates are taken from the national accounts statistics data available at www.mospi. nic.in. (accessed on February 15, 2007). Mid-year population figures for each state is taken from the CMIE data document ‘National Accounts Statistics, October 2006’. For some states, the mid-year populations for 2004-05 is estimated using the reported on per capita figures and verified using data in economic surveys of respective state governments. The growth rates of GSDP presented are compound growth rates based on two end points. This is done to maintain consistency with employment growth rates based on the NSS quinquennial EUS data in later sections.

6 The rank of states does not under go dramatic changes in India as shown by Shetty (2003). Some positions change only within groups.

7 There are others like the Gini coefficient for the inequality of sector shares used by Imbs and Wacziarg (2003). They also use HH index as an alternative index. There is actually no strong reason to prefer one or the other measure of dispersion. The HH index is the simplest and easiest to compute.

8 This is equal to 1/n, where n is the number of sectors and all the sectors having an equal share.

9 This has come into prominence in the new economic geography literature. The index of locational Gini is estimated to address this issue. See Amiti (1998) for a good discussion.

10 Here, I take the total to be the aggregate employment of the selected 14 states in order to maintain the focus on the 14 states.

11 These are accessible at http\\www.mospi.nic.in, retrieved on April 10, 2007.

12 I exclude form the non-agricultural sector mining and electricity as they are dominated by natural resource distribution and government ownership.

References

Ahluwalia, M S (2001): ‘State Level Performance under Economic Reforms in India’, Working Paper No 96, Centre for Research on Economic Development and Policy Reform, Stanford University.

Ahsan, Ahmad and Pages Carmen (2006): ‘Some Implications of Regional Differences in Labour Market Outcomes in India’ in Institute of Human Development, India, ‘Meeting the Employment Challenge’, conference on labour and employment issues, July 27-29, New Delhi.

Amiti, Mary (1998): ‘New Trade Theories and Industrial Location in the EU: A Survey of Evidence’, Oxford Review of Economic Policy, Vol 14, No 2, pp 45-53.

Bagchi, Amaresh and John Kurian (2005): ‘Regional Inequalities in India: Pre- and Post-reform Trends and Challenges for Policy’ in Jos Mooji (ed), The Politics of Economic Reforms in India, Sage Publications, New Delhi.

Besley, Timothy and Robin Burgess (2004): ‘Can Labour Regulation Hinder Economic Performance? Evidence from India’, The Quarterly Journal of Economics, Vol 119, No 1, pp 91-134.

Bharadwaj, Krishna (1982): ‘Regional Differentiation in India: A Note’, Economic and Political Weekly, Annual Number, April, pp 605-14.

Bhattacharya, B B and S Sakthivel (2004a): ‘Regional Growth and Disparity in India: Comparison of Preand Post-reform Decades’, Economic and Political Weekly, Vol 39, No 10, pp 1071-77.

– (2004b): ‘Economic Reforms and Jobless Growth in India’, Working Paper Series, No E/245, Institute of Economic Growth, New Delhi.

Bosworth B, Susan Collins and Arvind Viramani (2007): ‘Sources of Growth in the Indian Economy’, NBER Working Paper No 12091.

Chadha, G K and P P Sahu (2002): ‘Post-Reform Setbacks on Rural Employment: Issues that Need Further Scrutiny’, Economic and Political Weekly, Vol 37, No 21, pp 1998-2026.

Dreze, J and Amartya Sen (eds) (1996): Indian Development: Selected Regional Perspectives, Clarendon Press, Oxford.

Gordon, Jim and Poonam Gupta (2004): ‘Understanding India’s Services Revolution’, Working Paper No WP/ 04/171, International Monetary Fund, Washington DC.

Hasan, Rana, Devashsish Mitra and K V Ramaswamy (2007): ‘Trade Reforms, Labor Regulations and Labour Demand Elasticities’, Review of Economics and Statistics, Vol 89, No 3, pp 466-81.

Imbs, Jean and Romain Wacziarg (2003): ‘Stages of Diversification’, American Economic Review, Vol 93, No 1, pp 63-85.

Kochhar, Kalpana, Utsav Kumar, Raghuram Rajan, Arvind Subramanian and Ioannis Tokatlidis (2006): ‘India’s Pattern of Development: What Happened, What Follows’, Journal of Monetary Economics, Vol 53, pp 9811019.

Lewin, K and F Caillods (2001): Financing Secondary Education in Developing Countries: Strategies for Sustainable Growth, UNESCO, International Institute for Educational Planning, France.

Office of Registrar General and Census Commissioner, India (2006): ‘Census of India 2001, Population Projections for India and States 2001-2026 (Revised December 2006)’, report of the technical group on population projections, New Delhi, available at www. censusindia.net.

National Sample Survey Organisation (1997): Employment and Unemployment Situation in India, 1993-1994, NSS 50th Round, Report No 409, Government of India, New Delhi.

  • (2001): Employment and Unemployment Situation in India, 1999-2000, NSS 55th Round, Report No 458, Government of India, New Delhi.
  • (2006a): Employment and Unemployment Situation in India, January-June 2004, NSS 61th Round, Report No 515, Parts I and II, Government of India, New Delhi.
  • (2006b): Status of Education and Vocational Training in India, 2004-2005, Report No 517, Government of India, New Delhi.
  • Ramaswamy, K V (2007): ‘Growth and Employment in India: The Regional Dimension’, Working Paper No 22, Institute of South Asian Studies, National University of Singapore, Singapore.

    Rani, Geetha (2007): ‘Secondary Education in India: Determinants of Development and Performance’, Working Paper, NIEPA, New Delhi, available at http:\ www.esocialsciences.com

    Sachs, Jeffrey Nirupam, Bajpai and Ananthi Ramaiah (2002): ‘Understanding Regional Economic Growth in India’, Working Paper No 88, Centre for International Development, Harvard University.

    Shetty, S L (2003): ‘Growth of SDP and Structural Changes in State Economies: Interstate Comparisons’, Economic and Political Weekly, Vol 39, No 49, pp 5189-5200.

    Srivastava, Ravi (1994): ‘Planning and Regional Disparities in India’ in Terence Byres (ed), The State and Develop ment Planning in India, Oxford University Press, New Delhi.

    Wacziarg, Romain and Jessica Wallack (2004): ‘Trade Liberalisation and Intersectoral Labour Movements’, Journal of International Economics, Vol 64, No 2, pp 411-39.

    december 8, 2007 Economic & Political Weekly

    Dear reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here

    Comments

    (-) Hide

    EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

    Back to Top