Commentary
Are the Results of the Economic Census Robust?
While the buoyancy in the growth of employment in the economy since the year 2000 cannot be dismissed, this cannot be surmised from the results of the Economic Census 2005.
JEEMOL UNNI, G RAVEENDRAN
T
The Central Statistical Organisation (CSO) launched a plan scheme ‘Economic Census and Surveys’ in the year 1976 and it envisaged the conduct of countrywide census of all economic activities (excluding those engaged in crop production and plantation), which would be followed by detailed sample surveys of the unorganised segments of different sectors of the non-agricultural economy in a phased manner during the intervening period of two successive economic censuses. The basic purpose of conducting the census was to prepare a frame of non-farm enterprises to facilitate the conduct of followup surveys, intended to collect more detailed sector specific information. It is also expected to prepare a directory of establishments employing more than five workers and identify areas of industrial concentration.
The Economic Census of 2005 recorded a 2.4 per cent growth of employment in the non-agricultural sectors, including allied agriculture, compared to the previous Economic Census of 1998 (Table 1). This was much higher than the employment growth of 1.7 per cent in the previous period 1990 and 1998, recorded from the previous economic censuses. This spurt in the growth in nonagricultural employment is actually corroborated by the results of the 60th Round of the NSS. In fact the NSS of 2004 recorded a much higher growth of the non-agricultural sectors, 4.4 per cent over the previous employment-unemployment survey of 1999-2000. Again this was higher than the growth in employment during 1993-94 to 1999-2000, 3.2 per cent. Thus while the magnitude of growth recorded by the economic census is lower, the direction of change is similar. However, the economic census by itself is not an instrument for estimating the size of the workforce or for analysing the employment trends, as it is not designed for the purpose, as we shall discuss below.
NSS Employment Growth
The increase in employment growth over the two periods is pronounced when we consider the economy as a whole as well. The total workforce grew at 2.6 per cent during the recent period of 2000 to 2004 according to the 60th Round of the NSS, in contrast to the dismal 1 per cent growth in 1993-94 to1999-2000 (Table 2). The 60th Round has the limitation of being of six months duration as compared to the usual quinquennial rounds. However, the sample size was much larger than the thin rounds and was capable of providing reliable national level estimates, except for the seasonality factor.
Much of the deceleration in employment growth in the late 1990s is accounted for by the poor performance of the agricultural sector, which showed a negative growth in employment (Table 2) and jobless growth in the organised manufacturing sector (as we argue later). The agricultural sector picked up and had a positive 1.3 per cent annual growth in the period 2000-04. Another positive feature is also that the proportion of the workforce engaged in the agricultural sector declined from nearly 65 per cent in 1993-94 to less than 60 per cent in 1999-2000 and, in spite of the growth in the sector in 2004, remained at 58.4 per cent of the workforce.
The growth in the non-agricultural sectors was close to 3 per cent in industry
“Jobless growth is empty rhetoric, at least off the farm where the bulk of growth takes place. Off-farm jobs are growing faster than the workforce, finds the Fifth Economic Census, whose results were released by the government on Monday”
– Economic Times, June 13, 2006.
“The Fifth Economic Census Report, 2005, which is the complete enumeration of all economic activities in establishments and households, except crop production and plantation, allays, at least partially, fears of jobless growth in the economy. The evidence clearly shows that entrepreneurial skills unleashed in recent years have spurred massive growth, and dented unemployment rates”
– The Financial Express, June 13, 2006.
“To a certain extent, this (economic census) should put to rest the view that the country has been experiencing jobless growth. For, this rapid growth in employment, insofar as it is significantly faster than the rate of population growth, and in a segment that accounts for a fourth of the total workforce, points a promising picture”
– Business Standard, June 13, 2006.
Economic and Political Weekly August 19 2006
and services in the late 1990s, and has picked up to more than 4 per cent in 2004 (Table 2). The industry sector included the booming construction sector, and therefore, while manufacturing sector has also shown a growth in employment, it is the services sector that accounts for much of the buoyancy in employment.
It has been suggested that the decline in employment growth in 1999-2000 has been mainly due to the withdrawal of rural women in subsidiary activities from the labour force. This has been attributed to an income effect, that is, with rising household incomes, women prefer to withdraw from the workforce [Sundaram and Tendulkar 2006]. The postulation is not substantiated by the findings available from the surveys. First, the decline in workforce participation in 1999-2000 exists for men and women in rural areas, and for men, particularly in the principal status (Table 3). This can be attributed to the decline in the agricultural sector in that year. Among women, the phenomenon of decline in participation, while extended to the urban sector in 1999-2000, was found to have risen in 2004 in subsidiary status in rural and urban areas and in principal status as well in urban areas.
Further, if one were to observe the growth of the labour force in the two periods (Table 2), we find that it follows close to
Table 1: Annual Growth in the Non-agricultural Sector in NSS Roundsand Economic Census
NSS Rounds1 Economic Census2 Year Annual Year Annual Growth Growth
1983 to
1993-94 3.1 1980 to 1990 2.8 1993-94 to
1999-2000 3.2 1990 to 1998 1.7 1999-2000 to
2004 (January to June) 4.4 1998 to 2005 2.4
Notes: (1) Includes Assam and Jammu and Kashmir.
the growth rate of the workforce. This implies that the withdrawal of women from the workforce merely follows the business cycle. If there are opportunities available in the market, the women are available for work, and in the event of a depression in the non-agricultural sectors or poor agricultural performance, women withdraw from the workforce. This phenomenon of withdrawal of women from the workforce is therefore more a discouraged worker effect rather than an income effect. To emphasise, the spurt in the growth of employment in 2004 includes the increased participation of women who seize the opportunity available for work, in a subsidiary capacity in the rural and in principal activity as well in urban areas. The nature of work available to them and its quality are of course other issues which are not discussed in this note.
Under-Enumeration in the Economic Census
The reason for our scepticism about the employment growth estimates of the economic census is the gross underenumeration of employment. This is clearly depicted when we compare the estimate of absolute employment in the nonagricultural sectors thrown up by the NSSO
Table 3: Workforce Participation Rate byUsual Status by Location and Gender inthe NSS Rounds
Year Rural Female Urban Female P P + S P P + S
1983 24.8 34.0 12.0 15.1 1993-94 23.4 32.8 12.1 15.5 1999-2000 23.1 29.9 11.7 13.9 2004 22.8 31.5 12.1 15.0
Rural Male Urban Male 1983 52.8 54.7 50.0 51.2 1993-94 53.8 55.3 51.3 52.1 1999-2000 52.2 53.1 51.3 51.8 2004 52.7 54.2 53.1 54.0
Notes: P: Principal Status; P + S: Principal + Subsidiary Status.
Source: NSSO, Employment and Unemployment Situation in India, 60th Round, January to June 2004.
Table 2: Labour Force by Sectors and Annual Growth in the NSSO Rounds
Employment (Million) Annual Growth Rate (Per Cent) 1993-94 1999-2000 2004 1993-94 to 1999-2000 to (January-June) 1999-2000 2004
Agriculture 242.5 237.6 259.1 -0.3 1.3 Industry 58.2 69.2 80.3 2.9 4.2 Services 73.8 90.3 104.4 3.4 4.6 Workforce 374.5 397.0 443.8 1.0 2.6 Unemployed 7.5 9.1 10.5 3.2 3.4 Labour force 381.9 406.1 453.1 1.0 2.6
Note: Computed after adjusting for the population in corresponding population censuses.
employment-unemployment rounds and the economic census of various periods. Being a census conducted by ad hoc enumerators without any legal backing, it tends to miss out smaller own account enterprises and self-employed persons. As per the estimates available from the labour force survey of NSSO and the projected population estimates, the total workforce in the country as on January 1, 2000 was 397 million (Table 2). Of these, about 159 million workers were estimated to be in the non-agricultural sector (Table 4). However, the Fourth Economic Census, 1998 could list only 30.3 million
Table 4: Employment and in theNon-agricultural Sector in NSS Roundsand Economic Census
NSS Rounds1 Economic Census2
Year Employment Year Employment (Million) (Million)
1983 96.2 1980 53.2 (55.3)3 1993-94 132.0 1990 70.4 (53.3) 1999-2000 159.5 1998 80.6 (50.5) 2004 (January
to June) 191.9 2005 95.5 (49.7)
Notes: (1) Includes Assam and Jammu and Kashmir.
Table 5: Annual Growth in Non-agricultural Sector by States, NSS Rounds and Economic Census
(Per cent)
State NSS 1999-2000 Economic to 2004 Census 1998 to 2005
Andhra Pradesh 3.04 2.40 Arunachal Pradesh 8.57 4.17 Assam -1.45 3.19 Bihar 3.5 0.07 Goa 15.45 1.87 Gujarat 6.12 1.39 Haryana 4.24 5.12 Himachal Pradesh 5.96 2.09 Jammu and Kashmir 9.65 6.82 Karnataka 4.14 1.86 Kerala 5.69 5.39 Madhya Pradesh 4.91 1.49 Maharashtra 5.66 1.79 Manipur 12.63 2.25 Meghalaya 11.26 4.12 Mizoram 7.12 3.91 Nagaland 2.77 0.75 Orissa 5.17 2.08 Punjab 6.08 3.59 Rajasthan 6.16 2.60 Sikkim 1.92 5.52 Tamil Nadu 3.61 4.62 Tripura 7.98 5.07 Uttar Pradesh 3.61 3.14 West Bengal 4.72 0.87 All-India1 4.46 2.49
Notes: (1) All-India includes all states and union territories.
(2) Economic censuses estimates include employment in allied agricultural enterprises.
Economic and Political Weekly August 19, 2006 enterprises employing about 80.6 million workers. Of these, about 3.5 million enterprises employing 6.7 million workers were in agricultural enterprises. Thus the nonagricultural enterprises were 26.9 million with a total employment of about 76.5 million as per the Economic Census, 1998 as against about 159 million workers estimated from labour force survey 1999-2000.
Further, an informal sector enterprise survey conducted by the NSSO in the year 1999-2000 estimated the number of proprietary and partnership non-agricultural enterprises, excluding mining and quarrying and electricity, gas and water supply and manufacturing units covered under the Factories Act, as 44.4 million. This was about 65.3 per cent above the total nonagricultural enterprises of 26.9 million enlisted in the Economic Census, 1998. Obviously both non-agricultural enterprises and employment were grossly underenumerated in the Economic Census of 1998.
The Economic Census of 2005 estimated the non-agricultural enterprises, excluding Assam and Jammu and Kashmir, at
40.9 million generating employment for
95.5 million workers. In stark contrast the NSS Employment-Unemployment Survey, 2004 estimated employment in the nonagricultural sector (adjusted for the population and including Assam and Jammu and Kashmir) at nearly 192 million.
The economic census manages to capture only about 50 per cent of the employment in the non-agricultural sector in each successive attempt in comparison to the corresponding NSS employment-unemployment surveys (Table 3). To make matters worse, the percentage of coverage of employment in the non-agricultural sector in the economic census, compared to successive NSS rounds, has been reducing from 55 per cent in 1980 to 53 per cent in 1990, to 50 per cent in 1998 and 49.7 per cent in 2005.
The economic census made a major change in its enumeration procedure in 1998 when it de-linked the operation from the population census. The Economic Censuses of 1980 and 1990 were conducted along with the corresponding population censuses. This was therefore conducted with a statutory backing and relatively better qualified officials as investigators. In 1998 and 2005, the de-linking from the population census left the economic census without any statutory backing and it became almost the sole responsibility of the Directorate of Economics and Statistics of the respective states to manage the conduct of the economic census including the recruitment, training and supervision of ad hoc enumerators. This perhaps reduced the reliability of the economic census results even further. However, the extent of under-enumeration between the censuses was enormous even when it was linked to the population census, pointing to other reasons as well for this undercount.
There is, however, one redeeming feature of the economic census when the results are presented by all major states. In comparison to the respective NSS rounds, while the absolute undercount is significant in the economic census, the growth of employment during the period is similar across all major states. An intriguing feature is that the small states and the state of Jammu and Kashmir show exceptionally high growth of employment. Could this be a case of a small base or are there some genuine growth inducing factors operating?
Another interesting feature thrown up by the economic census is a decline in the employment per enterprise at the all India level and in all states. This result is also corroborated by successive Annual Surveys of Industry (ASI) for the organised sector and the Third All India Census of Small-Scale Industries, 2001-02 compared to second SSI census.
One explanation for this phenomenon is perhaps that the larger units, particularly in the organised and the SSI sectors, have advanced technologically and reduced the employment per unit, as seen in the ASI and SSI census. This would amount to what we referred to earlier as jobless growth in the organised manufacturing sector. A second possibility is that the economic census is not able to capture the total employment in the larger enterprises. These units may have started to hire a larger proportion of their workers on contract or casual labour and do not report them to the enumerators. This second phenomenon would provide part of the explanation for the under-enumeration in the economic census. The ASI in fact reports an increase in the proportion of contract workers to total workers from 13 to 20 per cent in 1995-96 to 2001-02 [Unni 2006].
In sum, while the buoyancy in employment growth in the economy since the year 2000 cannot be dismissed, this cannot be surmised from the economic census data per se. The second feature of the employment growth that is worrisome is the shedding of labour by the larger units, or rise in the smaller sized units, since this could be an indicator of work with less stable contracts and fewer benefits forthcoming from the enterprise.
EPWEmail: jeemolunni@yahoo.co.in
[The views presented here are those of the authors alone.]
References
Sundaram, K and Suresh D Tendulkar (2006): ‘Trends in Labour and Employment in India, 1983-2003: Some Fresh Results’, paper presented at the Conference on India: Meeting the Employment Challenge, organised by Institute for Human Development and WorldBank, July 27-29, New Delhi.
Unni, Jeemol (2006): ‘Employment Trends and Earnings in the Informal Sector’, paper presented at the Conference on India: Meeting the Employment Challenge, organised by Institute for Human Development and WorldBank, July 27-29, New Delhi.
REQUEST TO AUTHORS/CONTRIBUTORS
EPW seeks the help of contributors for quicker processing of articles submitted for publication.
Soft/Hard Copies
Contributions of less than 4,000 words for the Commentary/Discussion sections can be sent by email to. There is no need to send hard copies of short articles. Articles longer than 4,000 words for Perspectives/Special Article sections, however, must be sent in hard and soft copy formats. Soft copies can be sent by email or on a CD/floppy. Please note that EPW will not be able to process articles longer than 4,000 words unless both soft and hard copies have been received.
Revision of Articles
Authors are requested to send us articles that are complete and up to date in all respects. They are requested not to send us revised versions – by email or post – unless there are major developments that require changes to the article first submitted. Since revisions pose difficulties in managing submissions and processing the correct version, EPW requests authors to send in articles that have been finalised and are complete in all respects. Revised versions will not be processed other than in exceptional circumstances.
(Please also see Notes to Contributors for additional details)
Economic and Political Weekly August 19 2006
