Trade unions' influence on wages during 2004/05 --2011/12: evidence from NSS rounds.

AuthorMondal, Bidisha
PositionReport - Abstract


This paper has two questions to address. The first: how the effect of union-membership on wages has moved over the period 2004/05-2011/12 in the unionised enterprises and whether the trend differs among the occupational categories and with the type of enterprise? The second: how the presence of a union in an enterprise has influenced wages during the period and whether the trend differs among occupational categories and with the type of enterprises?

A few of the case studies have confirmed positive effect of union-membership on wages (Madheswaran & Shanmugam, 2003; Bhandari, 2010). In the background of the findings of a few other studies which found recent trends of declining strength of trade unions due to privatization of the economy, anti-labor policies of the government and declining membership-base of trade unions, it was interesting to look into the influence of union-membership on wages. The database of this paper has been the manufacturing sector workers in India during employment and unemployment surveys in the 61st and 68th NSS rounds. The methodologies have been the descriptive statistics, decomposition analysis and multivariate linear regression analysis. The findings confirm decline in effect of union-membership on wages in unionized enterprises, the decline being stronger for low-wage, low-skill workers and in public/private limited companies. But the presence of a union has become more important as wages of non-union members in unionized enterprises have increased relatively more than the workers in non-unionized enterprises during 2004/05-2011/12.


A study based on a sample survey of blue-collar male workers in manufacturing industries in Madras district of Tamil Nadu concluded that union members earned around 19% more than nonunion members and almost 42% of the wage gap with selectivity bias adjusted was due to discrimination against nonunion members (Madheswaran & Shanmugam, 2003). Another study based on field data collected from 271 workers in factories in selected industrial areas of West Bengal, Delhi and Haryana also concluded that the wage premium of union members compared to non-union members was almost 23.5% and around 40% of the wage gap was due to discrimination. Thus the study rejected the free-riding benefits of non-union members in getting the bargained benefits by the union-members (Bhandari, 2010)

A set of case-studies of trade unions working in companies in Mumbai and nearby found that privatization of public sector companies, outsourcing of production activities, anti-labor policies of the government and increasing flexibility in employment terms were the reasons for the declining base of trade unions (Srivastava, 2006). Another case study conducted in Tata Steel Company by interviewing the white-collar employees found out that management played the major role in recruitment decisions, job assignments and monetary decisions and trade unions had influence only on job security conditions. However the low-skill workers did not feel that trade unions were keen on protecting their jobs (Sarkar & Varkkey, 2008).

Mamkoottam (2006) found that trade unionism was more prevalent among the blue-collar workers before 1970s after which it became stronger among white-collar employees. Another analysis found that during the era of liberalization, privatization and globalization model of growth, the decentralized unions replaced the centralized trade unions affiliated to political parties mainly in newer capital-intensive industries. But the decentralized trade unions could not withstand the resistance of multinational companies in most of the cases (Bhattacharjee, 2000). Although there is evidence of positive impact of union-membership on wages, the case studies indicate declining strength of the trade unions due to several factors. So whether there is decline in the influence of trade unions on wages for the all India sample is an interesting question to look into.

Database & Variables

The database used here is the employment and unemployment surveys of 61st and 68th rounds of National Sample Survey Organization. These are quinquennial surveys that collect data on various socio-economic aspects and the units are the households. The geographical coverage in collecting data has been the whole of Indian union except inaccessible interior places. Various scientific sampling techniques like stratified multi-stage random sampling, simple random sampling without replacement and probability proportional to size with replacement has been applied to select the final households for surveying. So multipliers have been an integral part of the analysis attempted here. The datasets are cross-sectional.

The analysis has been on manufacturing sector industries only. Industrial classification in the 61st round (2004-05) has been done on the basis of 1998-NIC code. At two digit level, the manufacturing industries were coded from 15 to 37 according to 1998-NIC code. In the 68th round (2011-12) the industrial classification has been done according to 2008-NIC codes and the manufacturing sector industries were coded from 10 to 33. Concordance has been done between NCO code-1968 and NCO code-2004 with some subjective changes.

The definition of formal sector as mentioned in the Factories Act, 1948 has been followed here. The enterprises with 10 or more workers with use of electricity for production purpose or 20 or more workers without use of electricity for production purpose comprises the formal sector. All other enterprises are considered to be in the informal sector.

The analysis here focused on regular workers in the formal sector. Workers who are considered to be regular according to Usual Principal Status (UPS) have been taken into account for the analysis. UPS shows the main employment status of a person during the reference period of last 365 days.

Dummies were created for general education levels. General education level 1 stands for illiterates and literates without formal schooling. General education level 2 stands for below primary, primary and middle school level of education. General education level 3 stands for the secondary, higher-secondary and diploma courses and level 4 stands for graduate, post-graduate and above levels of education. Age and age-squared were used as proxies for experience level. The wage earnings of the workers has been calculated by dividing total wage and salary earnings received or receivable during a week including cash and kind and then by dividing by total days in that activity during the week in order to arrive at the average daily wage earnings. The real wage earnings are calculated using 2001 base prices. The logarithm of wage level has been used as dependent variable in decomposition analysis and multivariate regression analysis. The enterprises are divided into three groups according to the ownership patterns of the enterprises: group 1 consists of household-managed enterprises, partnership firms, co-operative societies; group 2 is the public sector...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT