Application of job security laws, workers’ bargaining power and employment outcomes in India

Abstract

This article empirically investigates two propositions normally advanced to explain employment stagnation in the organised sector of the Indian labour market namely, rigid labour laws regulating the ‘hiring and firing’ of workers (along with factory closures) and militant trade union activity boosting workers’ bargaining power. It is claimed that labour laws arrest employment creation by making the labour adjustment process difficult. Additionally, it is claimed that workers’ increased bargaining power raises the effective cost of labour, hindering employment creation. This article critically examines both these claims. First, it questions whether labour laws can be held responsible as an explanation for employment stagnation in the organised sector as a whole, since a close reading of the Indian legislation reveals that labour laws apply in less than 35% of aggregate employment in the organised sector in India. Next, it investigates whether employment stagnation in those industry segments where labour laws apply – namely, organised manufacturing – is due to restrictions on ‘hiring and firing’. We find no evidence to support this hypothesis. Finally, it analyses the trend in workers’ bargaining power through various indicators, and finds such power to be unambiguously declining. Thus, the study finds no empirical support for the two conventional arguments put forward to explain employment stagnation in the organised sector of the Indian labour market.

[I]t is argued that the legal provisions of job security and institutional factors like the pressure of trade unions make adjustment of the workforce of enterprises difficult, and discourage organized sector enterprises from expanding employment.

Sharma (2006: 2080)

Introduction

Though the author himself does not subscribe to it, the above quote succinctly captures a widely held view of the organised sector of the Indian economy: that, rigid labour laws and enhanced workers’ bargaining power deter employment creation. Commentators making this claim commonly draw attention particularly to Chapter VB of the Industrial Disputes Act (IDA, 1947), the statute that regulates the hiring and firing of workers and the closure of firms. Specifically, it stipulates that prior government permission is necessary for retrenching or laying off workers, and for closing down establishments. This, it is alleged, makes labour adjustment process practically impossible to undertake, which in turn hinders job creation in the first place. Rising trade union (TU) activity is also claimed to hamper labour hiring. Work stoppages due to industrial disputes, rising wage claims and demands for better working conditions inter alia are seen as intimately related to rising TU power – and this in turn raises the effective cost of labour. Consequently, the argument goes, Indian firms have responded by adopting capital intensive techniques rather than seeking more labour-focused growth. This article empirically investigates the validity of these two claims, concluding that the evidence goes against them.

In the next section, we briefly discuss the broad theoretical predictions and empirical evidence available internationally, on the impact of labour laws and collective bargaining on employment creation. Section ‘Evolution and application of labour laws’ sets the context of the case study with an overview of the evolution of the industrial relations system (IRS) and labour laws in India – with special focus on the current controversy around labour reforms. It then identifies the exact segments of the Indian workforce governed by job security regulation (JSR) and explains why employment growth in the organised sector is sluggish. Section ‘Evidence of employment adjustment in the presence of labour laws’ investigates the rigidity of the Indian labour market and finds no evidence of JSR arresting employment adjustment; on the contrary, indiscriminate uses of contract labour along with various means of circumventing JSR makes the labour market de facto flexible. Section ‘Assessing workers’ bargaining power in the organised sector with special reference to organised manufacturing’ assesses the comparative bargaining power of workers vis-à-vis employers and concludes that the bargaining strength of workers has unambiguously declined during the study period – which has implications for the class distribution of income and employment growth. The article ends by drawing the conclusion that there is no evidence that JSR hampered employment adjustment and production, while the strengthened bargaining position of employers has not prevented employment stagnation.

The impact of labour laws and collective bargaining on employment creation

Gordon Betcherman (2014: 12), conducting a comprehensive review of theoretical studies, concluded that ‘[e]conomic theory does not lead to clear predictions about the employment effects of (EPL) [employment protection law]’. In this section, I will outline the basic reasoning underpinning the post-Keynesian and new institutional arguments that have contested the neoclassical position that increased protection dampens employment creation, and draw attention to some empirical research that supports these claims.

The ‘standard economic analysis of labour law rules’ (Deakin, 2009: 3), which is to say a neoclassical approach, is based on ‘the idea of a self-equilibrating market which corrects itself in response to temporary dislocations’ (Deakin, 2009: 3). In this tradition, ‘wages and employment are set [entirely] by the interaction of supply and demand for labour’ and ‘labour law rules operate as an exogenous intervention in, or interference with, the operation of market forces’ (Deakin, 2009: 2). This interference with market forces leads to coordination failures, retarding economic growth and employment creation (Posner, 1984).

This reliance on the coordinating role of wages to attain full employment – in absence of worker-protective laws – has been questioned by at least two schools of economists – post-Keynesian and new institutional economics.

First, the concept of attaining full employment through wage adjustments – that is, in the absence of ‘hiring and firing costs’ imposed by labour laws (Lindbeck and Snower, 1988) – is reminiscent of pre-Keynesian thinking. It necessarily presumes that Say’s Law holds that savings are automatically channelled into investment (Dutt, 1986). This concept is obviously not accepted by post-Keynesians.

Second, according to the post-Keynesians, eliminating worker-protective laws – in order to institute labour market flexibility – is actually employment retarding. Patnaik (2011), for example, discusses the problem arising on the demand side from such a move:

[T]he introduction of labour market flexibility … necessarily entails a reduction in the share of wages in the net output of the economy. And since a rupee paid out as wages creates more demand than a rupee that accrues as profit (of which a larger proportion is saved), such a shift in income distribution against workers, quite apart from being regressive in itself, results in a constriction of the domestic market, with an adverse effect upon employment for this reason. Besides, … since goods demanded by workers typically tend to be produced by more employment-intensive methods, the generation of employment is constricted for this additional reason too in a regime of labour market flexibility.

New institutional economics, rather than looking at labour laws as an exogenous imposition on market operations, views them as a tool to correct market failures: Typically, labour laws are regarded ‘… as “endogenous” solutions to coordination problems … [matching] the expectations of actors [workers and employers] under conditions of uncertainty’ (Deakin, 2009: 3–4). In fact, ‘working times [regulation] and dismissal laws often refer back to standards based on workplace practice and self-regulation by industry, so that the distinction between external enforcement and reflection of pre-existing norms is blurred’ (Deakin, 2016: 4). Thus, endogenously evolving labour regulation is usually ‘complementary to private ordering, but … does not displace it’ (Deakin, 2016: 4). Therefore, labour laws, instead of hindering employment creation, actually promote it through resolving market failures. Moreover, labour laws encouraging collective bargaining can be envisaged as promoting cooperative behaviour between employer and workers, benefitting both parties (Simon, 1951).

Empirical evidence from cross-country studies also remains inconclusive; additionally, the quality of evidence is also a suspect. Betcherman (2012: 21) notes,

[n]ot only are the findings on employment impacts mixed, but the results also can be characterized as fragile. Findings are often sensitive to model specification and the treatment of data.

Furthermore, the implicit assumption in constructing the ‘strictness’ of employment protection index – used in the cross-country econometric analysis – is fundamentally problematic. Deakin (2016) notes that in cross-country regressions, the widely used Organisation for Economic Co-operation and Development (OECD) index on EPL and the World Bank’s Employing Workers Index are both open to question. He writes,

Part of the controversy around these indices arises from the premise underlying them, which is that worker-protective laws necessarily impose a burden on business. No account is taken of possible benefits to be derived from protective labour laws … [by] employers from enhanced labour-management cooperation or reduced levels of workplace conflict … [and unambiguous] identification of protection with negative economic effects. (Deakin, 2016: 14–15)

There is instead unambiguous ‘identification of protection with negative economic effects’ (Deakin, 2016: 14–15). Similar understanding permeates the construction of labour ‘rigidity’ indices used in individual country studies.

Interestingly, multiple empirical studies using different time-series datasets show that JSRs encourage employers to invest in productivity-enhancing technologies and inter alia boost employment in high technology sectors (Acharya et al., 2014). Evidence further suggests that labour laws uphold worker voice (within the enterprise) and collective bargaining (at firm and industry level), and that they encourage employment and productivity gains through their impact on worker motivation and devotion (Deakin et al., 2014). Empirical results show that protective labour laws supporting collective employee representation are associated with more egalitarian outcomes in wage-bargaining and higher share of labour income in national output (Deakin et al., 2014). These findings even forced The World Bank (2008: 20) to modify its view that ‘rigid labor regulation reduces jobs’ and to acknowledge that

Employment regulations are unquestionably necessary. They are needed to protect workers from arbitrary or unfair treatment and to ensure efficient contracting between employers and workers. They increase job stability and can improve productivity through employer-worker cooperation. They benefit both workers and firms. … [and their] impact can be negative where regulatory interventions are insufficient or excessive … (The World Bank, 2014: 231).

Thus, labour legislation can be efficiency enhancing, promoting employment and productivity growth. In the next three sections, we turn to a case study of India.

Evolution and application of labour laws

Since product and factor markets are interlinked, labour laws are an important instrument for the state to promote output and employment growth (Papola, 1994). Labour laws in India, it is alleged, have evolved in a manner that excessively promotes labour welfare at the cost of economic efficiency (Goldar, 2000). Is this reading correct?

Like other decolonised nations, India embarked on centralised economic planning in the 1950s in its pursuit of rapid economic development. Thus, policy makers preferred the state interventionist IRS model, consistent with a broader strategy of activist state regulation of the product market. With the state emerging as the chief mediator of capital-labour relations, its principal objective was to balance the interests of competing classes. This practice of promoting labour welfare alongside protecting the interests of capital ‘could be described as [a] corporatist social pact’ (Sundar, 2005: 922).

However, this IRS model was vulnerable to political manoeuvering and the state often intervened to further its own interest, ultimately weakening union power. This resulted to a situation in which,

there is no provision in [labour laws] either the Industrial Disputes Act or the Trade Union Act for formal recognition by the employer of a union as a collective bargaining agent on behalf of workers … [and this] has undermined the development of a culture of collective bargaining and bipartisan negotiation. (Hill, 2009: 402)

In a further blow to the workers’ movement, ‘conventions 87 and 98 of the ILO assuring freedom of association and collective bargaining to workers were not ratified’ (Sundar, 2005: 922).

Nevertheless, the state maintained its corporatist social stance, this coming out most clearly during National Emergency (1975–1977) when, ‘the right to strike was suspended, wages were frozen, annual bonuses were reduced … ’, along with the amendment to, ‘the Industrial Disputes Act making it a requirement that enterprises employing more than 300 people receive government permission before retrenching workers’ (Hill, 2009: 398).¹ Thus, the state was committed to the balancing of class interests.

The situation, however, changed dramatically with the advent of economic liberalisation in 1991. Employers raised concern ‘that India’s labour laws and IRS system were failing to meet the needs of a globalizing economy’ (Hill, 2009: 399). The bone of contention was JSR, as employers pointed out their inability to vary employment in the face of fluctuating world demand, which they considered essential to facing international competition. The Indian state realised that the ‘need for [corporatist] social pacts is [was] no longer there as labour is [emerged as] a weaker party and the state is [was] under pressure from capital [threatening to relocate]’ (Sundar, 2018: 930). Thus, while there was no effort to improve union recognition, the state decided to remove/dilute job security.

This dilution led to a number of government committees recommending complete removal of JSR in the early 1990s. In the face of stiff opposition, the Finance Minister, Yashwant Sinha, in his Budget Speech of 2001, proposed significant dilution in JSR, by raising an organisation’s threshold number of employees, under which the regulations do not apply, from 100 to 1000 workers. Failure to implement this change, led the National Labour Commission in 2002 to recommend a softer policy of lifting the threshold to 300 workers. Learning from past experience, the Modi government advised BJP-ruled sub-national governments to amend their respective JSRs, to prevent workers from becoming united.² This strategy finally succeeded, as the Rajasthan and Madhya Pradesh governments raised the employee number threshold to 300 (Roychowdhury, 2014).

In evaluating the claims that JSR had an arresting effect on Indian employment and that it dampened Indian producers’ capacities to compete internationally, the first thing to note is that the literature contributing to these debates often overstates the reach of Chapter VB. For example, the official Economic Survey published by the Government of India (2006) for 2005–2006 notes,

Indian Labour Laws are highly protective of labour, and labour markets are relatively inflexible. These laws only apply to the organized sector … and adversely affected the sector’s long-run demand for labour. (p. 209; emphasis added)

India’s overall organised sector indeed constitutes merely 7%–8% of its total workforce (Roychowdhury, 2014). Compounding this point, JRS does not apply universally within the organised Indian sector.

JSR applies only to permanent workers employed in manufacturing units, mines and plantations. These units must have more than 100 workers and must not operate seasonally. JSR do not apply to firms engaged in services or, agricultural activities.³ On the other hand, any enterprise employing 10 or more workers qualifies for inclusion in the ‘organised’ sector irrespective of the nature of its activity (Labour Bureau of the Ministry of Labour, Government of India, 2007b: 2). In fact, the share of organised manufacturing workers in the total organised workforce never exceeded 35% between 1980–1981 and 2012–2013 (except in 2011–2012; see supplementary file Figure A1). Therefore, JSR could not be the primary reason for employment stagnation in the overall organised sector.

What, then, explains employment stagnation in the overall organised sector? Table 1 compares the employment trends in different segments of the overall organised sector with the organised manufacturing sector.

Table 1. Employment growth in different segments of the organised sector, India – calculations based on alternative periodisations.

Author’s calculation. Data on Organised Manufacturing Sector from Annual Survey of Industries (ASI), various years and all organised sector employment data from Directorate General of Employment and Training (DGET) (1981–2012).

^a Due to changes in the coverage of ASI frame over the years, suitable adjustments were made in data to make it comparable over time, following the procedure suggested by Kannan et al. (2009: 82).

The first row shows that employment growth in the long period 1980–2008⁴ in the organised manufacturing sector (governed by JSR) was 1.31%, more than double that of the overall organised sector (0.59%). Disaggregating by private- and public-organised sector shows that employment growth in the private sector (0.88%) was almost double that of the public sector (0.45%). Since the public sector share in overall organised sector employment was above 60% during 1980–2008 (Roychowdhury, 2014), it is reasonable to say that sluggish growth in public sector employment slowed down overall organised sector employment.

However, the employment slowdown in the public sector is generally attributable to rationalising the public sector units (PSUs) workforce (a process typically associated with shedding workers), along with its large-scale privatisation, with the stated aim of ‘enhancing efficiency’ (Nagaraj, 2004). In addition, the slowdown could be due to re-classification of units following privatisation, leading to employment expansion in erstwhile PSUs, newly classified under the private sector. Therefore, it seems reasonable to conclude that the conscious policy of rationalising the PSU workforce and the associated privatisation explain the incapacity of the overall organised sector to generate employment.

Sub-period analysis confirms this hypothesis. During the 1980s, coinciding with the planning era, the state carried out its responsibility to create jobs in the public organised sector. This explains robust employment growth in the overall organised sector during the 1980s. However, with liberalisation of the economy in 1991, the major pillars of the new policy framework were downsizing public sector workforce and privatisation of PSUs (Ghosh and Chandrasekhar, 2002). This led to sluggish growth in public sector employment between 1991 and 1997, which contributed to the dismal growth in overall organised sector employment. After cautiously treading the path of market-driven reforms in the initial years owing to a lack of political consensus, India launched a second generation of reforms, wherein liberalisation policies were pursued with greater vigour and determination. This explains the negative employment growth in the public sector and hence overall organised sector during 1998–2008 (an alternative periodisation has been done as a robustness check – see Table 1). Note that employment growth in organised manufacturing – governed by JSR – was higher than the overall organised sector in the post-liberalisation period. From this, it can be inferred that rationalising the public sector workforce and not JSR emerges as the main reason for sluggish employment growth in the overall organised sector.

Evidence of employment adjustment in the presence of labour laws

Here, we provide evidence on labour adjustments in the presence of labour laws in India. A pervasive view exists in the literature suggesting that labour laws limit workforce adjustment and so impede employment creation. Fallon and Lucas (1991: 396) note that ‘[M]aking jobs more secure would make employers less able to make rapid adjustments to changing market conditions. … restricting employers’ ability to fire workers may actually reduce the size of the work force employers wish to maintain’. We check this claim, and find that labour-market regulation is not a prime determinant of employment variation.⁵

Before examining employment trends, we need to identify the time period to focus on. Employment in the organised manufacturing sector started to rise secularly from 2003 to 2004 onwards (except 2012–2013), despite JSR (see supplementary file Figure A2 for the index number of absolute employment). However, for more than two decades (1979–2002) employment stagnated in the sector such that the employment index which stood at 100 in 1979 merely increased to 114 in 2003–2004. Thus, if labour regulations negatively affected employment at all, it must have been during this period. Hence, our focus period is 1979–2002.

We use a simple indicator for capturing employment adjustment through its variation; namely, the coefficient of variation (CV).⁶ If JSR restricts employment adjustment, then it is expected that CV of employment in establishments with more than 100 workers should be consistently lower than establishments with less than 100 workers. This is known as the ‘threshold effect’ in the literature (Bhalotra, 1998: 8) and we check its existence (Table 2).

Table 2. Coefficient of variation by size of employment.

Source: Author’s calculation. Data from Annual Survey of Industries (ASI), various years.

Beside the employment (direct and contract workers) CV (Panel B), we calculated CVs of (real) output (Panel A) and number of factories (Panel C). These were selected as JSR is believed to impede output growth and closure of establishments. The first column arranges firms according to the size of their workforce.⁷ Thus, the labour law applies to all employment size-classes other than the 0–99 employee category. The second column looks at variations in the long period (1979–2002). In Panel A, the CV of the 0–99 employees category (52) is lower than for the 100–199 (62) and 200–499 (56) categories. However, the output CVs of the 500–999 employees and above 1000 categories are less than the 0–99 category. Such qualifications vanish in the employment and number of factories panels since the CVs in the 0–99 employee category are consistently below all other categories.

Next, we analyse sub-periods. Sub-periods were chosen considering fluctuations of the business cycle.⁸ For each sub-period (i.e. 1979–1990, 1991–1997 and 1998–2002), the output CV in the 0–99 employee category is consistently lower than or equal to those of the other categories, except for the above 1000 employee category. However, for all sub-periods, the employment and establishments CVs, in the 0–99 employee category are lower than those of the other categories.⁹ Hence, the empirical evidence does not appear to substantiate the claims that JSR hinders output, employment and factories’ adjustment flexibility.

Table 3 examines the effect of JSR on manufacturing activities. Panel A reports the growth rate in value added, Panel B reports employment growth, and Panel C reports the growth rate in factory numbers. The growth rate in value added distinctly increased from the first sub-period to the second sub-period. This was true for the total factory sector and also in each employment category. Did employment respond to output growth? For all employment categories, employment growth in the sub-period 1991–1997 was higher than in the sub-period 1979–1990. Interestingly, employment growth in the 0–99 employee category was 1%, whereas, for higher employment categories, it was greater.¹⁰ This casts doubt over the primacy of JSR in determining employment and suggests that employment is likely to be determined by a broader set of factors.

Table 3. Growth rates of output, employment and number of factories: By size of employment.

Author’s calculation. Data from Annual Survey of Industries (ASI), various years.

Similar trends emerge for the number of factories analysis. In Panel C, for the overall factory sector and for each employment category, entry of new firms increased with an acceleration in output growth in the second sub-period compared to the first. That is, the entry of firms was not stifled by closure restrictions. Hence, there is little to suggest that JSR hampers firms’ entry or employment therein. Interestingly, between sub-periods 1 and 2, for all three variables, growth in employment categories governed by JSR was mostly higher than for the 0–99 employee category (except 1000 and above).¹¹

The experience of the third sub-period confirms that the Indian labour market is de facto flexible. Between 1998 and 2002, output growth plummeted in the overall factory sector and in each category. Significant adjustments in both employment (Panel B) and number of factories (Panel C) occurred. Both variables registered negative growth in the overall factory sector. This holds true for both variables in most employment categories. Thus, firms could adjust – despite labour laws – both employment and production units unimpeded whenever necessary.

But, how were entrepreneurs able to adjust employment rapidly despite JSR? Primarily, such adjustment has occurred because the share of contract workers in the total workforce has continuously increased in the last two decades (see Figure 1).¹²

Figure 1. Proportion of contract to total workers in the organised manufacturing sector, India.

Source: Annual Survey of Industries (ASI), various years.

It is important to remember that JSR covers only regular workers, whereas contractual workers¹³ are out of its ambit. Thus, a secular rise in the share of contract workers provides an inbuilt flexibility and is recognised in the literature as introducing ‘reform by stealth’ (Nagaraj, 2004: 3388). Moreover, even regular workers employed in establishments with fewer than 100 workers – as restrictions on firing only apply to firms with 100 or more workers – are subject to arbitrary retrenchment. Hence, adding regular workers in the 0–100 employee category to the contract workers’ pool gives the true proportion of workers not protected by JSR (see supplementary file Figure A3 for details).

Interestingly, the industry lobby/employers themselves admit that ‘hiring contract labour is an important tool for labour market flexibility and this should be facilitated’ (Sabharwal and Arora, 2018: 303); for the ‘contract labour system’ helps to overcome the restrictions on firing set by JSR (Sundar, 2012).

A logical corollary of this argument might be that the abolition of JSR would help to reduce contract labour and register a rise in regular work. However, one remains sceptical about such an outcome for the following reasons. First, Sood et al. (2014: 60) note that between 2000 and 2009, the share of contract workers rose not only in firms employing 100 or more, but also in the 0–99 employee category where JSR does not apply. This implies that JSR is not the only reason for using contract workers. Further primary studies confirm this hypothesis. Barnes et al. (2015: 355), studying the automotive components industry in the national capital region of Delhi, document that the ‘contract labour system has enabled employers to keep wages low, increase firm flexibility, offload the burden of monitoring and controlling workers and undermine collective bargaining and TU rights’.¹⁴ Second, the abolition of JSR protection would itself change the meaning/nature of regular work, especially when there is evidence of high costs for workers seeking changes to their working conditions. Barnes et al. (2015: 364) report that an unprotected ‘contract worker who complained or asked for improvements was ignored or scolded harshly by managers … [and was given] precarious employment arrangements’. This high cost, they reported, was not tempered by membership of labour organisations, as ‘only regular workers were union members … [and even then] have been dismissed or victimised by employers … after they filed for [union] registration’ (Barnes et al., 2015: 366).

Third, the Contract Labour (Abolition and Regulation) Act (CLARA) 1970 that prohibits the use of contract workers in the main/core activities of a firm is routinely flouted with full impunity. Barnes et al. (2015: 363) note that ‘employers were openly violating the CLARA’, with contract workers being ‘recruited into ‘core’ business areas on a long-term or on-going basis’. Even then, they report, ‘we are not aware of cases in which the Government of Haryana has used … CLARA to prohibit the employment of contract labour’ (Barnes et al., 2015: 363). Moreover, in rare cases of enforcement, the paltry fines of Rs 100 or Rs 200 charged for violations hardly act as a deterrent (Sundar, 2008). For all these reasons, the abolition of JSR is unlikely to result in a rise of regular work.

In fact, the incremental use of contract labour is an effective way to introduce informalisation to the formal sector. It further inflates India’s already large informal workforce (93%) and is part of the broader trend to a rise in informality across the globe. Global supply chains may be traced back to precarious contract work, such as that undertaken by female homeworkers in Pakistan’s football stitching industry, characterised by conditions similar to those of contract workers in Indian manufacturing – absence of social security, uncertain work, no bargaining power, and wage theft (below-minimum payment) (Naz and Bögenhold, 2018). Although informal labour contracts are more pervasive in developing countries, they are detectable in certain sectors of developed nations as well. Delaney et al. (2018) show how Australian home-based garment and child-care workers receive significantly less than the statutory minimum pay, lack superannuation, and work excessive hours in risky environments with low occupational health and safety standards and no protection from unfair dismissal. At the risk of invisibilisation, contract workers everywhere are isolated and without scope for bargaining collectively. Through layers of subcontracting, their employment relationships are obscured, making it difficult to claim unpaid remuneration. Not recognised as employees but deemed ‘independent contractors’, they are denied proper working entitlements and regulatory protection. A recent manifestation is online ‘gig’ work, characterised by lack of control over conditions, piece rates and irregular income flows (Stewart and Stanford, 2017: 431). Such ‘sham contracting’ involves denial of rights such as paid annual leave, personal injury insurance claims, minimum wages, working hours restrictions, sick leave and superannuation benefits. Such work, governed by ‘commercial rather than employment law’ (Minter, 2017: 444), undermines legislated minimum labour standards, with broader ramifications for traditional sectors of the economy. Thus, the rise in informal work is a worldwide phenomenon and not unique to Indian manufacturing.

In India, although contract labour remains the main route to increased flexibility, employers adopt varied means to bypass JSR. One major escape clause from the requirement for employers to gain government permission to fire workers is for firms and workers to separate ‘voluntarily’. Employers frequently use coercive techniques to obtain employees’ consent and the government itself facilitated this method by adopting voluntary retirement scheme in 1999 (Nagaraj, 2004). Another method used in India relies on the fact that retrenchment law is only applicable to workers in continuous service for 240 days. Field studies report that employers ‘commonly terminated and re-hired workers before they complete 240 days of “continuous service” … to avoid this obligation’ (Barnes et al., 2015: 363). A third common way to get around JSR is to offer employment for a fixed period, known as ‘fixed-term employment’. Removal of workers due to non-renewal of contract – beyond the stipulated period – does not attract litigation (Sood et al., 2014) and is consciously and aggressively promoted by the Modi government as the principal means to achieve flexibility.¹⁵ Finally, there are widespread reports of the use of illegal means to circumvent the law. Employers resort to such means as violations attract the meagre penalty of Rs 5000 with no instance of any employer ever being sent to jail (D’Souza, 2010). Furthermore, ‘government authorities are often manipulated [by employers] to delay consent sought for retrenchments because after two months it is considered automatically to have been approved’ (D’Souza, 2010: 130).

The discussion above shows that JSR has largely been confined to the statute book. It follows that our empirical finding of labour laws not putting a drag on employment or output growth is hardly surprising. The analysis also casts doubt over how far the recent amendments to JSR in Rajasthan and Madhya Pradesh, raising the threshold number of workers from 100 to 300, are likely to yield the desired results. It should be remembered also that Uttar Pradesh amended its JSR along similar lines in 1983, but has not subsequently been among the industrially advanced states with significant employment growth (Bardhan, 2014). In fact, explaining sluggish employment growth in Indian manufacturing solely in terms of the few legislative restrictions is fraught with difficulties. For example, Bardhan (2014) cites the case of highly labour-intensive garment industry:

About 92% of garment firms in India have fewer than eight employees (the bunching of firms is around the eight-employee size, not the below-100-employee size, as one would have expected). Labour law cannot discourage an eight-employee firm from expanding to an 80-employee firm since Chapter VB of the IDA does not kick in until the firm reaches the size of 100 employees. So the binding constraints on the expansion of that eight-employee firm may have to do with inadequate credit and marketing opportunity, erratic power supply, wretched roads, bureaucratic regulations, and so on.

It follows that employment is actually determined by a combination of factors, and labour regulation is not its prime determinant.

Assessing workers’ bargaining power in the organised sector with special reference to organised manufacturing

Next we investigate workers’ bargaining power in the organised sector. Enhanced workers’ bargaining power through rising TU activity, it is claimed, can be expected to dampen employment creation by raising the effective cost of labour. Furthermore, JSR is supposed to bestow greater bargaining power by protecting employment.

To examine these claims, we examine trends in workers’ bargaining power in the organised sector as a whole to investigate whether it could contribute to the sector’s sluggish employment growth. We then separately examine workers’ bargaining power in the organised manufacturing segment. We adopt a method devised by Lucas (1988) for scrutinising labour market rigidities, ‘[f]ollowing Lucas in examining the trends in union power as a proxy for the alleged rigidities [in the labour market]’, to use the words of Nagaraj (1994: 180). In order to examine trends in union power we choose certain indicators (some of these used by Lucas) both for the total organised sector and its manufacturing subdivision.

At the outset, it is imperative to mention that the proportion of functional (submitting returns) TUs to total registered TUs, for all organised industries steadily declined from about 25% in the mid-1980s to around 10% in 2008–2009 (see Figure A4).

Furthermore, Figure 2 shows that the number of strikes – which are used as a proxy for TU power – has steadily fallen. However, the same is not true for lockouts – which showed a stable trend until 1998–1999, falling thereafter. In early 1980s, strikes were nearly four-times the lockouts; whereas from the mid-1990s, they tend to converge. Evidently, union power in recent years declined considerably relative to employers’ power.¹⁶ Sharma (2006: 2083) notes that the ‘fear of losing jobs has impelled unions to accept relocation, downsizing, productivity linked wages, freezes in allowances and benefits, voluntary suspension of TU rights for a specific period and commitment to modernisation’.

Figure 2. Number of strikes and lockouts in India: all industries.

Source: Labour Bureau of the Ministry of Labour, Government of India (1982b to 2014b).

However, it may be argued that the effectiveness of disputes does not depend on the number of disputes, but on the number of workdays lost during disputes. On account of the workdays lost per dispute (both strikes and lockouts), Figure 3 shows that lockouts have far more impact than strikes. The relative share of workdays lost due to strikes and lockouts reveals that the share of workdays lost due to lockouts has mostly exceeded that of strikes (see supplementary file Figure A5). Thus, the effectiveness of strikes is far less than that of lockouts.

Figure 3. Number of person days lost due to disputes (work stoppages): all industries, India.

Source: Labour Bureau of the Ministry of Labour, Government of India (1982b to 2014b).

Turning to the manufacturing sector, we investigate whether the 1982 IDA amendment fostered labour activism.

Figure 4 shows a stable trend in the unionised share of the workforce relative to total workers over 28 years, with some improvement coming in the last 2 years. There is no evidence of JSR translating to higher labour activism in the form of a rising proportion of unionised workers. However, this criterion may be considered insufficient in and of itself, as it says nothing of the rates of activism of existing unionised workers, which might have risen. Figure 5 plots the number of disputes (strikes plus lockouts) in the organised manufacturing sector and shows a drastic downwards trend during this period. Moreover, the number of workers involved in disputes and their share in the total organised manufacturing workers declined more or less consistently (see supplementary file Figure A6).

Figure 4. Unionised workers as a proportion of total workers: manufacturing sector, India.

Source: Labour Bureau of the Ministry of Labour, Government of India (1982b to 2014b); see also notes for Figure A4.

Figure 5. Number of disputes (strikes and lockouts) in the manufacturing sector, India.

Source: Labour Bureau of the Ministry of Labour, Government of India (1982a to 2015a, 1982b to 2014b).

Finally, small but strong unions may paralyse economic activity, entailing large losses to enterprises. To check this possibility of increasing union strength, we calculate the share of workdays lost to total workdays worked (Figure 6). This is unambiguously falling over time, leaving little room to suggest any rise in labour militancy.

Figure 6. Person days lost as a proportion of total person days worked: manufacturing sector, India.

Source: Labour Bureau of the Ministry of Labour, Government of India (1982a to 2015a, 1982b to 2014b).

The indicators above reveal that workers’ bargaining power and labour market rigidities were declining in both the organised sector as a whole as well as in the organised manufacturing segment, refuting any claim of a rising militant trade unionism hindering employment growth.

Next, we examine real wages (RW) and product wages (PW)¹⁷ in the organised manufacturing sector. Both metrics are included as the living standards of workers do not depend on PW but on RW, since industrial workers consume food, which is produced outside the manufacturing sector.

Wages are an important indicator of workers’ bargaining power. Nagaraj (1994: 179–180) argues,

[t]he increase in wage rate … is seen as an evidence of growing rigidities in the labour market – namely, minimum wage legislation, (growing) strength of trade unions and increasingly stringent job security laws – as reflected in the power of organized labour to appropriate a share of output disproportionate to their contribution, at the expense of additional employment generation and with a socially undesirable rise in capital intensity.

To check this claim, Figure 7 plots index numbers for RW and PW. Both were rising at a roughly comparable pace until 1995–1996, after which they diverged. The increasing gap between PW and RW signifies that although workers were somewhat successful in bargaining for higher money wages (based on their own product prices), nevertheless these increases were inadequate to keep pace with rising general prices. Especially after the downturn hit the manufacturing sector in 1997–1998, there was a sharp absolute decline in both PW and RW. Furthermore, as the pace of PW growth decelerated thereafter (up to 2008–2009, after which there were some signs of recovery), the RW virtually stagnated for the next one and a half decades. Chandrasekhar and Ghosh (2007) document the post-reform trends in the RW:

[T]he average real wage of workers in the organized manufacturing sector has been more or less constant right through the 1990s. Average RW increased in the early years of the 1990s, until 1995–1996, and then fell quite sharply. The subsequent recovery after 1998 has been muted, and RW have stagnated since 2000. As a result, RW in the triennium ending 2003–2004 were around 11% lower than RW in the triennium ending 1995–1996.

Figure 7. Index number of real wage (RW) and product wage (PW).

Source: Annual Survey of Industries (ASI), various years.

This analysis shows that unions could not successfully raise the RW and PW through bargaining. It has implications for the class distribution of output and employment growth – especially when labour productivity increased (Roychowdhury, 2017: 8). As wages could not keep pace with rising labour productivity – due to reduced TU power and effective labour flexibility – the wage share in manufacturing output fell sharply from around 27% in 1980 to merely 12% in 2013–2014 (Roychowdhury, 2017: 9). Since the decline in union power was not restricted to the manufacturing sector alone, the skewed distribution of output was more widespread – evident from the dramatic rise in income inequality. For example, Piketty and Chancel (2017) estimate that while the share of national income going to the top 1% of the population increased from 6% to 22% between early 1980s and 2015, during the same period the share of the bottom 50% of the population decreased from 23.6% to 14.9%. This growing disparity clearly could not have happened without wage compression and a rise in the profit/surplus share at the economy-wide level. This, following Patnaik (2011, see section 1), is likely to have constrained employment growth.

Taken as a proxy for workers’ bargaining power and labour market rigidities, it is clear that union power declined in the last three decades; in contrast, employers’ bargaining position strengthened. Thus, the proposition that a rise in the bargaining power of workers hampered employment growth does not withstand scrutiny.

Conclusion

We have examined the two propositions normally advanced in the literature to explain sluggish employment growth in the organised sector of India – namely, rigid labour laws and enhanced workers’ bargaining power. We showed that JSR applies to merely one third of the overall organised sector employment and that sluggish job growth is better explained by policy-induced privatisation and the restructuring of PSUs.

Concentrating on the organised manufacturing sector – that is governed by JSR – we found no evidence of JSR arresting employment adjustment and creation. Such adjustment was indeed made possible by the indiscriminate use of contract labour, circumvention of labour law and weak enforcement.

Finally, analysing the trend in workers’ bargaining power to ascertain if it contributed to employment slowdown, we showed that workers’ bargaining power and labour rigidities were declining, while the position of employers was being strengthened. Thus, the two conventional claims forwarded to explain employment stagnation in the organised sector of India found no empirical support.