Réglementation du travail et dynamiques de l’emploi au niveau de l’État en Inde
Résumé :
Au cours de la dernière décennie, les résultats du marché du travail se sont améliorés en Inde, l’emploi net augmentant de façon sensible dans l’ensemble de l’économie. Cependant, ces gains sont intervenus essentiellement dans les secteurs non organisé et informel de l’économie, où la productivité et les salaires sont généralement bien moindres que dans le secteur organisé formel. Seul le secteur organisé est assujetti à la réglementation du marché du travail et, dans ce secteur, l’emploi a diminué. L’incidence de la législation de protection de l’emploi sur la performance du marché du travail est sujette à controverses aussi bien dans la zone de l’OCDE qu’en Inde. Le présent article examine l’impact de cette législation et des réglementations associées sur la dynamique de l’emploi dans le secteur organisé de l’économie, en utilisant de nouvelles mesures de la réglementation nationale et des réformes au niveau des Etats. Il ressort de cette étude que si les réformes ont permis d’assouplir quelque peu la rigidité du droit fondamental du travail, des mesures plus approfondies sont nécessaires pour remédier aux distorsions qui sont apparues.
REGULATION AND EMPLOYMENT DYNAMICS AT THE STATE LEVEL IN INDIA
Employment has expanded in the economy as a whole, but not in the organized sector Contrary to widespread concerns about so-called jobless growth, employment expanded in five out of the last six years for India’s economy as a whole, as well as for manufacturing, according to updated estimates based on the most recent (2004/05) round of the Government of India’s National Sample Survey (NSS).2 Such estimates imply that employment growth has improved considerably, doubling the 1.3% pace of growth during the 1990s over the first five years of the 2000s. The pace of employment growth of over 2½ per cent per annum so far in the 2000s also exceeds the 1.9% pace of growth during the 1980s. Such an outturn is well above the previous expectations of many observers and represents a major upward shift in India’s labour market performance (see Nagaraj, 2004; Anant et al., 2006; Government of India, 2006).
Gains in employment from 1998 to 2005 came about nearly equally from increases in secondary and tertiary employment (representing 19% and 25% of the labour force, respectively). About 30 million net jobs were added in each of the two sectors, representing an annualized growth rate of about 6.2% for the secondary sector and 4.7% for the tertiary sector. These growth rates represent a turnaround from the anaemic pace of employment growth for much of the 1990s, and even exceed the employment growth rates of the 1980s. Net employment gains in agriculture on the other hand slowed to ½ per cent per year, as more workers moved to off-farm employment. Perhaps what is most surprising – given that industrial production has increased less than total output – is that the pattern of employment in industry (and manufacturing) has followed that of the economy as a whole, with overall net employment growth of 6½ per cent per year over the 1997 to 2004 period.
Despite strong gains in employment across the economy in recent years, a dichotomy has emerged with net increases in employment occurring almost exclusively in the least productive, unorganized, and often informal part of the economy. Meanwhile, the organized sector’s employment level has actually been shrinking.3 In industry, from 1997 to 2004, the overall 6½ per cent annual growth of employment disguises an average net loss of employment in the organized sector of about 1% per year over the period, while unorganized manufacturing employment grew at a rate of about 8% per year. As a result of these developments, the relative share of the organized sector in industrial employment fell from 15–20% in 1981–98 to only 10% in 2004, a decline that was much more rapid than for the economy as a whole.
The slight contraction of organized sector employment disguises considerable underlying job dynamics In order to better understand what has happened to the fortunes of the organized sector, this paper examines gross flows of jobs in organized manufacturing and the potential influence of labour regulations on these dynamics. Our methodology for measurement of job flows is described in Box 1. The resulting job creation, destruction and turnover figures capture the annual dynamics of employment at the level of detailed industry branches from 1998 to 2004 and represents the first study of India to have compiled gross job flow statistics using unit-level microdata. These figures reveal that the slight contraction of net employment in the organized sector disguises considerable underlying job dynamics, with rates of job destruction often higher than those for creation.
Measuring job flows in India
In empirical work carried out for this paper, seven years of Annual Survey of Industries (ASI) microdata covering the entire organized manufacturing sector were analysed to create a longitudinal sample of employment dynamics at the detailed industrial branch level, country-wide and for individual Indian states. This analysis used the basic methodology of OECD (1996) and Davis et al. (1998) to compute job creation and destruction statistics, with adjustments for the level of aggregation following the approach of Singh (2006). The resulting figures capture changes in employment levels for detailed manufacturing industry branches between adjacent years from 1998/99 to 2003/04, relying upon plant-level microdata.
The preferred approach for computing such job creation and destruction statistics is to use longitudinal microdata, matched over time at the establishment level. This was impossible because the CSO does not disclose plant-level identifiers that can be matched across time. So instead, a pseudo-longitudinal panel was constructed that uses detailed (three and five-digit) industry branch codes based on each plant’s industry of operation. Such s should in principle capture all job reallocation across industries, although they are likely to underestimate the actual turnover of jobs within-industry, as plant-level flows cannot be directly observed. Therefore, they represent a probable lower bound on job creation and destruction rates, although measurement error from non-responding plants could raise measured job flow rates to the extent they are not fully captured in sampling weights (see Ahmad, 2006).
Although this means that the resulting figures are not comparable with other countries’ job flow statistics, they are internally consistent and thus can be analysed across time and sector.
Our estimates of job flows at the national level finds considerable employment dynamics at the five-digit, and even three-digit, levels of aggregation (Table 1). At close to 4% per year through 2004, the annual rate of job creation has eased slightly since 1985. At the same time, the rate of job destruction has picked up from about 4% per year to 5%, resulting in a deterioration of net employment growth from about +1% per year pre-1998 to -1% per year in the most recent period. This shift in employment outcomes has coincided with a slowing of job turnover and reallocation.
Using a more detailed five-digit industry branch level of aggregation, additional job dynamics are picked up in the analysis, implying a 14% rate of job creation and 15% rate of job destruction since 1998. Such rates are as high as plant-level job flows in the US and France, although they are lower than (firm-level) estimates for China, where there has been more job destruction (Deng and McGuckin, 2005).
Given this apparently high degree of jobs dynamics in the organized sector, one would expect that the regulatory stance of labour laws would be weak. However, India’s labour market regulations for the formal sector are much more stringent than those of most OECD countries and several major developing countries as well. A standardized evaluation of India’s employment protection legislation (EPL) using the OECD EPL methodology was carried out using legal information supplied by the Government of India. The OECD methodology, outlined in the 1999 edition of the OECD Employment Outlook, evaluates the stringency of a country’s labour regulations in three areas: those for regular (indefinite) contracts, for temporary or fixed-term contracts, and for collective dismissal. Following this approach, India’s federal labour laws were scored, as they stood in early 2007.
For regular contracts, India’s labour laws are stricter than those of all but two OECD countries, Portugal and the Czech Republic, This stringency comes in large part from the procedural difficulty of having to obtain prior government permission to lay-off just one worker for plants covered by the Industrial Disputes Act (IDA). Chapter V-B of this Act, which requires such approval for all factories with more than 100 workers, removes much of the distinction between the EPL for regular employment and for collective dismissals. If this provision were not in force, the EPL for regular contracts would fall to the OECD average, which happens to be nearly the same as China’s score. This is indeed the situation for most employees in India’s tertiary sector, yet the concern remains for the secondary sector.
Even then, though, dismissing workers in non-IDA plants on performance grounds is difficult, as standard notification and severance obligations tend to be quite high, and courts often order reinstatement rather than compensation for unfair dismissals.
The stance of India’s EPL regime is more relaxed only for temporary and fixed-term contracts, placing it just above the mean of OECD countries’ score for such contracts. The moderate score of India for such contracts reflects the main area of labour reform that has come about in recent years, allowing employees to work on temporary work agency contracts to carry out a range of “non-core” activities, a concept that is defined in various ways across states but is usually restricted to activities that are ancillary to production.4 Standard fixed-term contracts are allowed for white-collar workers as well as, in principle, for regular workers, following a federal notification in 2003, and about half the states have implemented analogous provisions, but public awareness of the state-level provisions is low and repeated attempts to
reverse the central notification may be deterring its use.
The indicator score runs from 0-6, representing the least to most restrictive recent shift in the temporary EPL stance was a recent Supreme Court decision (2006) that clarified the
minimal restrictions on renewal of temporary contracts (although severance payments proportional to job tenure are required). This decision represented an important shift from the prior possibility of contractors being converted to indefinite contracts which could then cause the restrictions on regular contracts and collective dismissals under Chapter V-B of the IDA to come into force.
The third measure of the economy-wide stance of EPL is for collective dismissals.
Here, India’s stance is especially onerous, more-so than any other comparison country. Nearly every aspect of this measure rates at the top end of the scale, from the definition of collective dismissal (one worker), to notification and delays involved, given requirements for obtaining prior government consent. Most of these rules are encompassed in the IDA legislation, and come into force for any plant with more than 100 workers, encouraging many plants to stay small and restrict their scale of operation.
As a whole, India’s federal EPL stance is highly restrictive for the organized sector that it covers, given its interference in a number of important respects with firms’ hiring and severance decisions.
Nevertheless, job flow statistics suggest that considerable dynamics exist despite the regulations. This seeming mystery could be a result of enterprises altering their behaviour to reduce the impact of the rules, or it could be that in practice EPL application or enforcement is less onerous than the regulations suggest – a possibility that is likely to vary considerably across states, given that constitutional responsibility for state labour regulation is shared between the centre and the states. Both of these possibilities are explored.
Job dynamics under heavy labour regulation are uneven and come at a great cost
While the degree of job dynamics appears to be considerable, it varies widely across different segments of the organized labour force. The unbalanced coverage of employment protection legislation (EPL) in India appears to be important in explaining the variation of job flows across segments, suggesting that EPL is distorting the decisions of employers and accentuating labour market duality:
• The EPL index for staff employed with fixed-term or temporary contracts – in India typically supplied by a labour contracting agency – showed a moderate degree of restrictiveness, similar to the average for OECD countries. Correspondingly, over the period 1998 to 2004, job creation amongst contractual staff consistently dominated job destruction, allowing the net employment of contractual staff to rise by 10% annually, while it declined by 2½ per cent a year for regular workers, for whom there was much less job creation (Table 2).
• For small firms employing 100 or fewer workers, the EPL index is also similar to that found in OECD countries and, over the same period, job creation rates were more than double job destruction rates for this group of firms, causing employment in these smaller firms to rise by more than 20% a year. Moreover, despite the absence of restrictive dismissal laws, the job destruction rate in small plants was actually less than that in larger ones.
• In contrast, in large firms with over 100 employees (where the EPL index is high), the job creation rate for (non-supervisory and non-contractual) regular workers was low and well below the job destruction rate, with the result that the net employment of such staff fell by more than 5% per year in the period 1998 to 2004. In contrast, job creation for contract staff in large firms was twice the rate for large firms’ regular workers, and despite a higher destruction rate as well, their net employment increased almost 4% per year.
The tendency for large firms to restrain job creation is associated with a marked substitution of capital for labour that is not seen in small firms. Over the 1998 to 2004 period, constant price fixed assets per employee in plants with more than 100 workers rose by 21%, while they fell by 14% in smaller firms, according to tabulations by OECD using the Annual Survey of Industries (Table 3). While the large plants were already 28% more capital intensive, this raised their capital intensity relative to small plants by more than half plants over the 1999 to 2004 period. While public plants also have higher rates of job creation than private plants, public plants’ net contribution to employment is highly negative, on the order of a 10% drop in employment each year. This result may appear to be strange given the presumably strict compliance of public firms with the IDA legislation and the usual motive of public firms to preserve employment.
How could turnover in public enterprises be so high? The main explanation for this extensive (public plant) turnover is that voluntary retirement schemes (VRS) have been accepted by workers and unions as a mutually agreeable mechanism for downsizing. VRS schemes became common after the government set up a National Renewal Fund in 1991 as part of the structural reform program, to fund restructuring of public sector enterprises; only later did the private sector begin to make use of VRS (Nagaraj, 2004).
The VRS scheme has made employment adjustment feasible; however, it comes at a great cost. Since it requires that settlements are collectively agreed upon, high levels of compensation are made to groups of workers. Typically, the reservation payoff to the most recalcitrant marginal worker determines the outcome of the negotiations. OECD interviews with employers and local governments suggest that such payoffs can easily reach levels of two to three months of compensation for each year of work.6 A financial analysis of public sector redundancy schemes suggests that total average payments to staff that left ranged between almost four to over six years’ pay (Morrison, 2006). The redundancy compensation component (as opposed to normal leaving benefits such as accrued leave and provident fund balances) averaged between 10 and 36 months, with central government enterprises tending to pay the most. Such a cost places India among the most expensive in the world in terms of cost of dismissal, and is far above the statutory compensation that was intended by the IDA act (15 days of compensation per year of service). So while employment flexibility is feasible, it comes with an increased cost of severance – and not just to the employer who would normally have to pay for the VRS, but to the worker as well, in part because some businesses avoid payment altogether, but also because the costs create behavioural distortions that have detrimental economic effects.
High labour costs in the formal sector increase capital intensity and the size of the informal sector
Such high potential compensation for dismissal and the associated uncertainty raises the cost of employment, inducing larger employers to substitute more capital for labour than would appear justifiable given the apparently low wages that prevail in India.8 This effect can be observed in the low – and falling – labour share in organized industry, which can be observed in all two-digit industries from the 1980s. The share of total labour costs in value added has fallen from 36% in the early 1990s to 31% in 1999/2000 and 29% in 2003/04, with the steepest fall among large enterprises.9 The share of wages (including bonus payments) in value added has fallen similarly, from 25.7% in 1998/99 to 23.5% in 2003/04.10 Average wages have risen at a slower pace than labour productivity, so that labour compensation relative to Germany has been stagnant since the early 1990s. Moreover, wage inequality has grown considerably, with most gains going to the top two deciles, while median wages have remained stagnant (see Narain, 2006).11 In addition to causing business to substitute capital for labour, high implicit costs of employment for large-scale firms have induced many entrepreneurs to start small and stay small. Businesses in the unorganized (and often informal) sector with fewer than 10 or 20 workers (depending on the law) are subject to very few labour regulations and can employ casual or contract labour freely, as can most information technology (IT) and business process outsourcing (BPO) companies. Even businesses in the organized sector with fewer than 50 workers face no legal restraints on collective dismissals, while those having 100 or fewer workers are not subject to the IDA requirement to obtain prior permission before dismissing workers (or shutting down). As a consequence, the size distribution of businesses in India is highly skewed toward small units, with visible peaks in the distribution of plant sizes at ten and 20 workers. Considerable anecdotal evidence illustrates the extent to which entrepreneurs deliberately disintegrate their business activities into numerous small-scale units (Mukerji, 2006).12 As discussed in Dougherty et al. (2008), such small scale production reduces the efficiency of the economy and limits the productivity and incomes of businesses and workers.
The direct impact of India’s labour regulations are subject to considerable debate
The suggestion was made earlier in this Survey that the local business environment, and stringent labour regulations in particular, may have a deterrent effect on firm entry and investment, and thus could reduce potential output growth and possibly job creation as well. However, the empirical evidence on the role of labour regulations in India has been subject to a vigorous debate.
A series of economy-wide studies have examined the role of two key amendments to the Industrial Disputes Act (IDA) affecting collective dismissals (see Anant et al., 2006; Bhattacharjea, 2006). The 1976 amendment mandated prior government permission for layoffs, and the 1982 amendment lowered the size threshold from establishments with more than 300 workers to those with more than 100. Roy (2002) examined accessions and separations of individuals from firms from the 1960s through 1992, and she found that the two amendments had a strong and adverse effect on labour market flexibility, a result consistent with the cross-country evidence. However, flexibility may have increased for other reasons, given the growing use of contract labour from the early 1980s. The amendments did not appear to have an impact on the slow speed of employment adjustment to shocks in India (5-6 years), a finding that may reflect the broader range of rigidities that exist in India: especially weaknesses in the insolvency framework and urban land laws (Anant and Goswami, 1995; Roy, 2004).
The variation that exists in India’s labour laws across its states has offered fertile ground for empirical exploration by labour economists from around the world. Several well-known studies have been carried out, the best-known being the study by Besley and Burgess (BB) (2002, 2004), who compiled an index of state-level amendments to the IDA from 1949 through 1992, and classified them as to whether they were pro-worker or pro-employer. Their analysis of the indicator suggests that pro-worker amendments resulted in lower output, investment, employment and productivity in formal manufacturing although output in the informal sector increased. Their results also suggest that the more restrictive labour amendments had a perverse effect on the poor as they tended to increase urban poverty rates.
The BB index and its interpretation have turned out to be highly controversial and have been utilized as well as critiqued in multiple papers. Studies such as Hasan et al. (2003), Topalova (2004), and Sanyal and Menon (2005) use modified versions of the index (through 1992) to classify states as pro-worker or pro-employer, and find several things: states with more pro-worker labour legislation have lower elasticity of demand for labour, higher rates of poverty, and less investment, in part through interactions with trade liberalisation and state-level institutions. Further work by Aghion et al. (2005) updated the index (through 1997) and, using the time series of the index, found higher rates of growth following industrial de-licensing in states with more pro-employer labour institutions. Ashan and Pagés (2007) carried out a full re-scoring of the index to measure transaction costs and found at least as detrimental an impact on output and employment as the original BB study. Ashan and Pagés also separated the index into several subcomponents and found that regulations that increase the cost of settling disputes appeared to be more costly for employment than the IDA’s restrictions on firing and closure. On the other hand, a separate study by Purfield (2006) that further updated the BB index (through 2002) did not find significant detrimental effects of more pro-employer labour institutions on state-level growth.
Despite the robustness of the link between the BB index and poor labour market outcomes across different states in most of these studies, there are reasons to think that the index could be improved due to the sparseness of underlying amendments and its focus on just one law. Anant et al. (2006) and Bhattacharjea (2006) point out that many of the formal state-level amendments to the IDA used in the index would appear to be of minimal significance since they are infrequent, sometimes pertain to obscure procedural matters, and are likely to be overshadowed by court decisions and informal changes in the application of the laws. An update of the BB index carried out as background to this Survey (through 2005), using the most recent edition of the original data source for the index (Malik, 2006), shows that there have been only eight amendments to the IDA in any Indian state since 1990, and they relate to only three states, of which only the 2004 amendments in Gujarat would appear to be of any consequence to labour market outcomes.
Several prominent observers of India’s labour markets now argue that regulations other than the IDA are likely to be more important and that the IDA may be something of a distraction in the debate over labour reforms, given the four dozen central laws and hundreds of state laws that govern labour issues (Nagaraj, 2004; Debroy, 2005; TeamLease, 2006). There have been far fewer studies that look more broadly at India’s labour laws in systematic empirical analysis, perhaps due to their complexity and overlapping coverage. Moreover, some of the most important changes appear to have occurred in interpretation of the laws by the Supreme Court and through changes in the way the laws are enforced, particularly obvious in the case of the use of contract labour, which has become increasingly widespread in many states (Anant et al., 2006). Such state-level reforms have been the subject of speculation, but there has been little systematic regulatory information collected beyond the IDA amendments in the BB index.14
State-level labour reforms
A new OECD index of state-level labour reforms in India may illuminate the way forward
In order to better understand how state-level reforms may have affected labour markets, the OECD developed a customized survey instrument to identify the areas in which Indian states have made specific discrete changes to the implementation and administration of labour laws. This state-level survey covered eight major labour legal areas, identifying 50 specific subjects of possible reform, many of which could be implemented by administrative procedure rather than through formal amendments to the laws. Since few of these changes have been systematically documented, the assistance of the All-India Association of Employers (AIOE) was used to survey 21 states, about half of which were visited in person by members of the survey team. Preliminary answers to the questionnaires were drafted in most states by a labour expert designated by the AIOE or the Federation of Indian Chambers of Commerce and Industry (FICCI) affiliate in the state capital. The answers to this draft questionnaire were then corrected through discussions with local union leaders, independent labour experts, employers and state labour commissioners. Finally, the completed questionnaires were submitted to each state’s labour secretary for their final review.15 The verified responses to the survey were compiled by the OECD into an index that reflects the extent to which procedural changes have reduced transaction costs, through limiting the scope of regulations, providing greater clarity in their application, or simplifying compliance procedures. Answers were then scored as “1” if they reduced transaction costs, “0” if they did not, and (for two questions) “2” for a further reduction, with a maximum score of 50. A list of the subjects and summary scores are shown in Annex 1.
The reforms covered in the index concern eight specific areas: the Industrial Disputes Act (IDA), Factories Act, State Shops and Commercial Establishments Acts, Contract Labour Act, the role of inspectors, the maintenance of registers, the filing of returns and union representation.
A breakdown of the average index score across all states is show, which illustrates that the largest number of reforms that were identified concern contract labour, an area where many reforms have been carried out (especially for IT/BPO sectors and for export processing and special economic zones), but also an area for which there were more questions on the survey than for other areas.
In order to adjust for the number of questions in each area, the number of cases where reform had occurred was taken as a proportion of the number of possible subjects for reform in a given reform area (Panel B, out of 100). This proportional index shows that, with the exception of reforming the role of inspectors and the rules concerning the role of unions, there was a fairly similar share of the reforms carried out across the different areas surveyed.
What is perhaps most notable about the indexes, though, is the very modest degree of reform that has actually been carried out since, even in the area of contract labour, there have been reforms in only about half of the subject areas. Moreover, no state has a raw score of more than 28 out of 50 of the possible reform subjects, and for seven of these, no more than two states made any reforms. The principal areas where there was little change were: key sections of the IDA (that concern collective layoffs); rules governing working hours; union recognition and reduction in the number of inspections.16 Nevertheless, looking across states, enormous variation in the extent of reform appears across all of the regulatory areas. Ranked lowest in terms of the number of reforms carried out are Chhattisgarh, Goa and Bihar; at the same time, Gujarat, Uttar Pradesh and Andhra Pradesh are ranked highest. There is no dominant pattern in terms of the areas where the highest-ranked states have reformed, as the reforms are quite broad-ranging. However, it is notable that the three lowest-ranked states have made virtually no reforms in the areas of inspections, simplification of registers or the filing of returns. It is also of some significance that all states have made at least some reforms in the areas of the IDA and Contract Labour Act, for which the variation is the lowest among the reform areas.
State-level labour reforms do seem to explain differences in flexibility
The extent of labour reform captured in the above cross-state indices does appear to be linked to the flexibility of the labour market. In order to measure flexibility, gross job flow statistics at the five-digit industry level are computed using the plant-level microdata for formal sector enterprises in each Indian state, separately, over the period 2000 to 2004. The resulting state-level job turnover rates are then compared with the state-level labour reform index to see if there is a relationship. As the trend-line illustrates, the correlation between the reform index and job turnover rates is very high (R2 = 0.37), suggesting that states that have made more reforms to their labour rules and regulations indeed obtain more flexibility (in terms of job dynamics) in their labour markets and vice-versa: fewer reforms significantly limit flexibility.
The increased job dynamics come from higher rates of both job creation and destruction, which tend to offset each other almost entirely in terms of net employment.
There is broad cross-country evidence to suggest that job security regulations raise the cost of employment and have a detrimental effect on the creative-destruction process that is vital to economic growth and technological progress (OECD, 2004 and 2006). Job security regulations limit job turnover in sectors that for technological or market structure reasons need more frequent adjustment of employment to be competitive. In the case of India, the results show that state-level labour reforms appear to increase job turnover (or flexibility), and a lack of such reforms limits turnover. Across Indian states in the 2000 to 2004 period, an increase in turnover rates can be shown to limit the fall in the labour share of value added that has been observed nationwide. Yet no state is able to do better than to stabilize their labour share. More broadly, increases in job turnover can help productivity and output growth (in part through better job matching), and therefore increase the pace at which incomes converge to those of more developed countries (see Caballero et al., 2004; Micco and Pagés, 2004; Haltiwanger et al., 2006).
The OECD job strategy re-assessment also found that a lack of employment flexibility accentuates labour market inequalities, as it makes it more difficult for vulnerable workers such as women, minorities and the youth to enter the workforce, and the formal sector in particular. In the case of India, the effects of the overall EPL stance are rigid enough that the marginal state-level reforms may only have moderating effects on the labour market as a whole. Thus, while we cannot identify a direct relationship between the labour reform index and the size of the informal sector in given states, we do observe that higher rates of organized sector job creation stem the rise in the size of the unorganized sector. Again, no state is able to do better than to stabilise the size of its formal sector.
Of course, many other factors also affect the incidence of informal employment, from competitive conditions and product market regulations to the extent of illiteracy (see Dougherty and Herd, 2008).
What can feasibly be done to improve India’s labour markets?
The priority should be to reduce the cost of formal employment
Labour reforms are urgently needed in India to reduce the relatively high cost of formal employment.
Reforms could serve to limit the detrimental effects that high costs have on the dynamism of the labour market, as well as improve the wage share and enhance economy-wide productivity. The above analysis indicates that India’s labour laws create a degree of employment protection (EPL) that is highly uneven, and among large firms and for collective dismissals it is higher than that of any OECD member or major emerging market. The negative consequences of this high level of protection are in line with the experiences of many OECD member countries. Considerable benefits for all workers could be had from a re-design and moderation of the current stance of India’s EPL. This would allow business to become more competitive though greater economies of scale, which would not only improve the welfare of workers through higher productivity and wages thereby better benefiting even those segments of the workforce that the laws were designed to protect, as well as the vast (low productivity) workforce in the unorganized and informal sectors that they virtually ignore.20
… through broad-based labour reforms
Our analysis of state-level labour reforms suggests that it is not the IDA alone that is harming labour market outcomes – it is the wider range of labour legislation.21 This result is consistent with the views of labour experts in India that cite the enormous complexity and uncertainty caused by the manifold overlapping laws and antiquated (often colonial-era) provisions, that are in dire need of simplification (Debroy, 2005; ADB, 2005; Sharma, 2006; Anant et al., 2006; TeamLease, 2006; World Bank, 2006).
As a general principle, a country like India with a large informal economy should have less, not more, labour regulation than a typical OECD country. Stringent labour regulations exacerbate labour segmentation, and the negative consequences are very clear in the case of India. While a long-term goal should be to formalise all jobs, labour laws should not seek to impose conditions too different from what employers and workers would agree upon voluntarily.
Major efforts have been made over the years in an attempt to simplify the laws, but, they have met with little success. Most notable is the Second National Labour Commission’s (SNLC) report in 2002, whose recommendations were widely vetted, highly detailed and for the most part quite moderate and well-regarded; yet few of its recommendations have been implemented in the past five years. While this is not entirely surprising as labour reform is perhaps the most intractable area to reform in any country, it raises the question of whether there are reform
strategies which could be more effective.
Strategies to make EPL more well-balanced should focus on improving fairness
A critical challenge for Indian labour markets is to ensure that there is greater fairness for all types of workers – whether in the organized or the unorganized sector, and regardless of what types of establishments they work in or what type of contract they are employed under. At the same time it is critical that labour policies support the healthy functioning of markets so that overall welfare is not hurt.
Current policies work against large classes of workers, from females, whose unemployment rates of 9% (for urban females) are double the more moderate rates for males, to scheduled castes and tribes (Borooah, 2005) and even out-of-state migrant workers.22 The bulk of these under-represented groups earn below the minimum wage,23 and many of them are faced with the unattractive prospect of casual employment at a low wage, with neither social protections nor stability.24 In part these limited options are a result of low levels of education and skills, but they are accentuated by the strength of employment protection, a problem that is observed for women and youth in many OECD countries.
A broader and more even-handed approach for India’s labour laws and regulations is needed, in keeping with the recommendations of the recent re-assessment of the OECD’s Jobs Strategy (OECD, 2006). This could address not only the cost of employment, but the unbalanced regulatory impact of current law. One promising approach would be to introduce a comprehensive new labour law (or set of only a few laws) that offers basic (moderate) protections to all types of workers and would not discriminate by size of enterprise. This law could improve statutory compensation for redundancy, provide for retraining and assistance in job search, thereby providing a balance between flexibility and security. If this law were to supersede the patchwork of existing laws and offer greater certainty and reasonable levels of compensation for dismissals, the cost of employment could perhaps be brought down enough to reverse the fall of the wage share and stem the ongoing expansion of the size of the informal sector. Efforts have been underway by the government in consultation with unions and employers to draft legislation that would offer some types of basic protections for the informal sector, including social insurance. If such vital consensus-based efforts are combined with a broad-based rationalisation of the EPL framework, considerable gains could be achieved for all workers.
Practical strategies are already being pursued and may be expanded, even if they are second-best
Given the slow progress in implementing the SNLC’s reasonable recommendations since 2002, second-best strategies for carrying forward labour reform must and are being considered. These reforms are necessarily incremental, but may offer ways of demonstrating the benefits of better-crafted regulations, and thus pave the way for more comprehensive consensus-based reforms in the future:
• Further reforms at the state level:
Reforms at the level of individual states are already occurring, as is documented in the state-level reform index. Given that the average value of the (count) index is only 21 out of 50, and the most reform-minded state only has a 28, there are many areas in which procedural or rule changes could be made at the state level to ease the burden of these labour regulations that raise the effective cost of labour. For contract labour, this is the route that has already been followed, with a considerable expansion of these arrangements in many states, although there remains ambiguity about when it can be utilised. In several other areas of labour regulation, there is even more that could be done, such as in rationalising inspections and streamlining union rules, which, for instance, still allow for a multitude of unions in an enterprise without clear guidelines for worker representation. While many of these reforms can be done without substantial central government involvement, states may feel restrained in what they are willing to carry out by the centre’s policies. Moreover, the potential benefits of state-level reforms under the current framework may not be enough to actually reverse the ongoing decline in the labour share (or the shrinking of the formal sector), even if they may be able to stabilize it. Greater autonomy for individual states to make more significant legal changes to labour laws may be necessary, although this could require a constitutional amendment making labour a state issue rather than a concurrent one (see TeamLease, 2006). In the meantime, though, there would seem to be considerable scope for reform under the present governance arrangements.
• Expansion of contract labour and fixed term contracts:
The ongoing expansion of contract labour (and potentially fixed-term contracts) is having a major impact on India’s effective EPL stance, and contract labour’s growth in popularity may serve to encourage its even more widespread use (see Anant et al., 2006). With last year’s clarification by the Supreme Court of the legal status of contract labour as distinct from regular employment, it is likely to be used even more widely, further reducing the share of the workforce facing the highest level of EPL. While this direction of reform is generally a move in a useful direction, the widening of the extent of differences in the degree of flexibility between different types of employees (workers and contractors especially) creates distortions that segment the labour force. A broader range of reforms would seem to be needed, initially giving firms clear authority to determine what types of work are suitable for contract labour. Even less distortionary would be greater use of fixed-term contracts for workers, which are allowed under revisions to central standing orders from 2003. Although half of states have issued similar enabling provisions, interviews suggest that businesses have only made limited use of these types of fixed-term contracts.
• Provide labour reforms in demonstration areas:
The development of export processing zones (EPZs) and special economic zones (SEZs) has created an enabling environment for local experimentation and reform of labour laws in many states. In fact, three-quarters of states have granted general procedural exemptions for use of contract labour in EPZs and SEZs, although only four have made formal amendments to their legislation for such purposes, according to information provided by the Ministry of Law & Justice. To the extent that SEZs provide a demonstration effect on employment illustrating the benefits of less costly regulation, they may be useful in promoting more general reforms, despite the second-best nature of such partial reforms in relatively small jurisdictions. Already, the information technology and business processing outsourcing sectors (IT/BPO) have enjoyed such exemptions in virtually all states, and the benefits to this broad sector have been visible in its tremendous growth, although they remain a very small share of the labour force.
• Enhance competition in product markets:
Recent examinations of OECD countries’ experience with the political economy of structural reforms demonstrates a strong link between the extent of product market deregulation and the progress of trade liberalization and labour market reforms (Høj et al., 2006). Such experience suggests that besides reducing tariffs and easing FDI restrictions, facilitating business entry and exit, lowering administrative burdens, and further progress on privatization would be useful in creating a more competitive operating environment that could make labour market reforms more politically feasible. In part, this feasibility reflects the fact that uncompetitive markets with high margin businesses make it easier for workers and unions to appropriate and sustain high economic “rents”, which are present to a significant degree in India (see Dougherty et al., 2008, Table 4). More competitive product markets and associated improvements in labour market conditions, on the other hand, can make reforms easier to undertake (Bassanini and Duval, 2006). In fact, the recent upswing in Indian growth, as well as the strong employment growth highlighted earlier, would suggest that now may be as propitious a time as any to make some serious attempts at labour reforms.
Policy recommendations for labour market reform
• Reduce barriers to formal employment: redesign and moderate the framework for employment protection, making it neutral across different types of firms and employees. This would at least include substantial revisions to the Industrial Disputes Act, removing the most restrictive provisions that require prior government permission for employment termination and exit decisions. As an interim step, the threshold for government permission could be raised to 300 employees and then phased out over a several year period.
At the same time, compensation for retrenchment could be raised from 15 days per year of service to closer to 45 days, coming closer to actual negotiated levels, as well as norms in the OECD.
• Revamp, consolidate and simplify all labour laws: reduce the number of labour laws to just a few simplified laws that provide basic legal coverage and protections for all employees, increasing fairness. Beyond this, decisions about how labour is utilized within the enterprise should be left to the managers of the firm.
Provisions that interfere with such decisions should be removed (i.e. Section 9A of the IDA).
• Improve labour market data: move toward a quarterly small-sample labour force survey so that employment outcomes can be better assessed between large-scale National Sample Survey rounds.
• Further state-level reforms: states should accelerate their own labour reforms in all areas, including offering special treatment for Special Economic Zones, which will not only attract investment but also provide a laboratory for demonstrating the benefits of reform. Broadening the allowances for the use of contract labour and fixed-term contracts could also effectively reduce the overall level of EPL. While some central flexibility exists at present for states to make labour reforms, they would be aided through a constitutional amendment to shift the jurisdiction of labour regulation from a concurrent central-state to just a state issue. In the absence of such a provision, greater clarity from the centre for states to make amendments in specific areas should be granted (i.e., for union recognition in a multi-union situation).
• Lay the groundwork for future reform: Further deepening of product, service and financial market reforms could improve competition and facilitate future labour reforms.
Source : Organisation de Coopération et de Développement Économiques
ECONOMICS DEPARTMENT
LABOUR REGULATION AND EMPLOYMENT DYNAMICS AT THE STATE LEVEL IN INDIA
ECONOMICS DEPARTMENT WORKING PAPERS No. 624
By Sean M. Dougherty
Résumé :
Au cours de la dernière décennie, les résultats du marché du travail se sont améliorés en Inde, l’emploi net augmentant de façon sensible dans l’ensemble de l’économie. Cependant, ces gains sont intervenus essentiellement dans les secteurs non organisé et informel de l’économie, où la productivité et les salaires sont généralement bien moindres que dans le secteur organisé formel. Seul le secteur organisé est assujetti à la réglementation du marché du travail et, dans ce secteur, l’emploi a diminué. L’incidence de la législation de protection de l’emploi sur la performance du marché du travail est sujette à controverses aussi bien dans la zone de l’OCDE qu’en Inde. Le présent article examine l’impact de cette législation et des réglementations associées sur la dynamique de l’emploi dans le secteur organisé de l’économie, en utilisant de nouvelles mesures de la réglementation nationale et des réformes au niveau des Etats. Il ressort de cette étude que si les réformes ont permis d’assouplir quelque peu la rigidité du droit fondamental du travail, des mesures plus approfondies sont nécessaires pour remédier aux distorsions qui sont apparues.
REGULATION AND EMPLOYMENT DYNAMICS AT THE STATE LEVEL IN INDIA
Employment has expanded in the economy as a whole, but not in the organized sector Contrary to widespread concerns about so-called jobless growth, employment expanded in five out of the last six years for India’s economy as a whole, as well as for manufacturing, according to updated estimates based on the most recent (2004/05) round of the Government of India’s National Sample Survey (NSS).2 Such estimates imply that employment growth has improved considerably, doubling the 1.3% pace of growth during the 1990s over the first five years of the 2000s. The pace of employment growth of over 2½ per cent per annum so far in the 2000s also exceeds the 1.9% pace of growth during the 1980s. Such an outturn is well above the previous expectations of many observers and represents a major upward shift in India’s labour market performance (see Nagaraj, 2004; Anant et al., 2006; Government of India, 2006).
Gains in employment from 1998 to 2005 came about nearly equally from increases in secondary and tertiary employment (representing 19% and 25% of the labour force, respectively). About 30 million net jobs were added in each of the two sectors, representing an annualized growth rate of about 6.2% for the secondary sector and 4.7% for the tertiary sector. These growth rates represent a turnaround from the anaemic pace of employment growth for much of the 1990s, and even exceed the employment growth rates of the 1980s. Net employment gains in agriculture on the other hand slowed to ½ per cent per year, as more workers moved to off-farm employment. Perhaps what is most surprising – given that industrial production has increased less than total output – is that the pattern of employment in industry (and manufacturing) has followed that of the economy as a whole, with overall net employment growth of 6½ per cent per year over the 1997 to 2004 period.
Despite strong gains in employment across the economy in recent years, a dichotomy has emerged with net increases in employment occurring almost exclusively in the least productive, unorganized, and often informal part of the economy. Meanwhile, the organized sector’s employment level has actually been shrinking.3 In industry, from 1997 to 2004, the overall 6½ per cent annual growth of employment disguises an average net loss of employment in the organized sector of about 1% per year over the period, while unorganized manufacturing employment grew at a rate of about 8% per year. As a result of these developments, the relative share of the organized sector in industrial employment fell from 15–20% in 1981–98 to only 10% in 2004, a decline that was much more rapid than for the economy as a whole.
The slight contraction of organized sector employment disguises considerable underlying job dynamics In order to better understand what has happened to the fortunes of the organized sector, this paper examines gross flows of jobs in organized manufacturing and the potential influence of labour regulations on these dynamics. Our methodology for measurement of job flows is described in Box 1. The resulting job creation, destruction and turnover figures capture the annual dynamics of employment at the level of detailed industry branches from 1998 to 2004 and represents the first study of India to have compiled gross job flow statistics using unit-level microdata. These figures reveal that the slight contraction of net employment in the organized sector disguises considerable underlying job dynamics, with rates of job destruction often higher than those for creation.
Measuring job flows in India
In empirical work carried out for this paper, seven years of Annual Survey of Industries (ASI) microdata covering the entire organized manufacturing sector were analysed to create a longitudinal sample of employment dynamics at the detailed industrial branch level, country-wide and for individual Indian states. This analysis used the basic methodology of OECD (1996) and Davis et al. (1998) to compute job creation and destruction statistics, with adjustments for the level of aggregation following the approach of Singh (2006). The resulting figures capture changes in employment levels for detailed manufacturing industry branches between adjacent years from 1998/99 to 2003/04, relying upon plant-level microdata.
The preferred approach for computing such job creation and destruction statistics is to use longitudinal microdata, matched over time at the establishment level. This was impossible because the CSO does not disclose plant-level identifiers that can be matched across time. So instead, a pseudo-longitudinal panel was constructed that uses detailed (three and five-digit) industry branch codes based on each plant’s industry of operation. Such s should in principle capture all job reallocation across industries, although they are likely to underestimate the actual turnover of jobs within-industry, as plant-level flows cannot be directly observed. Therefore, they represent a probable lower bound on job creation and destruction rates, although measurement error from non-responding plants could raise measured job flow rates to the extent they are not fully captured in sampling weights (see Ahmad, 2006).
Although this means that the resulting figures are not comparable with other countries’ job flow statistics, they are internally consistent and thus can be analysed across time and sector.
Our estimates of job flows at the national level finds considerable employment dynamics at the five-digit, and even three-digit, levels of aggregation (Table 1). At close to 4% per year through 2004, the annual rate of job creation has eased slightly since 1985. At the same time, the rate of job destruction has picked up from about 4% per year to 5%, resulting in a deterioration of net employment growth from about +1% per year pre-1998 to -1% per year in the most recent period. This shift in employment outcomes has coincided with a slowing of job turnover and reallocation.
Using a more detailed five-digit industry branch level of aggregation, additional job dynamics are picked up in the analysis, implying a 14% rate of job creation and 15% rate of job destruction since 1998. Such rates are as high as plant-level job flows in the US and France, although they are lower than (firm-level) estimates for China, where there has been more job destruction (Deng and McGuckin, 2005).
Given this apparently high degree of jobs dynamics in the organized sector, one would expect that the regulatory stance of labour laws would be weak. However, India’s labour market regulations for the formal sector are much more stringent than those of most OECD countries and several major developing countries as well. A standardized evaluation of India’s employment protection legislation (EPL) using the OECD EPL methodology was carried out using legal information supplied by the Government of India. The OECD methodology, outlined in the 1999 edition of the OECD Employment Outlook, evaluates the stringency of a country’s labour regulations in three areas: those for regular (indefinite) contracts, for temporary or fixed-term contracts, and for collective dismissal. Following this approach, India’s federal labour laws were scored, as they stood in early 2007.
For regular contracts, India’s labour laws are stricter than those of all but two OECD countries, Portugal and the Czech Republic, This stringency comes in large part from the procedural difficulty of having to obtain prior government permission to lay-off just one worker for plants covered by the Industrial Disputes Act (IDA). Chapter V-B of this Act, which requires such approval for all factories with more than 100 workers, removes much of the distinction between the EPL for regular employment and for collective dismissals. If this provision were not in force, the EPL for regular contracts would fall to the OECD average, which happens to be nearly the same as China’s score. This is indeed the situation for most employees in India’s tertiary sector, yet the concern remains for the secondary sector.
Even then, though, dismissing workers in non-IDA plants on performance grounds is difficult, as standard notification and severance obligations tend to be quite high, and courts often order reinstatement rather than compensation for unfair dismissals.
The stance of India’s EPL regime is more relaxed only for temporary and fixed-term contracts, placing it just above the mean of OECD countries’ score for such contracts. The moderate score of India for such contracts reflects the main area of labour reform that has come about in recent years, allowing employees to work on temporary work agency contracts to carry out a range of “non-core” activities, a concept that is defined in various ways across states but is usually restricted to activities that are ancillary to production.4 Standard fixed-term contracts are allowed for white-collar workers as well as, in principle, for regular workers, following a federal notification in 2003, and about half the states have implemented analogous provisions, but public awareness of the state-level provisions is low and repeated attempts to
reverse the central notification may be deterring its use.
The indicator score runs from 0-6, representing the least to most restrictive recent shift in the temporary EPL stance was a recent Supreme Court decision (2006) that clarified the
minimal restrictions on renewal of temporary contracts (although severance payments proportional to job tenure are required). This decision represented an important shift from the prior possibility of contractors being converted to indefinite contracts which could then cause the restrictions on regular contracts and collective dismissals under Chapter V-B of the IDA to come into force.
The third measure of the economy-wide stance of EPL is for collective dismissals.
Here, India’s stance is especially onerous, more-so than any other comparison country. Nearly every aspect of this measure rates at the top end of the scale, from the definition of collective dismissal (one worker), to notification and delays involved, given requirements for obtaining prior government consent. Most of these rules are encompassed in the IDA legislation, and come into force for any plant with more than 100 workers, encouraging many plants to stay small and restrict their scale of operation.
As a whole, India’s federal EPL stance is highly restrictive for the organized sector that it covers, given its interference in a number of important respects with firms’ hiring and severance decisions.
Nevertheless, job flow statistics suggest that considerable dynamics exist despite the regulations. This seeming mystery could be a result of enterprises altering their behaviour to reduce the impact of the rules, or it could be that in practice EPL application or enforcement is less onerous than the regulations suggest – a possibility that is likely to vary considerably across states, given that constitutional responsibility for state labour regulation is shared between the centre and the states. Both of these possibilities are explored.
Job dynamics under heavy labour regulation are uneven and come at a great cost
While the degree of job dynamics appears to be considerable, it varies widely across different segments of the organized labour force. The unbalanced coverage of employment protection legislation (EPL) in India appears to be important in explaining the variation of job flows across segments, suggesting that EPL is distorting the decisions of employers and accentuating labour market duality:
• The EPL index for staff employed with fixed-term or temporary contracts – in India typically supplied by a labour contracting agency – showed a moderate degree of restrictiveness, similar to the average for OECD countries. Correspondingly, over the period 1998 to 2004, job creation amongst contractual staff consistently dominated job destruction, allowing the net employment of contractual staff to rise by 10% annually, while it declined by 2½ per cent a year for regular workers, for whom there was much less job creation (Table 2).
• For small firms employing 100 or fewer workers, the EPL index is also similar to that found in OECD countries and, over the same period, job creation rates were more than double job destruction rates for this group of firms, causing employment in these smaller firms to rise by more than 20% a year. Moreover, despite the absence of restrictive dismissal laws, the job destruction rate in small plants was actually less than that in larger ones.
• In contrast, in large firms with over 100 employees (where the EPL index is high), the job creation rate for (non-supervisory and non-contractual) regular workers was low and well below the job destruction rate, with the result that the net employment of such staff fell by more than 5% per year in the period 1998 to 2004. In contrast, job creation for contract staff in large firms was twice the rate for large firms’ regular workers, and despite a higher destruction rate as well, their net employment increased almost 4% per year.
The tendency for large firms to restrain job creation is associated with a marked substitution of capital for labour that is not seen in small firms. Over the 1998 to 2004 period, constant price fixed assets per employee in plants with more than 100 workers rose by 21%, while they fell by 14% in smaller firms, according to tabulations by OECD using the Annual Survey of Industries (Table 3). While the large plants were already 28% more capital intensive, this raised their capital intensity relative to small plants by more than half plants over the 1999 to 2004 period. While public plants also have higher rates of job creation than private plants, public plants’ net contribution to employment is highly negative, on the order of a 10% drop in employment each year. This result may appear to be strange given the presumably strict compliance of public firms with the IDA legislation and the usual motive of public firms to preserve employment.
How could turnover in public enterprises be so high? The main explanation for this extensive (public plant) turnover is that voluntary retirement schemes (VRS) have been accepted by workers and unions as a mutually agreeable mechanism for downsizing. VRS schemes became common after the government set up a National Renewal Fund in 1991 as part of the structural reform program, to fund restructuring of public sector enterprises; only later did the private sector begin to make use of VRS (Nagaraj, 2004).
The VRS scheme has made employment adjustment feasible; however, it comes at a great cost. Since it requires that settlements are collectively agreed upon, high levels of compensation are made to groups of workers. Typically, the reservation payoff to the most recalcitrant marginal worker determines the outcome of the negotiations. OECD interviews with employers and local governments suggest that such payoffs can easily reach levels of two to three months of compensation for each year of work.6 A financial analysis of public sector redundancy schemes suggests that total average payments to staff that left ranged between almost four to over six years’ pay (Morrison, 2006). The redundancy compensation component (as opposed to normal leaving benefits such as accrued leave and provident fund balances) averaged between 10 and 36 months, with central government enterprises tending to pay the most. Such a cost places India among the most expensive in the world in terms of cost of dismissal, and is far above the statutory compensation that was intended by the IDA act (15 days of compensation per year of service). So while employment flexibility is feasible, it comes with an increased cost of severance – and not just to the employer who would normally have to pay for the VRS, but to the worker as well, in part because some businesses avoid payment altogether, but also because the costs create behavioural distortions that have detrimental economic effects.
High labour costs in the formal sector increase capital intensity and the size of the informal sector
Such high potential compensation for dismissal and the associated uncertainty raises the cost of employment, inducing larger employers to substitute more capital for labour than would appear justifiable given the apparently low wages that prevail in India.8 This effect can be observed in the low – and falling – labour share in organized industry, which can be observed in all two-digit industries from the 1980s. The share of total labour costs in value added has fallen from 36% in the early 1990s to 31% in 1999/2000 and 29% in 2003/04, with the steepest fall among large enterprises.9 The share of wages (including bonus payments) in value added has fallen similarly, from 25.7% in 1998/99 to 23.5% in 2003/04.10 Average wages have risen at a slower pace than labour productivity, so that labour compensation relative to Germany has been stagnant since the early 1990s. Moreover, wage inequality has grown considerably, with most gains going to the top two deciles, while median wages have remained stagnant (see Narain, 2006).11 In addition to causing business to substitute capital for labour, high implicit costs of employment for large-scale firms have induced many entrepreneurs to start small and stay small. Businesses in the unorganized (and often informal) sector with fewer than 10 or 20 workers (depending on the law) are subject to very few labour regulations and can employ casual or contract labour freely, as can most information technology (IT) and business process outsourcing (BPO) companies. Even businesses in the organized sector with fewer than 50 workers face no legal restraints on collective dismissals, while those having 100 or fewer workers are not subject to the IDA requirement to obtain prior permission before dismissing workers (or shutting down). As a consequence, the size distribution of businesses in India is highly skewed toward small units, with visible peaks in the distribution of plant sizes at ten and 20 workers. Considerable anecdotal evidence illustrates the extent to which entrepreneurs deliberately disintegrate their business activities into numerous small-scale units (Mukerji, 2006).12 As discussed in Dougherty et al. (2008), such small scale production reduces the efficiency of the economy and limits the productivity and incomes of businesses and workers.
The direct impact of India’s labour regulations are subject to considerable debate
The suggestion was made earlier in this Survey that the local business environment, and stringent labour regulations in particular, may have a deterrent effect on firm entry and investment, and thus could reduce potential output growth and possibly job creation as well. However, the empirical evidence on the role of labour regulations in India has been subject to a vigorous debate.
A series of economy-wide studies have examined the role of two key amendments to the Industrial Disputes Act (IDA) affecting collective dismissals (see Anant et al., 2006; Bhattacharjea, 2006). The 1976 amendment mandated prior government permission for layoffs, and the 1982 amendment lowered the size threshold from establishments with more than 300 workers to those with more than 100. Roy (2002) examined accessions and separations of individuals from firms from the 1960s through 1992, and she found that the two amendments had a strong and adverse effect on labour market flexibility, a result consistent with the cross-country evidence. However, flexibility may have increased for other reasons, given the growing use of contract labour from the early 1980s. The amendments did not appear to have an impact on the slow speed of employment adjustment to shocks in India (5-6 years), a finding that may reflect the broader range of rigidities that exist in India: especially weaknesses in the insolvency framework and urban land laws (Anant and Goswami, 1995; Roy, 2004).
The variation that exists in India’s labour laws across its states has offered fertile ground for empirical exploration by labour economists from around the world. Several well-known studies have been carried out, the best-known being the study by Besley and Burgess (BB) (2002, 2004), who compiled an index of state-level amendments to the IDA from 1949 through 1992, and classified them as to whether they were pro-worker or pro-employer. Their analysis of the indicator suggests that pro-worker amendments resulted in lower output, investment, employment and productivity in formal manufacturing although output in the informal sector increased. Their results also suggest that the more restrictive labour amendments had a perverse effect on the poor as they tended to increase urban poverty rates.
The BB index and its interpretation have turned out to be highly controversial and have been utilized as well as critiqued in multiple papers. Studies such as Hasan et al. (2003), Topalova (2004), and Sanyal and Menon (2005) use modified versions of the index (through 1992) to classify states as pro-worker or pro-employer, and find several things: states with more pro-worker labour legislation have lower elasticity of demand for labour, higher rates of poverty, and less investment, in part through interactions with trade liberalisation and state-level institutions. Further work by Aghion et al. (2005) updated the index (through 1997) and, using the time series of the index, found higher rates of growth following industrial de-licensing in states with more pro-employer labour institutions. Ashan and Pagés (2007) carried out a full re-scoring of the index to measure transaction costs and found at least as detrimental an impact on output and employment as the original BB study. Ashan and Pagés also separated the index into several subcomponents and found that regulations that increase the cost of settling disputes appeared to be more costly for employment than the IDA’s restrictions on firing and closure. On the other hand, a separate study by Purfield (2006) that further updated the BB index (through 2002) did not find significant detrimental effects of more pro-employer labour institutions on state-level growth.
Despite the robustness of the link between the BB index and poor labour market outcomes across different states in most of these studies, there are reasons to think that the index could be improved due to the sparseness of underlying amendments and its focus on just one law. Anant et al. (2006) and Bhattacharjea (2006) point out that many of the formal state-level amendments to the IDA used in the index would appear to be of minimal significance since they are infrequent, sometimes pertain to obscure procedural matters, and are likely to be overshadowed by court decisions and informal changes in the application of the laws. An update of the BB index carried out as background to this Survey (through 2005), using the most recent edition of the original data source for the index (Malik, 2006), shows that there have been only eight amendments to the IDA in any Indian state since 1990, and they relate to only three states, of which only the 2004 amendments in Gujarat would appear to be of any consequence to labour market outcomes.
Several prominent observers of India’s labour markets now argue that regulations other than the IDA are likely to be more important and that the IDA may be something of a distraction in the debate over labour reforms, given the four dozen central laws and hundreds of state laws that govern labour issues (Nagaraj, 2004; Debroy, 2005; TeamLease, 2006). There have been far fewer studies that look more broadly at India’s labour laws in systematic empirical analysis, perhaps due to their complexity and overlapping coverage. Moreover, some of the most important changes appear to have occurred in interpretation of the laws by the Supreme Court and through changes in the way the laws are enforced, particularly obvious in the case of the use of contract labour, which has become increasingly widespread in many states (Anant et al., 2006). Such state-level reforms have been the subject of speculation, but there has been little systematic regulatory information collected beyond the IDA amendments in the BB index.14
State-level labour reforms
A new OECD index of state-level labour reforms in India may illuminate the way forward
In order to better understand how state-level reforms may have affected labour markets, the OECD developed a customized survey instrument to identify the areas in which Indian states have made specific discrete changes to the implementation and administration of labour laws. This state-level survey covered eight major labour legal areas, identifying 50 specific subjects of possible reform, many of which could be implemented by administrative procedure rather than through formal amendments to the laws. Since few of these changes have been systematically documented, the assistance of the All-India Association of Employers (AIOE) was used to survey 21 states, about half of which were visited in person by members of the survey team. Preliminary answers to the questionnaires were drafted in most states by a labour expert designated by the AIOE or the Federation of Indian Chambers of Commerce and Industry (FICCI) affiliate in the state capital. The answers to this draft questionnaire were then corrected through discussions with local union leaders, independent labour experts, employers and state labour commissioners. Finally, the completed questionnaires were submitted to each state’s labour secretary for their final review.15 The verified responses to the survey were compiled by the OECD into an index that reflects the extent to which procedural changes have reduced transaction costs, through limiting the scope of regulations, providing greater clarity in their application, or simplifying compliance procedures. Answers were then scored as “1” if they reduced transaction costs, “0” if they did not, and (for two questions) “2” for a further reduction, with a maximum score of 50. A list of the subjects and summary scores are shown in Annex 1.
The reforms covered in the index concern eight specific areas: the Industrial Disputes Act (IDA), Factories Act, State Shops and Commercial Establishments Acts, Contract Labour Act, the role of inspectors, the maintenance of registers, the filing of returns and union representation.
A breakdown of the average index score across all states is show, which illustrates that the largest number of reforms that were identified concern contract labour, an area where many reforms have been carried out (especially for IT/BPO sectors and for export processing and special economic zones), but also an area for which there were more questions on the survey than for other areas.
In order to adjust for the number of questions in each area, the number of cases where reform had occurred was taken as a proportion of the number of possible subjects for reform in a given reform area (Panel B, out of 100). This proportional index shows that, with the exception of reforming the role of inspectors and the rules concerning the role of unions, there was a fairly similar share of the reforms carried out across the different areas surveyed.
What is perhaps most notable about the indexes, though, is the very modest degree of reform that has actually been carried out since, even in the area of contract labour, there have been reforms in only about half of the subject areas. Moreover, no state has a raw score of more than 28 out of 50 of the possible reform subjects, and for seven of these, no more than two states made any reforms. The principal areas where there was little change were: key sections of the IDA (that concern collective layoffs); rules governing working hours; union recognition and reduction in the number of inspections.16 Nevertheless, looking across states, enormous variation in the extent of reform appears across all of the regulatory areas. Ranked lowest in terms of the number of reforms carried out are Chhattisgarh, Goa and Bihar; at the same time, Gujarat, Uttar Pradesh and Andhra Pradesh are ranked highest. There is no dominant pattern in terms of the areas where the highest-ranked states have reformed, as the reforms are quite broad-ranging. However, it is notable that the three lowest-ranked states have made virtually no reforms in the areas of inspections, simplification of registers or the filing of returns. It is also of some significance that all states have made at least some reforms in the areas of the IDA and Contract Labour Act, for which the variation is the lowest among the reform areas.
State-level labour reforms do seem to explain differences in flexibility
The extent of labour reform captured in the above cross-state indices does appear to be linked to the flexibility of the labour market. In order to measure flexibility, gross job flow statistics at the five-digit industry level are computed using the plant-level microdata for formal sector enterprises in each Indian state, separately, over the period 2000 to 2004. The resulting state-level job turnover rates are then compared with the state-level labour reform index to see if there is a relationship. As the trend-line illustrates, the correlation between the reform index and job turnover rates is very high (R2 = 0.37), suggesting that states that have made more reforms to their labour rules and regulations indeed obtain more flexibility (in terms of job dynamics) in their labour markets and vice-versa: fewer reforms significantly limit flexibility.
The increased job dynamics come from higher rates of both job creation and destruction, which tend to offset each other almost entirely in terms of net employment.
There is broad cross-country evidence to suggest that job security regulations raise the cost of employment and have a detrimental effect on the creative-destruction process that is vital to economic growth and technological progress (OECD, 2004 and 2006). Job security regulations limit job turnover in sectors that for technological or market structure reasons need more frequent adjustment of employment to be competitive. In the case of India, the results show that state-level labour reforms appear to increase job turnover (or flexibility), and a lack of such reforms limits turnover. Across Indian states in the 2000 to 2004 period, an increase in turnover rates can be shown to limit the fall in the labour share of value added that has been observed nationwide. Yet no state is able to do better than to stabilize their labour share. More broadly, increases in job turnover can help productivity and output growth (in part through better job matching), and therefore increase the pace at which incomes converge to those of more developed countries (see Caballero et al., 2004; Micco and Pagés, 2004; Haltiwanger et al., 2006).
The OECD job strategy re-assessment also found that a lack of employment flexibility accentuates labour market inequalities, as it makes it more difficult for vulnerable workers such as women, minorities and the youth to enter the workforce, and the formal sector in particular. In the case of India, the effects of the overall EPL stance are rigid enough that the marginal state-level reforms may only have moderating effects on the labour market as a whole. Thus, while we cannot identify a direct relationship between the labour reform index and the size of the informal sector in given states, we do observe that higher rates of organized sector job creation stem the rise in the size of the unorganized sector. Again, no state is able to do better than to stabilise the size of its formal sector.
Of course, many other factors also affect the incidence of informal employment, from competitive conditions and product market regulations to the extent of illiteracy (see Dougherty and Herd, 2008).
What can feasibly be done to improve India’s labour markets?
The priority should be to reduce the cost of formal employment
Labour reforms are urgently needed in India to reduce the relatively high cost of formal employment.
Reforms could serve to limit the detrimental effects that high costs have on the dynamism of the labour market, as well as improve the wage share and enhance economy-wide productivity. The above analysis indicates that India’s labour laws create a degree of employment protection (EPL) that is highly uneven, and among large firms and for collective dismissals it is higher than that of any OECD member or major emerging market. The negative consequences of this high level of protection are in line with the experiences of many OECD member countries. Considerable benefits for all workers could be had from a re-design and moderation of the current stance of India’s EPL. This would allow business to become more competitive though greater economies of scale, which would not only improve the welfare of workers through higher productivity and wages thereby better benefiting even those segments of the workforce that the laws were designed to protect, as well as the vast (low productivity) workforce in the unorganized and informal sectors that they virtually ignore.20
… through broad-based labour reforms
Our analysis of state-level labour reforms suggests that it is not the IDA alone that is harming labour market outcomes – it is the wider range of labour legislation.21 This result is consistent with the views of labour experts in India that cite the enormous complexity and uncertainty caused by the manifold overlapping laws and antiquated (often colonial-era) provisions, that are in dire need of simplification (Debroy, 2005; ADB, 2005; Sharma, 2006; Anant et al., 2006; TeamLease, 2006; World Bank, 2006).
As a general principle, a country like India with a large informal economy should have less, not more, labour regulation than a typical OECD country. Stringent labour regulations exacerbate labour segmentation, and the negative consequences are very clear in the case of India. While a long-term goal should be to formalise all jobs, labour laws should not seek to impose conditions too different from what employers and workers would agree upon voluntarily.
Major efforts have been made over the years in an attempt to simplify the laws, but, they have met with little success. Most notable is the Second National Labour Commission’s (SNLC) report in 2002, whose recommendations were widely vetted, highly detailed and for the most part quite moderate and well-regarded; yet few of its recommendations have been implemented in the past five years. While this is not entirely surprising as labour reform is perhaps the most intractable area to reform in any country, it raises the question of whether there are reform
strategies which could be more effective.
Strategies to make EPL more well-balanced should focus on improving fairness
A critical challenge for Indian labour markets is to ensure that there is greater fairness for all types of workers – whether in the organized or the unorganized sector, and regardless of what types of establishments they work in or what type of contract they are employed under. At the same time it is critical that labour policies support the healthy functioning of markets so that overall welfare is not hurt.
Current policies work against large classes of workers, from females, whose unemployment rates of 9% (for urban females) are double the more moderate rates for males, to scheduled castes and tribes (Borooah, 2005) and even out-of-state migrant workers.22 The bulk of these under-represented groups earn below the minimum wage,23 and many of them are faced with the unattractive prospect of casual employment at a low wage, with neither social protections nor stability.24 In part these limited options are a result of low levels of education and skills, but they are accentuated by the strength of employment protection, a problem that is observed for women and youth in many OECD countries.
A broader and more even-handed approach for India’s labour laws and regulations is needed, in keeping with the recommendations of the recent re-assessment of the OECD’s Jobs Strategy (OECD, 2006). This could address not only the cost of employment, but the unbalanced regulatory impact of current law. One promising approach would be to introduce a comprehensive new labour law (or set of only a few laws) that offers basic (moderate) protections to all types of workers and would not discriminate by size of enterprise. This law could improve statutory compensation for redundancy, provide for retraining and assistance in job search, thereby providing a balance between flexibility and security. If this law were to supersede the patchwork of existing laws and offer greater certainty and reasonable levels of compensation for dismissals, the cost of employment could perhaps be brought down enough to reverse the fall of the wage share and stem the ongoing expansion of the size of the informal sector. Efforts have been underway by the government in consultation with unions and employers to draft legislation that would offer some types of basic protections for the informal sector, including social insurance. If such vital consensus-based efforts are combined with a broad-based rationalisation of the EPL framework, considerable gains could be achieved for all workers.
Practical strategies are already being pursued and may be expanded, even if they are second-best
Given the slow progress in implementing the SNLC’s reasonable recommendations since 2002, second-best strategies for carrying forward labour reform must and are being considered. These reforms are necessarily incremental, but may offer ways of demonstrating the benefits of better-crafted regulations, and thus pave the way for more comprehensive consensus-based reforms in the future:
• Further reforms at the state level:
Reforms at the level of individual states are already occurring, as is documented in the state-level reform index. Given that the average value of the (count) index is only 21 out of 50, and the most reform-minded state only has a 28, there are many areas in which procedural or rule changes could be made at the state level to ease the burden of these labour regulations that raise the effective cost of labour. For contract labour, this is the route that has already been followed, with a considerable expansion of these arrangements in many states, although there remains ambiguity about when it can be utilised. In several other areas of labour regulation, there is even more that could be done, such as in rationalising inspections and streamlining union rules, which, for instance, still allow for a multitude of unions in an enterprise without clear guidelines for worker representation. While many of these reforms can be done without substantial central government involvement, states may feel restrained in what they are willing to carry out by the centre’s policies. Moreover, the potential benefits of state-level reforms under the current framework may not be enough to actually reverse the ongoing decline in the labour share (or the shrinking of the formal sector), even if they may be able to stabilize it. Greater autonomy for individual states to make more significant legal changes to labour laws may be necessary, although this could require a constitutional amendment making labour a state issue rather than a concurrent one (see TeamLease, 2006). In the meantime, though, there would seem to be considerable scope for reform under the present governance arrangements.
• Expansion of contract labour and fixed term contracts:
The ongoing expansion of contract labour (and potentially fixed-term contracts) is having a major impact on India’s effective EPL stance, and contract labour’s growth in popularity may serve to encourage its even more widespread use (see Anant et al., 2006). With last year’s clarification by the Supreme Court of the legal status of contract labour as distinct from regular employment, it is likely to be used even more widely, further reducing the share of the workforce facing the highest level of EPL. While this direction of reform is generally a move in a useful direction, the widening of the extent of differences in the degree of flexibility between different types of employees (workers and contractors especially) creates distortions that segment the labour force. A broader range of reforms would seem to be needed, initially giving firms clear authority to determine what types of work are suitable for contract labour. Even less distortionary would be greater use of fixed-term contracts for workers, which are allowed under revisions to central standing orders from 2003. Although half of states have issued similar enabling provisions, interviews suggest that businesses have only made limited use of these types of fixed-term contracts.
• Provide labour reforms in demonstration areas:
The development of export processing zones (EPZs) and special economic zones (SEZs) has created an enabling environment for local experimentation and reform of labour laws in many states. In fact, three-quarters of states have granted general procedural exemptions for use of contract labour in EPZs and SEZs, although only four have made formal amendments to their legislation for such purposes, according to information provided by the Ministry of Law & Justice. To the extent that SEZs provide a demonstration effect on employment illustrating the benefits of less costly regulation, they may be useful in promoting more general reforms, despite the second-best nature of such partial reforms in relatively small jurisdictions. Already, the information technology and business processing outsourcing sectors (IT/BPO) have enjoyed such exemptions in virtually all states, and the benefits to this broad sector have been visible in its tremendous growth, although they remain a very small share of the labour force.
• Enhance competition in product markets:
Recent examinations of OECD countries’ experience with the political economy of structural reforms demonstrates a strong link between the extent of product market deregulation and the progress of trade liberalization and labour market reforms (Høj et al., 2006). Such experience suggests that besides reducing tariffs and easing FDI restrictions, facilitating business entry and exit, lowering administrative burdens, and further progress on privatization would be useful in creating a more competitive operating environment that could make labour market reforms more politically feasible. In part, this feasibility reflects the fact that uncompetitive markets with high margin businesses make it easier for workers and unions to appropriate and sustain high economic “rents”, which are present to a significant degree in India (see Dougherty et al., 2008, Table 4). More competitive product markets and associated improvements in labour market conditions, on the other hand, can make reforms easier to undertake (Bassanini and Duval, 2006). In fact, the recent upswing in Indian growth, as well as the strong employment growth highlighted earlier, would suggest that now may be as propitious a time as any to make some serious attempts at labour reforms.
Policy recommendations for labour market reform
• Reduce barriers to formal employment: redesign and moderate the framework for employment protection, making it neutral across different types of firms and employees. This would at least include substantial revisions to the Industrial Disputes Act, removing the most restrictive provisions that require prior government permission for employment termination and exit decisions. As an interim step, the threshold for government permission could be raised to 300 employees and then phased out over a several year period.
At the same time, compensation for retrenchment could be raised from 15 days per year of service to closer to 45 days, coming closer to actual negotiated levels, as well as norms in the OECD.
• Revamp, consolidate and simplify all labour laws: reduce the number of labour laws to just a few simplified laws that provide basic legal coverage and protections for all employees, increasing fairness. Beyond this, decisions about how labour is utilized within the enterprise should be left to the managers of the firm.
Provisions that interfere with such decisions should be removed (i.e. Section 9A of the IDA).
• Improve labour market data: move toward a quarterly small-sample labour force survey so that employment outcomes can be better assessed between large-scale National Sample Survey rounds.
• Further state-level reforms: states should accelerate their own labour reforms in all areas, including offering special treatment for Special Economic Zones, which will not only attract investment but also provide a laboratory for demonstrating the benefits of reform. Broadening the allowances for the use of contract labour and fixed-term contracts could also effectively reduce the overall level of EPL. While some central flexibility exists at present for states to make labour reforms, they would be aided through a constitutional amendment to shift the jurisdiction of labour regulation from a concurrent central-state to just a state issue. In the absence of such a provision, greater clarity from the centre for states to make amendments in specific areas should be granted (i.e., for union recognition in a multi-union situation).
• Lay the groundwork for future reform: Further deepening of product, service and financial market reforms could improve competition and facilitate future labour reforms.
Source : Organisation de Coopération et de Développement Économiques
ECONOMICS DEPARTMENT
LABOUR REGULATION AND EMPLOYMENT DYNAMICS AT THE STATE LEVEL IN INDIA
ECONOMICS DEPARTMENT WORKING PAPERS No. 624
By Sean M. Dougherty
Aucun commentaire:
Enregistrer un commentaire