Professional Documents
Culture Documents
1 Introduction
The abysmal performance of Indian manufacturing sector for the last couple of
decades has dismayed the economists and policy makers who expected that it
(manufacturing sector) would serve as an engine for economic growth, and would
*Corresponding author: Irfan Ahmad Sofi, School of Humanities and Social Science, Indian
Institute of Technology Indore, M-Block, IET DAVV Campus Khandwa Road Indore, Madhya
Pradesh, Pin Code: 452017, India, Cell no: +918370038163, e-mail: sofiirfan.irfan@gmail.com
Pritee Sharma: School of Humanities and Social Science, Indian Institute of Technology Indore,
M-Block, IET DAVV Campus Khandwa Road Indore, Madhya Pradesh, Pin Code: 452017, India
productivity. Therefore, the question arises: why then employers prefer to choose
contractual labour rather than the regular labour?
Under this backdrop, this study mainly focuses on investigating whether
there is a link between employment protection legislation (EPL) and informalisation. Besides, we link contractual employment with labour bargaining power and
volatility as well.
The study is based on 3-digit level data on 28 industrial sectors across 13
major states of India for time period 19992000 to 20072008 t hree-dimensional
panel data. We exploit the state level variation in EPL and quantify it, using leximetrics approach, to capture its impact. Given a dearth of literature on the issue,
this study significantly adds to the existing body of empirical literature. Unlike
most of the existing studies in India and/or internationally, which could not
exploit the fixed effect strategy owing to the time-invariant nature of EPL indexes
available, we are able to provide the evidences, in fixed effect framework. The
results show that informal employment is positively linked with EPL. Further,
our results show that with the increase in workers bargaining power, employers
tend to substitute informal-employment for the formal. Finally, we also provide
the evidence showing a positive association between volatility of demand and
informal-employment. Our results are robust to endogeneity that we remove by
following the instrumental variable two stage least square (IV 2SLS) estimation.
3 Literature Review
The impact of Employment Protection Legislation (EPL) on industrial performance is documented well in the theoretical literature (see, e.g., Nickell 1986;
Bertola 1990; Hamermesh 1993; Hopenhayn and Rogerson 1993). Bertola (1990)
developed a dynamic partial-equilibrium model in which he shows how the firing
publishing
Printing and
spining, weaving
Other food
Special purpose
other textiles
General purpose
Plastic products
Electric
Rubber products
other fabric
Insulated wire
other chemical
Furniture
Basic precious
Casting of metals
basic chemicals
Refined
Electricity dist
Grain mill
Dairy products
Structural metal
ppp of
Beverages
Non metalic
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
costs imposed by EPL can distort the optimal employment policy of a firm. the
model shows that at times when market demand is subdued and the labour force
available with employers become unprofitable, the firing costs associated with
EPL has the potential to hinder the layoffs, resulting into redundancy in firms and
thus lead to inefficiencies. The model also shows that these firing costs also serve
as prospective costs, which are taken into consideration by employer while hiring
fresh workers. Therefore, during favorable economic conditions, the actual hiring
of worker may be lesser than the desired. In this way, the EPL is shown to have
the potential of creating inefficiencies and stifling employment generation. Likewise, Hopenhayn and Rogerson (1993) developed a general equilibrium model
based on US economy that accounts for entry and exit of firms. They presuppose
a context, wherein job creation and destruction of firms take place every period in
response to firm-specific shocks; and then explain that the average employment
in US would be positively linked with firing costs as the latter will result in lesser
firing. However, this model also shows that the firing costs would impinge upon
firm entry and job creation in newly created or existing firms. In their empirical analysis, they find that higher firing costs in US negatively affected average
employment. In his model, Kugler (2000) shows that job security regulations
would induce the firms with higher labour-turnover to operate in informal sector.
Also, the effect of EPL is also drawn from the insider/outsider literature (see, e.g.,
Lindbeck and Snower 1987).
The increasing trend in informal employment across the world predominantly
in developing countries like India drew attention from researchers into studying
the various aspects of the informal employment system [see, e.g., Autor (2003)
for America; Pierre and Scarpetta (2013) for cross-country level; Almeida and
Almeida and Carneiro (2009) for Brazil, Gimpelson, Kapelyushnikov, and Lukyanova (2010) for Russia; and Ronconi (2010) for Russia]. However, the literature
IDA. Along the similar methodological line, Roy (2004) finds in his analysis that
there were significant rigidities in employment adjustment. However, he notes
that the rigidities existed even before 19751976.
Besley and Burgess (2004) constructed an index, henceforth BB index, using
the amendments undertaken by state governments to Industrial Disputes Act
(IDA). The study classifies the amendments into three categories pro-worker,
pro-employer, and neutral and assigns the scores 1, 1, and 0, respectively.
The scoring/coding is based on reading all the state level amendments to the
Industrial Disputes Act of 1947 from Malik (1997). By cumulating the scores over
time, they construct a stringency measure of EPL. Using the index along with
control variables, they investigate the impact of EPL on industrial performance.
The study concludes that EPL has a negative impact on output, employment,
and investment. However, the study drew many criticisms from researchers on
grounds such as false interpretation of certain amendments, and faulty coding
and cumulation procedure (Bhattacharjea 2006). However, after making the
changes in the BB index as per Bhattacharjea (2006), many studies find the
results are still in line with Besley and Burgess (2004) see, e.g., Ahsan (2009);
Adhvaryu, Chari, and Sharma (2014) etc. Using OECD index, Dougherty, Roubles,
and Krishna (2013) carried out a firm-level analysis and find higher productivity in firms that are operating in flexible states. Similarly, Mitra and Ural (2008),
using BB index, finds a positive effect of industrial-de-licensing in flexible states
on labor productivity.
Our study is important because it contributes in the existing literature by overcoming the above-mentioned research gaps in the literate. Besides, we also study
the link between informalisation, and labour bargaining power and volatility.
4 M
easuring Employment Protection
Legislation(EPL)
To capture the impact of employment protection legislations on the incidence of
informal employment, we exploit the state-level variation in these legislations to
categorize the states into rigid, flexible, and neutral. Using Besley and Burgess
(2004) database on labour laws, OECD index (2007), and Bhattacharjea (2006)
we derive the quantitative index capturing the state-wise stringency of EPL, using
majority principle. We designate a state as rigid and assign it score 1 if majority of the three sources (or measures) treat the given state as rigid. For example,
under BB index and OECD index (2007), West Bengal is picked up as a state with
flexible labour market. Therefore, we assign score 1 to it. Likewise, under BB and
BB Index
Battacharjea
OECD Index
Andhra Pradesh
Gujarat
Haryana
Karnataka
Kerala
Madhya Pradesh
Maharashtra
Orissa
Punjab
Rajasthan
Tamil Nadu
Utter Pradesh
West Bengal
Flexible
0*
0
Flexible
Flexible
0*
Rigid
Rigid
0
Flexible
Flexible
0
Rigid
0
0
0
Flexible
0
0
Rigid
Rigid
0
0
Flexible
Flexible
0
Flexible
Flexible
Flexible
0
Rigid
0
Rigid
0
0
Flexible
0
Flexible
Rigid
EPLI
1
0
0
1
0
0
1
1
0
1
1
1
1
5 Empirical Strategy
5.1 Econometric Model: Three-Dimensional Panel Regression
To investigate under fixed effect model1 whether employment protection legislation (EPL) and bargaining power constitute disincentives to employment of formal
workers, we first identify a relevant time-variant variable with which the EPL
index (EPLI) can be interacted to make it compatible with the fixed effect model.2
The time variant variable must, in this context, be strategically related with the
share of contractual worker, preferably negatively. For this purpose, we draw on
the empirical evidence in Sapkal (2014) which shows a decreasing impact of fixed
capital a proxy for technology on the share of contract workers, suggesting
that the industries using more technology prefer less contractual workers and
more regular worker. To confirm this, we run a regression of the following form:
Where, (CW/TW)ist is the log ratio of contractual worker to total number of workers
in industry i, state s and year t the total number of workers is equal to the
sum of contractual and regular workers. K/List is the log ratio of total fixed capital
to total labour force proxy for technology. The coefficient of our interest (1)
captures the impact of technology on the relative share of contract workers. Xkk
is a vector of state-specific and industry-specific control variables. i, i and t are
the industry, state and year fixed-effects, respectively. The results in column 1,
Table 3, show that the coefficient on K/L is negative and highly significant, implying that the higher use of technology discourage the employment of contractual
workers and encourage that of regular workers.
Now, in order to estimate with fixed effect model the impact of EPL on relative
share of contract workers, we interact the EPLI with K/L, as follows:
( CW / TW )ist = o + 1 K / List + 2 K / List EPLI s + 3 S / Lst + Xk k + i + s + t + ist(2)
1The importance of using fixed effect model arises from the fact that under the Hausman test,
the null hypothesis of zero correlation between error term and explanatory variables is rejected.
In other words, the hypothesis of coefficient estimates of random of random effect model and
fixed effect model are equal to one another is rejected, which suggests that random effect estimator is inconsistent. Besides, as our data is three-dimensional panel involving cross-industry
as well as cross-state variation, therefore it is important to include fixed effects in the model.
2Since the EPLI is a time-invariant variable, it cannot be included separately in the fixed effect
model. We must interact it with a relevant time-variant variable, so as to include it in the model.
( S / L ) FSLS st = 0 + Zn n + st n 1, 2, , n(3)
2 SLS
ist
= o + n Znist + kk Xkist + vist (5)
Under the Null hypothesis, H0 : Cov(znist, ist)=0, N*R2X2 with L-K degrees of
freedom. Where, L is the number of instruments and K is the number of endogenous right hand side variables in the original equation. If the null-hypothesis (H0)
is not rejected based on the Sargan statistics, then we may consider the instruments as valid.
To identify the relevant instruments for endogenous ratio of strikes to
lockouts, we draw on the information in Botero et al. (2004), which shows
that the left of the centre political parties across the world lead more stringent
pro-worker regulations and happen to be more inclined towards labour unions.
Likewise, in India too, the left of the centre political parties are claimed to be
pandering to labour unions, and are biased more in favour of workers than
employers (Aghion etal. 2008; Cali and Sen 2011). Therefore, we exploit the relative share of electoral seats occupied by various groups of political parties in
state legislature, between 19992000 and 20072008. Between this time-period,
there had been at least two rounds of state elections, causing a pronounced variation in the electoral seat share. For classifying the political parties according
to their fundamental ideologies left or right, we follow Aghion etal. (2008).
The political parties are classified into three categories soft left, hard Left,
6 Empirical Results
6.1 I mpact of EPL and Labour Bargaining Power
onInformalisation
The policy implications of our study are based on the 2SLS equation equation
4. But before estimating it, we first estimate OLS equation equation 2 to see
how the results turn up when the endogeneity concerns are not taken care off.
As discussed in Section 5.1, the existing literature finds that in capital-intensive
4Soft Left includes the Indian National Congress and the National Congress Party. Hard Left
includes the CPI and the CPI-M and other socialist parties. And, Right of the centre party includes the Bharatiya Janata Party.
5The data on contract workers is available at state-wise only since 1998.
Mean
Standard
Deviation
267,878
2675.85
8538.07
103,749
11,213.93
0.4917997
0.1538462
1,966,562
20,212.21
3.104701
641,706.2
5098.25
19,010.79
262,460.3
21,544.55
0.9979931
0.7693482
4,344,532
6843.574
0.1948673
Variation
By Industry, year and state
By Industry, year and state
By Industry, year and state
By Industry, year and state
By Industry, year and state
By Industry, year and state
By state and year
By state and year
By state and year
By state and year
industries using more technology, the employer tends to usher in regular workers.
Therefore, to use the EPLI in the fixed effect model, we interact it with K/L ratio,
which is a proxy for technology. As shown in column I, Table 3 the coefficient
on K/L ratio is negative and highly significant, confirming the evidence in the
existing literature mentioned in the Section 5.1. In column II, Table 3 we include
the interaction of EPLI and K/L ratio, and number of strike per lockouts. The
coefficient on the interaction is positive but not significant. We control for state
development expenditure per capita per million on economic services as it has a
bearing with state infrastructural facilities. Besides, we control for per capita net
state domestic product and real output, which determine external and internal
economies of scale, respectively (Mitra and Ural 2006).
Before coming to the 2SLS equation, we first discuss the FSLS estimates. The estimates are presented in Table 4. With highly significant coefficients, well in line with
our expectations drawn from the literature as stated in Section 5.3, the instruments
seem to be strong enough. As expected, the number of strikes per lockouts increases
with the increase in the share of electoral seats of Congress and the Hard left political parties, while Bharatiya Janata Partys seat share being negatively affecting the
number of strikes per lockouts. All these findings are well in line with the implications of the literature mentioned in the Section 5.3. Moreover, the instruments are
valid as suggested by the higher p-value against Sargan test shown in Table 5.
Now we come to the equation of our interest, i.e., equation 4, from which we
derive the policy implications. The estimates are presented in Table 5. In column
I, we are interested in examining the impact of instrumented ratio of strikes to
lockouts, and that of K/L ratio. The coefficient on strikes per lockouts is positive
and highly significant, suggesting that the informalisation is positively linked
with labour bargaining power. This finding is in line with the implications of
Constant
Ratio of strikes to lockouts (log)
Capital-labor ratio (log)
K-L ratio*EPLI
Control variables
Development exp. per capita pm
Per capita NSDP (log)
Real output (log)
Year effects
Industry effects
State effects
R2
No. of observations
(I)
OLS Results
(II)
OLS Result
5.075***
(1.743)
0.155***
(0.045)
5.123***
(1.744)
0.008
(0.025)
0.141***
(0.046)
0.055
(0.051)
0.021
(0.092)
1.041***
(0.393)
0.065*
(0.044)
Yes
Yes
Yes
0.697
2772
0.038
(0.093)
1.048***
(0.393)
0.065*
(0.044)
Yes
Yes
Yes
0.697
2772
theoretical model propounded by Sen (2013). The finding implies that employers
use informal employment to debilitate the labour bargaining power and to curb
the consequences of labour union movement, given that the informal workers are
incapable of forming or joining the union. The coefficient on K/L ratio, as shown
in column I, Table 5 is quite in line with that in OLS regression negative and significant. Now, in column II Table 5 we introduce the interaction between K/L ratio
and EPLI, so as to capture the impact of rigidity on informalisation. As shown,
the coefficient on the interaction is positive and significant suggesting that in
the rigid states, if technology increases, the substitution takes place in favour of
contractual employment. That is, the additional perspective costs, in the form
of firing costs associated with EPL, serves as a disincentive to employment of
regular workers. This finding supports the main proposition of the theoretical
Standard Errors
are Not Clustered
Standard Errors
are Clustered
R2
No. of observations
Year dummies
Industry-by-state dummies
State dummies
0.038
(0.043)
0.415***
(0.116)
0.471***
(0.056)
0.033
(0.100)
0.755
2772
Yes
Yes
Yes
0.038
(0.046)
0.415***
(0.213)
0.471***
(0.083)
0.033
(0.145)
0.755
2772
Yes
Yes
Yes
Constant
Dependent variable: Ratio of number of strikes to number of lockouts proxy for labour
bargaining power.
Figures in parenthesis represents robust standard errors, and (c) *=p<0.10, **=p<0.05,
***=p<0.01.
literature (mentioned in literature review section) that the EPL, for it imposes cost
on employers to employer regular workers, would give rise to informalisation. The
increase in the share of contractual worker when the employer uses more technology in rigid states as suggested by the positive and significant coefficient on
the interaction indicates that the employer, perhaps, hires skilled contractual
worker which may not necessarily be at par with the regular worker, just as to
evade the brunt of labour laws. The coefficients on the strikes per lockouts and on
the interaction term in 2SLS vary significantly from those in the OLS regression,
reason being the presence of endogeneity in the latter model.
Constant
Ratio of strikes to lockouts (log)
Capital-labor ratio (log)
K-L ratio*EPLI
Control variables
Development exp. per capita pm (log)
Per capita NSDP (log)
Real output (log)
Year effects
Industry effects
State effects
R2
No. of observations
Sagan test (p-value)
(I)
2SLS Resutls
(II)
2SLS Results
7.907***
(1.896)
0.283***
(0.092)
0.137***
(0.043)
7.921***
(1.895)
0.280***
(0.091)
0.113**
(0.045)
0.105**
(0.051)
0.018
(0.088)
1.682***
(0.428)
0.079**
(0.042)
Yes
Yes
Yes
0.681
2772
0.242
0.049
(0.089)
1.677***
(0.427)
0.078**
(0.042)
Yes
Yes
Yes
0.682
2772
0.230
Constant
High volatility
Ratio of strikes to lockouts (log)
Capital-labor ratio (log)
Development exp. per capita pm
Per capita NSDP (log)
Real output (log)
Year dummies
Industry dummies
State dummies
R2
No. of observations
(I)
Without Controls
(II)
With Controls
0.747***
(0.125)
0.419***
(0.174)
Yes
Yes
Yes
0.662
3276
8.683***
(1.983)
0.775***
(0.214)
0.283***
(0.092)
0.137***
(0.043)
0.018
(0.088)
1.682***
(0.428)
0.079*
(0.042)
Yes
Yes
Yes
0.681
2772
of EPL, which, analysts claim, does neither ameliorate the plight of the workers
nor let the business grow. Despite the substantial body of contract labor laws
strictly prohibiting contractual employment in core activities, the incidence of
contractual employment has registered an unprecedented growth over the last
two decades, cutting across non-core as well as core activities. Provoked by the
unfair exploitation against contract workers, the labour unions in India have been
strongly resisting the contractualisation, demanding job security for the workers.
Though alarming, the uncontrolled rise in informalisation has not received adequate attention of policy makers. What is more worrying is the fact that there
are evidences suggesting that the overdependence of informal employment may
impinge on efficiency and productivity growth of industries, and thus drive them
out of the competitive markets, in the long run.
In this study, we investigate the factors driving the use of contractual workers
in Indian manufacturing sector. Using three-dimensional panel data on 28 industrial sectors, between 19992000 and 20072008, across 13 major states of India,
References
Adhvaryu, A., A. V. Chari, and S. Sharma. 2014. Firing Costs and Flexibility: Evidence from
Firms Employment Responses to Shocks in India. Review of Economics and Statistics 95
(3): 725740.
Aghion, P., R. Burgess, S. Redding, and F. Zilibotti. 2008. The Unequal Effects of Liberalization:
Evidence from Dismantling the License Raj in India. American Economic Review 98 (4):
13971412.
Ahsan, A., and C. Pags. 2009. Are all Labor Regulations Equal? Evidence from Indian
Manufacturing. Journal of Comparative Economics 37: 6275.
Almeida, R., and P. Carneiro. 2009. Enforcement of Labor Regulation and Firm Size. Journal of
Comparative Economics 37 (1): 2846.
Autor, D. 2003. Outsourcing at will: Unjust Dismissal Doctrine and the Growth of Temporary
Help Employment. Journal of Labor Economics 21 (1): 142.
Bertola, G. 1990. Job Security, Employment and Wages. European Economic Review 34 (4):
851856.
Bertrand, M., E. Dufflo, and S. Mullainathan. 2004. How much Should We Trust Difference in
Differences Estimates? Quarterly Journal of Economics CXIS: 249275.
Besley, T., and R. Burgess. 2004. Can Labor Regulation Hinder Economic Performance?
Evidence from India. The Quarterly Journal of Economics 119 (1): 91134.
Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.