Employment and Earning Differentials Among Vocationally Trained Youth: Evidence from field studies in Punjab and Haryana in India

Abstract

This article examines the employability of vocationally trained youth and estimates their earning functions and existing wage differentials by gender and social groups in both Punjab and Haryana. Using primary data from two selected districts (one from each state) with a sample size of 914 pass-outs from 19 training institutes, we find that lack of demand and skill issues restrict vocationally trained youth to obtain quality jobs in these districts. Hence, most of them are either found in contractual jobs with a lower level of earning or remain unemployed from the time of completion of their training. Moreover, significant earning/wage differences are noted across gender and social groups. Female and Scheduled Caste (SC) workers are discriminated by getting paid lesser than their male and upper caste counterparts, despite their similar training and skill endowments. Hence, it is suggested that government intervention is necessary to lower the existing skill gap and to improve the quality of jobs for vocationally trained youth to reduce the rising unemployment problem.

Keywords

Vocational education employability wage differential decomposition

I. Introduction

Punjab and Haryana became relatively well-off states of India due to the partial success of the Green Revolution (see Dasgupta, 1977; Pal & Birthal, 1992; Ravallion & Datt, 1996; Singh & Kohli, 1997; Singh, 2000; Singh, 2001). Subsequently, the process of structural transformation began in these states with the growth of mechanization in agriculture (Mehrotra et al., 2014; Shetty, 2003). While the state economies of both Punjab and Haryana, in recent years, have moved towards industrialization with a growth in the share of their output and employment in industry, the rising open unemployment among vocationally (or industrial) trained youth in these states is a matter of great concern. The rising unemployment is not only likely to boost the process of out-migration (including illegal emigration) and domestic labour scarcity, but it also has implications on social issues such as creating pressure in terms of reservation of government jobs through violent means, increasing crime rates, rising incidence of illegal migration and drug addiction (see Advani, 2013; Parida & Raman, 2018; Saha, 2012; Smith, 2014). Hence, it is important to know why there is a high unemployment rate among vocationally trained youth in Punjab and Haryana.

Earlier studies on vocational education and employment in India, by and large, have focused on issues such as poor quality of vocational education (Agrawal 2012; Mehrotra, 2014; Mehrotra et al., 2015; World Bank, 2008), rising unemployment of trained youth (Ahmed, 2016; Mehrotra & Parida, 2017, 2019; Singh et al., 2020a; World Bank, 2008), poor quality of teachers and trainers (see Ajithkumar, 2017), lower returns to vocational training (Ahmed, 2016; Tilak, 2003) and high levels of skill mismatches (Hajela, 2012; Mehrotra et al., 2013; Mitra 2013; NSDC, 2013; Singh et al., 2020a; World Bank, 2008). However, there is a dearth of studies which have examined the employment patterns of vocationally trained youth in India and selected provincial states like Punjab and Haryana. Moreover, literature on the low level of earnings of vocationally trained workers and its likely consequences on their living conditions is very scarce. Therefore, the major objective of this study is to explore the employment patterns of vocationally trained youth, their earning structure and earning differentials by gender and social groups.

This article is organized into five major sections. Section II provides a brief review literature and highlights the research gaps, questions and contribution of this article. In Section III, data collection methods, variables and econometric methods are discussed. Section IV provides results and discussion. It has three subsections. While the first subsection explains the employment and unemployment scenario of vocationally trained youth, the second subsection provides their earning structure and explores the factors determining earnings of vocationally trained youth. The third subsection estimates the earning/wage differentials and decomposes it to understand its endowment effect and discrimination component across the earning distribution. Finally, Section V concludes the article along with the policy suggestions.

II. A Brief Review of Literature

The review of earlier studies on vocational education and employability across the countries of the world can be divided into two main categories. The first strand of literature argues in favour of vocational education and training due to its positive effects on employability, productivity and growth. However, the second strand of studies highlights higher unemployment rates and the skill mismatch issues. We have reviewed both theoretical and empirical studies from India and abroad.

Theoretical studies such as Nelson and Phelps (1966), Romer (1990), Mankiw et al. (1992), and Benhabib and Spiegel (1994) have argued that skilled workforce is the most crucial input to increase production and to sustain growth of output at the macro level, while at the micro level, human capital theorists such as Schultz (1960), Becker (1993), Mincer (1958, 1962, 1981), Lucas (1988), Azariadis and Drazen (1990), Glomm and Ravikumar (1992), and Benabou (1996) argued that improved levels of human capital investment would have positive impacts on employment, labour productivity and earnings for individuals.

The empirical studies conducted in various countries of the world (e.g., see Psacharopoulos and Chu [1994] in Latin America; Frost [1991], Middleton et al. [1993] from cross-country evidence; Neuman and Ziderman [1991] in Israel; Arriagada et al. [1992] in Brazil; Chung [1990] in Hong Kong; Bellew and Moock [1990] in Peru; Hinchliffe [1990] in Botswana; Ashton et al. [1999] in Singapore, South Korea, Taiwan and Hong Kong; Silverberg et al. [2004] in the USA, Hoeckel [2008] in OECD countries) have found that improved vocational and technical education has not only positive effect on both technological innovations and its adaptation, but it also helps to obtain better and qualitative jobs and a higher level of earnings.

However, studies such as Zymelman et al. (1976), Psacharopoulos (1987, 1994) and Tilak (1988), based on their reviews of studies from Malaysia, Barbados, China, Columbia, Jordan, Kenya and India, have concluded that in comparison to general education, the cost of vocational education is much higher, but its benefits are not much higher. Furthermore, studies such as Foster (1965) in Ghana and Blaug (1973) and Grubb (1985) in developing countries have found that unemployment among vocationally trained graduates is much higher.

Earlier study conducted in India such as the World Bank (2008), Hajela (2012), Agrawal (2012), Mehrotra et al. (2013), Mitra (2013), NSDC (2013), Ahmed (2016), Mehrotra (2014), Ajithkumar (2017), Pilz (2016), Mitra (2018) and Singh et al. (2020), by and large, have found a higher rate of unemployment among vocationally trained youth. According to Mehrotra (2014), Ajithkumar (2017) and Singh et al. (2020), this high unemployment is mainly due to the existing skill demand and supply mismatches. The existing skill gaps not only restrict many vocationally trained youth to participate in the labour market, but it also offers them a low and subsistence level of earning, which could be the main reason for the existing wage inequality among them (Singh et al., 2020; World Bank, 2007). In the case of Punjab, the study of IAMR (2011, 2013) and NSDC (2013) highlighted the problem of high unemployment among vocationally trained youth due to the same skill issues.

Although a few studies on vocational education (discussed above) have highlighted low level of earnings among vocationally trained workers, none of these studies have raised the issue of earning–wage inequality. On the other hand, existing earning–wage inequality studies in India, by and large, have focused on entire labour force and the issues such as labour market segmentation based on informality, government versus private jobs, occupations, etc. Existing studies on earning differentials by gender (Kingdon, 1998; Kingdon & Unni, 2001; Duraisamy & Duraisamy, 2005; Agrawal, 2014; Deshpande et al., 2015) and social group (Agrawal, 2014; Madheswaran & Attewell, 2007; Madheswaran & Singhari, 2016; Singhari & Madheswaran, 2017; Thorat et al., 2009) have not examined the case of vocationally trained youth categorically. Except Singh et al. (2020b), there is a dearth of studies that examine wage inequality among vocationally trained youth in India, including the states of Punjab and Haryana.

In this milieu, the article intends to study employment pattern of vocationally trained youth in Punjab and Haryana. It also measures the existing earning inequality among them and decomposes it to explore whether it is due to the difference in productivity (or endowment) or due to prevailing labour market discrimination. This article is going to answer the following research questions: (a) does vocational education and training help the youth to get appropriate jobs in Punjab and Haryana? (b) If not, then why? (c) Is the average earning/wage of vocationally trained workers equal or above the existing floor wage rates of the state government of Punjab and Haryana? Whether vocationally trained workers are paid equally as per their qualifications or there exists any kind of discrimination based on their gender or social group? Apart from its contribution to the employment and wage discrimination literature in the selected states, this study is privileged to have a large tracer survey and the use of advanced econometric tools for substantiating our arguments.

III. Data and Methodology

This study is based on primary data, which were collected using a structured interview schedule. This survey was conducted during the period from 29 June 2017 to 31 March 2018, using a convenient sampling method to include one district from each of these selected states, viz. Punjab and Haryana. Secondary data from Economic Census, Department of Industrial Training and Technical education, Government of Punjab and Haryana, are used to select the sample districts and the vocational training institutions (ITIs and Polytechnic institutions).

Selection of Sample Districts and Institutions

The districts with the highest number of establishments (factories) were chosen for the collection of primary data. The district Ludhiana from Punjab and Hisar from Haryana were selected. As per the sixth Economic Census, these districts ranked the topmost in their respective states in terms of the number of establishments (see Annexure I for the detailed list). While Ludhiana is known for cycle, hosiery and auto part manufacturing, Hisar is well known for steel, textile and automobile industries. Since the main objective of the article is to examine the employment and earning aspects of the vocationally trained youth, these districts were selected. The reasons for inclusion of these selected districts in the sample could also be clear from their profiles.

District Profiles

The district Ludhiana ranks the topmost in terms of number of establishments (2.18 lakhs) in Punjab. As per total population, it also ranks on the top with a population of about 3.5 million (as per Census, 2011). While about 47 per cent of the total population is female, about 26.4 per cent belong to the Scheduled Caste (SC) category. It is further noted that the youth (population in the age group 15–29 years) constitute about 30.6 per cent of the total population in the Ludhiana district. Out of the total of 1,320,000 workers (aged 15 years and above), a very large share (about 88 per cent as per PLFS, 2017–2018 data) depends on the non-agricultural sector. It is also interesting to note that a high share of the total workforce (about 61.3%) has either technical or formal vocational training. Despite a higher share of technically and vocationally qualified workforce, the open unemployment rate among vocationally and technically qualified youth is still very high in this district (17.4% as per PLFS, 2017–2018 data). Apart from the number of factory/establishment criterion, this profile also justifies the inclusion of this district in the sample.

Similarly, the study has included the Hisar district as it ranks the topmost in terms of the number of establishments (about 97,000). But this district ranks second in Haryana with a total population of about 1.75 million (as per Census, 2011). About 47 per cent of the total population are female, whereas about 23 per cent belong to SC category. The youth constitute 26.8 per cent of the total population in Hisar district. As per the PLFS (2017–2018), the total workforce (aged 15 years and above) in Hisar is about 3.8 lakh. The service sector is the major source of employment in this district as it stands at 48.5 per cent of the total employment, while the share of employment of the manufacturing sector is only about 6 per cent. On the basis of the skill profile of the workforce in this district, it is noted that 56.1 per cent of the total workforce has technical and formal vocational skills. Furthermore, out of the total technically and formally trained workforce, about 21.2 per cent are female and about 39.65 percent belong to SC category. But once again, the open unemployment rate among vocationally and technically trained youth is very high (19.8% as per PLFS, 2017-18 data) in this district. Once again, we would like to mention that apart from the number of factory/establishment criterion, this profile also justifies the inclusion of this district in our sample.

Selection of Institutions and Sample Size

Out of 85 vocational training institutions (37 from Ludhiana and 48 from Hisar), the survey has been conducted on only 19 institutes purposively (10 from Ludhiana and 9 from Hisar), which roughly constitutes about 22 per cent of the total available institutions (population) in the selected sample area. These institutions include both Industrial Training Institute (ITI) and Polytechnic colleges (see Annexure II for details). It was intended to cover at least two¹ institutes from each category.² First, it was planned to include only 8 institutions from each district (both 4 ITIs and 4 Polytechnic colleges) and to cover only 16 institutions (50% each from government and private institutions). As the female enrolment in co-educational vocational institutions was quite low, we included female training institutions to obtain a fair female sample size, which helped us to explore the employability and earning aspects of females in great detail. While two women’s vocational training institutions (one ITI and one polytechnic) from Ludhiana were covered, due to the unavailability of women’s polytechnic collages (see Annexure II) in Hisar, only one women’s training institution (one ITI college) was covered. As a result, the final sample size totalled 19 institutions. Moreover, it should be noted that this sample size is quite larger than the sample institutions covered by any of the existing studies that were reviewed in the previous section.

It should be noted that while the principals of the government institutions cooperated and provided the required information without any hesitation, some of the selected private institutions did not allow interviews to be conducted. Hence, other private institutions had to be approached as substitutes to the ones that did not allow the interviews.

From the 19 institutions, the contact information of 2,754 pass-outs were obtained from their placement registers. But, it should be noted that full and detailed information from only 914 pass-outs was available during the survey. This is again a very large sample in the case of tracer studies. The trained youth were asked to provide the socio-economic and demographic profile of their family, and their personal information, including current employment status, nature of employment, apprenticeship training, scholarship details, on-the-job training, total job market experience, earnings, levels of job satisfaction, occupational mobility, etc.

Ethical research practices were followed by not compelling respondents to answer questions that were uncomfortable.

Econometric Techniques

Both descriptive statistics (percentages and averages, including arithmetic mean and median) and advanced econometric tools have been used to analyse empirical results. While percentage and average values are used to explore the employability aspects, an augmented Mincerian earning equation (see Equation [1]) is used to find out the factors determining earning/wage of the vocationally trained workers:

\log (W_{i}) = X_{i} β + Z_{j} γ + δ M + θ S + μ L + ε

(1)

where log(W_i) is the natural logarithm of monthly wages/earning, X_i is a vector of individual-level factors, Z_j is a vector of occupational characteristics, M is the gender dummy (assumes the value of 1 for a female, and zero otherwise), S is social group dummy (assumes the value of 1 for socially backward castes like SC and ST, and 0 otherwise). All these variables are assumed to be exogenous. β, γ, δ and θ are vectors of constant parameters to be estimated, while ε is an unobserved stochastic disturbance term. First, the equation (see Table 1) is estimated. Furthermore, an additional regressor (L—lambda) has been used to overcome the problem of self-selection (due to Heckman, 1979).

Table 1.

Estimation of Earning Function (regression results)

Variables	OLS Results					Heckman Regression Results
Variables	Model 1 (overall)	Model 2 (male)	Model 3 (female)	Model 4 (SC)	Model 5 (non-SC)	Model 1 (overall)	Model 2 (male)	Model 3 (female)	Model 4 (SC)	Model 5 (non-SC)
Job experience	0.109*** (4.1)	0.078*** (3.32)	0.20*** (2.6)	0.038 (0.91)	0.142*** (3.84)	0.104*** (4.2)	0.077*** (3.5)	0.20*** (2.7)	0.037 (0.86)	0.138*** (4.4)
Experience square	0.006 (1.4)	0.002 (0.80)	0.002** (1.9)	0.001 (0.46)	0.009 (0.92)	0.005 (1.0)	0.003 (0.70)	0.002 (1.31)	0.002 (0.31)	0.008 (1.0)
Age	0.023** (2.1)	0.024** (1.95)	0.036** (1.9)	0.073*** (3.7)	0.002 (0.13)	0.020* (1.9)	0.019** (1.8)	0.040* (1.9)	0.071*** (3.7)	0.001 (0.05)
Higher caste dummy	0.057 (0.77)	0.038 (0.57)	0.147 (0.69)			0.076 (1.12)	0.043 (0.72)	0.14 (0.78)
OBCs dummy	0.050 (0.68)	0.126** (1.99)	–0.085 (–0.43)			0.048 (0.67)	0.113* (1.7)	-0.095 (-0.52)
Male dummy	0.598*** (6.3)			0.60*** (4.3)	0.576*** (4.3)	0.625*** (8.1)			0.649*** (5.2)	0.58*** (5.9)
Urban dummy	0.012 (0.21)	0.07 (1.25)	0.053 (0.30)	0.01 (0.10)	0.031 (0.4)	0.040 (0.74)	0.03 (0.53)	0.06 (0.34)	0.051 (0.5)	0.043 (0.60)
Polytechnic dummy	0.432*** (5.7)	0.289*** (4.52)	1.84*** (5.4)	0.303** (2.5)	0.495*** (5.1)	0.34*** (3.5)	0.183** (2.2)	1.96*** (4.6)	0.313** (1.9)	0.396*** (3.3)
Higher secondary dummy	0.012 (0.19)	0.006 (0.11)	0.075 (0.35)	–0.082 (–073)	0.04 (0.54)	0.033 (0.47)	0.019 (0.30)	0.014 (0.06)	–0.065 (–0.48)	0.062 (0.76)
Graduation and above dummy	0.236 (1.5)	–0155 (–1.12)	0.954*** (3.4)	–0.087 (–0.41)	0.40** (1.9)	0.21 (1.36)	–0.218 (–1.3)	0.97*** (2.8)	–0.185 (–0.72)	0.376** (1.9)
On the job training dummy	0.077 (1.09)	–0116* (–1.74)	0.65*** (3.4)	0.045 (0.38)	0.121 (1.4)	0.077 (1.17)	–0.107* (–1.8)	0.615*** (3.2)	0.032 (0.30)	0.118 (1.46)
Campus selection dummy	0.252*** (4.2)	0.306*** (5.3)	0.136 (0.61)	0.40*** (2.8)	0.209*** (3.2)	0.211** (2.7)	0.263*** (4.2)	0.22 (0.84)	0.377** (2.11)	0.19** (2.19)
Apprenticeship dummy	0.127* (1.6)	0.131** 2.02)	–0.78** (–2.4)	0.248** (1.9)	0.081 (0.82)	0.215** (2.2)	0.179** (2.5)	–1.11** (–2.4)	0.29* (1.8)	0.16 (1.42)
Changed job once dummy	0.04 (0.62)	0.065 (1.03)	–0.10 (–0.50)	–0.04 (–0.03)	0.027 (0.34)	0.066 (0.86)	0.077 (1.22)	–0.13 (–0.54)	–0.026 (–0.17)	0.06 (0.73)
Changed job twice dummy	0.033 (0.22)	0.02 (0.16)	0.30 (0.10)	0.05 (0.22)	0.08 (0.04)	0.01 (0.06)	0.01 (0.01)	0.21 (0.37)	0.03 (0.14)	–0.01 (0.01)
Migration dummy	0.918*** (4.3)	0.983*** (4.37)		1.35*** (10.2)	0.827*** (3.6)	1.037*** (3.5)	1.08*** (4.8)		1.28** (2.22)	1.012*** (2.9)
Constant	7.44*** (25.3)	8.33*** (29.7)	6.55*** (11.9)	6.52*** (13.8)	7.84*** (20.3)	7.636*** (30.1)	8.643*** (34.7)	6.40*** (11.4)	6.60*** (14.6)	8.021*** (26.7)
F-statistics	16.20***	10.60***	10.04***	8.13***	11.51***
R ²	0.420	0.276	0.464	0.477	0.403
Lambda (inverted Mills ratio)						–4.5***	–6.31***	1.35	–1.73*	–3.45***
Censured observation						472	359	113	146	326
Uncensored observation						442	343	99	130	312
Total observation	442	343	99	130	312	914	702	212	276	638

Source: Primary Survey 2017–2018.

Note: Z- and t-statistics are given in the parentheses. ***, ** and * imply statistical significance at 1%, 5% and 10% levels respectively.

Moreover, the Oaxaca–Blinder (OB) decomposition method (Oaxaca, 1973; Blinder, 1973) is used to estimate the wage differential between male and female, and between socially backward and forward castes. The simplest form of the OB decomposition is presented in the following equation:

\log (W_{A}) - \log (W_{B}) = \sum^{​} (X_{A} β_{A} - X_{B} β_{B}) + \sum^{​} (Z_{A} γ_{A} - Z_{B} γ_{B})

(2)

where alphabets A and B indicate groups A and B, respectively, for which wage differential is to be estimated. Equation (2) can further be rewritten to show the productivity and discrimination components of wage differentials distinctly. However, Reimer (1983), Cotton (1988), Neumark (1988) and Madheswaran and Attewell (2007) have stated that the OB decomposition method suffers from index number issues. That is, undervaluation of one of the groups (say A) will always come along with overvaluation of the other (say B). Hence, it is very difficult to conclude a priori which one of these two wage equations is the non-discrimination wage equation. Therefore, proper adjustment of weights is necessary, that is, if group A is paid as per the group B wage structure and vice versa, or they are paid based on an average rate, what would be the likely wage difference? The decomposition method (pooled) developed by Cotton (1988) and Neumark (1988) provides the possible answer.³ The estimated results (see Tables 5 and 6) and its discussion are given in the third subsection of Section IV.

Although OB decomposition results provide the average earning difference, it is not helpful for examining earning differences across the wage distribution. To explore the earning difference across the wage distribution, the quintile decomposition methods developed by Machado and Mata (2005) and Melly (2005, 2006) have been used. While Machado and Mata (2005) provide a simulation-based estimator where the counterfactual unconditional wage distribution is constructed by generating a random sample, Melly (2005, 2006) proposes estimating the unconditional distribution by integrating the conditional distribution over a range of covariates. The conditional quantile regression methodology proposed by Melly (2006) is very similar to the decomposition technique proposed by Machado and Mata (2005). But the major limitation of both these techniques is that it does not allow for computing detailed decompositions of the effect of each covariate on the unconditional quantile wage distribution. Chernozhukov et al. (2009) discuss a variety of methods based on conditional distributions that attempt to address this limitation.

The formal derivations of these three methods are based on the quantile regressions. The simple conditional quantile function can be expressed using a linear specification:

Q_{θ} (W | X) = X_{i} β_{θ} f o r e a c h θ ε (0, 1)

(3)

where W is the dependent variable denoting natural logarithm earnings/wages,⁴ X_i represents the set of covariates for each individual i and β_θ are the coefficient vectors to be estimated for the different θth quantiles. These quantile regression coefficients can be interpreted as the returns to different characteristics at given quantiles of the wage distribution. It is assumed that all quantiles of W, conditional on X, are linear in X. The conditional quantile function for each group can be expressed in two separate equations like Equation (3).

The conditional quantile function is not necessarily monotonic; hence, it might not be possible to invert it (Melly, 2005). In addressing this problem, Melly (2005, 2006) proposes integrating the entire conditional distribution function by integrating the full set of covariates.

θ = F_{W} (Q_{θ}) = E [F_{W | X} {Q_{θ} (W | X)}] = \int^{​} F_{W | X} {Q_{θ} (W | X)} d F_{X} (X)

(4)

where F_W(Q_θ) represents the conditional cumulative distribution of wages and the inverse of the distribution function, $F_{W}^{- 1} (θ)$ is ultimately the quantile function. First, the entire conditional distribution is estimated by quantile regression. Then, the unconditional distribution function is obtained by integrating the conditional distribution function over a range of covariates. Finally, by inverting the unconditional distribution function, the unconditional quantiles of group A or B can be obtained. To obtain the key counterfactual quantile of Group A, we need to invert the counterfactual distribution of group A characteristics with the wage structure of group B as:

F_{W_{A θ}}^{- 1} (W^{C}) = \int^{​} F_{W_{A}, θ | X_{A}} {(W | X)} d F_{X} (X)

(5)

Once the key counterfactual is estimated using either of these quantile techniques, the decomposition of wage gaps of the unconditional quantile function between groups A and B is estimated using Equation 6:

Δ_{θ} = (Q_{A, θ} - Q_{A, θ}^{C}) + (Q_{A, θ}^{C} - Q_{B, θ})

(6)

The first bracketed term represents the effect of characteristics (or the quantile endowment effects) and the second represents the effect of coefficients (or the quantile treatment effects). The estimated results (see Figure 3) and its discussion are presented in the third subsection of Section IV.

IV. Results Based on Descriptive Statistics

Employability of Vocationally Trained Youth

The employment status of the vocationally trained youth shows that (see Table 2) the workforce participation rate is quite low (only 54.5%), whereas the open unemployment rate among them is very high (19.2%). Out of the total sample, about 35 per cent are engaged as wage employees in the private sector, while about 13 per cent are self-employed, about 6 per cent are doing agriculture (an indication of employment distress) and only about 0.5 per cent are working in government/public sector enterprises as wage employed; but still about 45.5 per cent remain unemployed. This pattern of employment is noted for each category of trained (both the ITI and polytechnic) pass-outs and in both the sample states (see Table 2).

Table 2.
Employment Pattern of Vocationally Trained Youth

Employment Status Punjab
Haryana
Total

ITI Polytechnic Total ITI Polytechnic Total ITI Polytechnic Total

Government job 1.5 0.0 0.7 0.4 0.0 0.2 0.9 0.0 0.4

Private jobs 38.0 37.7 37.9 25.5 42.3 32.6 30.9 39.7 35.2

Self-employed (non-farm sector) 17.5 11.3 14.0 15.2 6.2 11.4 16.2 9.1 12.7

Unemployed (searching job) 15.5 13.2 14.2 27.4 19.6 24.1 22.3 16.0 19.2

Agriculture and allied (searching job) 8.5 6.2 7.2 6.8 3.1 5.3 7.6 4.9 6.2

Attending further training (higher) 4.3 10 9 8.4 9.9 7 7.3 9.9 8.8

Participating in general education 14.7 21.5 17 16.3 19 19.5 15 20.5 17.5

Total 100 100 100 100 100 100 100 100 100

Employment Status	Punjab	Haryana	Total
Government job	1.5	0.0	0.7	0.4	0.0	0.2	0.9	0.0	0.4
Private jobs	38.0	37.7	37.9	25.5	42.3	32.6	30.9	39.7	35.2
Self-employed (non-farm sector)	17.5	11.3	14.0	15.2	6.2	11.4	16.2	9.1	12.7
Unemployed (searching job)	15.5	13.2	14.2	27.4	19.6	24.1	22.3	16.0	19.2
Agriculture and allied (searching job)	8.5	6.2	7.2	6.8	3.1	5.3	7.6	4.9	6.2
Attending further training (higher)	4.3	10	9	8.4	9.9	7	7.3	9.9	8.8
Participating in general education	14.7	21.5	17	16.3	19	19.5	15	20.5	17.5
Total	100	100	100	100	100	100	100	100	100

Source: Primary Survey, 2017–2018.

Although the open unemployment among vocationally trained youth is high in both the sample states (Table 2), it is relatively higher in Haryana (24%) than in Punjab (14%). Moreover, the comparison of the type of vocational training reveals that the open unemployment among ITI pass-outs is higher (about 22%) than that of the polytechnic pass-outs (about 16%). This result is also substantiated by the overall state-level unemployment rates (see Figure 1), which is equally very high and has risen most recently—a very disturbing news, indeed.

Figure 1.

Sources: Authors plot based on NSS (2004–2005 and 2011–2012) and PLFS (2017–2018) data.

These results are very surprising as most recently (since the past decade), both the central and state governments have focused on the skill development of youth by increasing their investment on skill development and training heads. Hence, it can be argued that this higher unemployment rates among vocationally trained pass-outs is a sign of failure in the government’s policy. During the survey, it was observed that high unemployment is due to unavailability of appropriate number of jobs or skill issues. It is observed that the private vocational training institutions still use old and outdated training instruments although their syllabus has been revised most recently (since past 2–3 years) due to government’s intervention. Moreover, a large number of private institutions in both the states neither have industrial collaborations for placements nor apprenticeship training arrangements. Although government-owned training institutes provide apprenticeship training (due to existing apprenticeship norms), there is little attention given to job placements. This is really very painful from the trainees’ perspective.

It could be argued that the lack of employment opportunities in government sectors is also one of the major reasons for higher unemployment, rising participation in general education (a retrograde movement) and taking up distressed agricultural jobs among vocationally trained youth in both Punjab and Haryana (see Table 2). Out of the total sample of pass-out trainees, only about 0.5 per cent are found to be employed in government jobs (see Table 2). Hence, it is expected that an increase in the employment opportunities in government sectors is not only likely to reduce the volume of open unemployment, but it will also reduce the extent of distressed movement of trained youth to agriculture and general education substantially.

Furthermore, it should be noted that a number of vocationally trained pass-outs are eager to open their own business (as self-employed), but they are highly constrained by their family’s financial condition. Although the Indian government, most recently, has implemented a number of schemes to promote self-employment, the respondents do not find these schemes useful mainly because of the following reasons: (a) insufficient funds to initiate the business and (b) lack of working capital to sustain the business for at least 2 years. Hence, the efficacy of these schemes needs to be examined and evaluated for their better implementations and outcomes, as it would benefit a larger section of the society that are poor and socially marginalized like women and SCs.

Earning Structure and Its Determinants

The existing earning differences among vocationally trained youth by their gender and social groups have already been explored. The kernel density estimation (see Figure 2) depicts a clear picture of the earning difference between the male and female (Figure 2, panels A and B) and among the workers belonging to both ITI and Polytechnic institutions. While the earning distribution of the workers with ITI education, shows wage difference across the distribution; for the polytechnic pass-out workers, it is reflected by a relatively shorter tail (Glass Ceiling phenomenon) of the female earning distribution.

Similarly, we have noted significant earning difference across various social groups of the vocationally trained workers (Figure 2, panels C and D). In the case of workers trained from ITI, the earning difference is noted at the upper end of the earning distribution. Whereas in the case of workers with polytechnic degrees, it is noted throughout the distribution, but a relatively higher earning differences at the lower end. This earning difference does not only affect youth employability, it also has implications on poverty and the level of social inequality in both the sample states.

Figures 2.

Source: Primary Survey, 2017–2018.

To explore the pattern of earning difference further, the average (arithmetic mean) monthly earnings of male and female by their employment types, trades of training received and level of general education have been computed (see Table 3). It is noted that earning inequality between male and female vocationally trained workers is a bit higher in Haryana than in Punjab. While male workers are paid about 1.6 times higher than the female workers in Punjab, male workers are paid 2.2 times higher than their female counterparts in Haryana. The wage–earning difference is relatively more among ITI workers than that of their polytechnic counterparts holding degrees.

Table 3.

Average Monthly Earnings of Vocationally Trained Youth by Gender

	Groups	ITI			Polytechnic			Total
	Groups	Male	Female	M/F	Male	Female	M/F	Male	Female	M/F
State	Punjab	11720.1	6745.9	1.7	13919.7	11916.7	1.2	13033.0	8012.2	1.6
	Haryana	10665.2	4546.4	2.4	15334.3	13500.0	1.1	13306.9	5979.0	2.2
General education	Secondary	10795.2	4028.6	2.7	14667.7	13166.7	1.1	13878.2	8246.2	1.7
	Higher secondary	11357.8	5000.8	2.3	14402.2	12285.7	1.2	12873.4	6360.7	2.0
	Graduation and above	10500.0	9754.5	1.1	15250.0			12227.3	9754.5	1.3
Trade group	Mechanical	10240.8			13236.3	15000.0	0.9	11914.0	15000.0	0.8
	Electrical	13144.0	7125.0	1.8	13531.3	20000.0	0.7	13295.1	9700.0	1.4
	Textile	12516.7	4733.1	2.6	17830.0	10250.0	1.7	16801.6	4905.5	3.4
	Computer/IT	11071.4	9400.0	1.2	13237.5	11769.2	1.1	12675.9	10683.3	1.2
	Civil	10214.3			14942.9			11790.5
	Automobile	11726.2			11958.3			11850.0
	Food technology				16444.4			16444.4
	Pharmacy/health	16000.0	7600.0	2.1	13600.0	14166.7	1.0	14000.0	11540.0	1.2
Employment types	Government job	25500.0	15400.0	1.7				25500.0	15400.0	1.7
	Private job	10910.0	7639.5	1.4	13888.7	12550.0	1.1	12704.0	9332.8	1.4
	Self-employed	10980.0	3371.3	3.3	17007.3			14231.6	3371.3	4.2
	Total	11233.2	5576.6	2.0	14528.0	12550.0	1.2	13154.4	6985.4	1.8

Source: Primary Survey, 2017–2018.

There exists a glaring wage–earning inequality between male and female workers in government jobs, which is higher than the prevailing wage inequality in the private sector (see Table 3). It is noted that female workers are mostly engaged in low-paid occupations because of their types of vocational training, whereas male workers are found to be engaged in relatively higher positions. Hence, the existing earning inequality in the government sector is mainly due to the difference in skill endowment or because of the prevailing occupational hierarchy. A similar observation is made in the case of wage inequality in the private sector, where female workers are mostly engaged on contract basis with a relatively lower salary structure than their male counterparts. The earning difference between self-employed male and female workers is also because of the nature of business. While female self-employed workers are mostly found to be running ‘Boutique Shops’ and ‘Beauty Parlours’, male workers are working as ‘Motor mechanics’, ‘Electricians’ and ‘Owners of manufacturing workshops’.

Moreover, a significant wage–earning difference is noted across the social group categories. The workers belonging to the SC group earn less than both Other Backward Classes (OBCs) and Other Higher Caste workers on average. It is noted across the sample states and by the types of employment (viz., government, private and self-employment). While the higher caste workers earn 1.2 times higher than the SC workers on average In Punjab, the average wage of higher caste workers is 1.6 times higher than the workers belonging to SC. Similarly, OBC workers earn 1.3 times higher than the workers belonging to SC in Punjab, whereas they tend to earn about 1.2 times higher than the workers belonging to SC in Haryana.

While the low level of earning/wage in case of wage employment is an indication of the existing discrimination practices in the vocational labour market, the lower level of earning of self-employed SC workers reflects the poor quality of business in which they are engaged in. These workers have reported that they are unable to take up good quality business due to financial constraints.

But it is important to note that with an increase in the level of general education, there in an increase in the average level of earning/wage. It is observed in case of all the vocationally trained workers irrespective of their gender and social groups. Although this observation is consistent with the argument of Human Capital Theory, this is explored further.

Factors Determining Earnings of Vocationally Trained Youth

To provide comparative estimates of the earning function, both ordinary least square (OLS) and Heckman selectivity regression models have been estimated. Ten separate regression models (five OLS and five Hackman regression models) have been estimated for comparison (see Table 4). While the OLS models are based on the portion of sample for which we have positive earning/wage, the Heckman regression models make use of the entire sample with an additional regressor called the self-selection component, lambda (popularly known as the Inverted Mill Ratio). The statistically significant coefficients of lambda indicate that vocationally trained youth are self-selected into the labour market, and hence, the OLS regression models (ignoring self-selection bias) likely produce biased estimates. The Heckman regression models use a two-step procedure. First, it runs a probit model for estimating factors determining labour force participation. From this first-stage regression model, it computes the lambda and uses it as an additional regressor in the second-stage regression (OLS) to estimate the wage–earning equation. In this process, it overcomes the self-selection bias. Heckman command in stata has been used to obtain these results. Statistically significant coefficients are obtained with expected signs (as per the theory of human capital) for most of the important regressors (see Table 4).

Table 4.

Average Monthly Earnings of Vocationally Trained Youth by Social Group

		ITI					Polytechnic					Total
		Others	SC	OBC	OTR/SC	OBC/SC	Others	SC	OBC	OTR/SC	OBC/SC	Others	SC	OBC	OTR/SC	OBC/SC
State	Punjab	10010.2	8688.8	14737.5	1.2	1.7	14300.7	12980.0	13020.0	1.1	1.0	12548.8	10617.4	13906.5	1.2	1.3
	Haryana	10280.0	7306.7	7899.1	1.4	1.1	16614.0	13413.6	14120.0	1.2	1.1	14285.3	8945.1	10574.7	1.6	1.2
General education	Secondary	10370.0	8222.2	8577.8	1.3	1.0	15380.2	13156.3	13875.0	1.2	1.1	14516.4	11380.0	12430.3	1.3	1.1
	Higher secondary	9812.1	8276.2	9512.3	1.2	1.1	14792.3	13289.1	13777.4	1.1	1.0	12443.9	9929.9	11086.3	1.3	1.1
	Above	12450.0	6600.0	10442.9	1.9	1.6	30000.0	10333.3		2.9	0.0	14957.1	8000.0	10442.9	1.9	1.3
Trade groups	Mechanical	8154.5	9610.0	11650.0	0.9	1.2	13797.3	10029.2	14628.6	1.4	1.5	12504.2	9796.3	12777.0	1.3	1.3
Trade groups	Electronic	13958.8	10883.3	9083.3	1.3	0.8	13000.0	15000.0	14350.0	0.9	1.0	13708.7	11471.4	12375.0	1.2	1.1
	Textile	6006.3	6664.3	4969.6	0.9	0.7	19000.0	16981.3	14630.0	1.1	0.9	14050.0	9901.0	7897.0	1.4	0.8
	Computer/IT	13240.0	6900.0	11900.0	1.9	1.7	13486.1	12571.4	10875.0	1.1	0.9	13432.6	9546.7	11269.2	1.4	1.2
	Civil	10250.0	8000.0	12375.0	1.3	1.5	18750.0	12742.9	15000.0	1.5	1.2	12375.0	10213.3	13090.9	1.2	1.3
	Automobile	10063.6	11864.3	17500.0	0.9	1.5	12730.8	9928.6	13000.0	1.3	0.0	11508.3	10896.4	14928.6	1.1	1.4
	Food technology						18800.0		13500.0			18800.0		13500.0
	Health service	12250.0		6700.0			14000.0	13000.0	13750.0	1.1	1.1	13500.0	13000.0	11400.0	1.0	0.9
Employment types	Government job	16450.0		29500.0								16450.0		29500.0
Employment types	Private job	11042.0	9370.9	9693.4	1.2	1.0	14679.4	11989.5	13534.8	1.2	1.1	13380.4	10519.9	11797.0	1.3	1.1
	Self-employed	7386.4	5397.9	7831.0	1.4	1.5	17025.0	18900.0	15277.8	0.9	0.8	12415.2	8773.4	9594.7	1.4	1.1
	Total	10101.4	8163.9	9484.8	1.2	1.2	15173.3	13073.5	13820.0	1.2	1.1	13176.9	10090.0	11407.7	1.3	1.1

Sources: Primary Survey, 2017–2018.

The coefficient of job market experience is positive and statically significant in all the models. As expected, the coefficient of age is also positive and significant. These two coefficients reflect the fact that all other things remaining unchanged/constant, with increasing age and job experience, the earning/wage of the vocationally trained workers also increases. The coefficient of general and vocational education dummies are used to demonstrate the returns to skill endowments. The general education dummies have positive coefficients. This reflects the fact that with an increase in the level of general education, there is an increase in the earning levels. Moreover, the positive polytechnic dummy coefficients indicate that workers having polytechnic training, on average, earn more than those trained at ITI (the reference category). The impact of on the job training, attainment of apprenticeship programme and the campus selection dummies, all of them have also reflected positive and statistically significant coefficients. This result is similar to the findings of earlier studies such as Ryan (1998, 2001), Bonnal et al. (2002) and Parey (2016).

The impact of characteristics such as gender, social group and migration is also estimated using dummy variables. The positive coefficient of the male dummy reflects the fact that male workers, on average, earn more than their female counterparts, all other characteristics being constant. But the estimated value of the social group dummies indicates that earning differential between SC and OBC and between SC and other caste (normally general castes) workers are statistically insignificant. However, we have found that foreign migration experience has positive influence on the average earning levels of the vocationally trained youth.

The comparison of the regression coefficients across the models reveals that the coefficients of a few important endowments and characteristic variables are statistically insignificant. This is mainly due to the fact that even with the same levels of education and training, few workers, particularly women and pass-outs belonging to SC, are not able to get jobs. Many of them have reported that they could not acquire the required level of skills during their training because of their poor knowledge on basic sciences and English language skill (as the medium of instructions and learning).

Earning Differentials and Its Decomposition

To explore whether the existing earning–wage differential between male and female workers and between SC and non-SC workers is due to productivity (endowment) or discrimination, their wage differentials are decomposed using O–B decomposition method. The results using Reimer (1983), Cotton (1988) and Neumark (1988) methodology are presented for comparison (see Tables 5 and 6). First, the male–female wage differential results are explained, and then the caste-wise differentials are discussed (by SC and non-SC).

The twofold decomposition result suggests (see Table 5, panel A) that if female ITI-trained workers are paid according to their own wage structure (equation), that is, D = 0, then they are facing an 89% wage discrimination. This is a clear indication that female workers are not paid as per their skill endowments. However, if female workers are paid as per the male wage structure (i.e., D = 1), then they would face even more discrimination (about 94%). But, if the average wage structure prevails (i.e., D = 0.5) in the vocational labour market, then female workers would face relatively less discrimination (92.1%). However, if the wage structure is determined according to their sample proportion as weights in the wage equation, then the discrimination would increase slightly to 92.1 per cent. However, the pooled wage structure over both groups would decrease the wage discrimination significantly to 75.5 per cent, but it is still much higher.

Comparison of the wage decomposition results of the polytechnic degree holder reveals that the earning discrimination components are much less than that of the ITI-trained youth. The earning decomposition of female workers would be about 35.5 per cent (if D = 0), 37.7 per cent (if D = 1), 36.6 per cent (if D = 0.5), 37.5 per cent (as per Cotton, 1988) and 34.8 per cent (as per pooled wage structure), respectively (see Table 5, panel B). Hence, it could be stated that female vocationally trained workers earn less because of the prevailing wage discrimination against them in the vocational labour market.

Table 5.

Decomposition of Earning Differentials by Sex (Oaxaca–Blinder method)

Components of Wage Differential	Oaxaca (1973) D = 0	Oaxaca (1973) D = 1	Reimers (1983) D = 0.5	Cotton (1988) D = 0.776	Neumark (1988) Pooled
Panel A: Summary of decomposition results of ITI pass outs by gender
Mean prediction high (H)	9.153	Difference due to endowments (E)		0.103
Mean prediction low (L)	8.242	Difference due to coefficient (C)		0.857
Raw differential (R, H-W)	0.911	Difference due to interaction (CE)		–0.048
Unexplained difference	0.809	0.857	0.833	0.840	0.688
Explained difference	0.103	0.055	0.079	0.072	0.223
% of unexplained difference	88.7	94.0	91.4	92.1	75.5
% of explained difference	11.3	6.0	8.6	7.9	24.5
Panel B: Summary of decomposition results polytechnic ITI pass outs by gender
Mean prediction high (H)	9.489	Difference due to endowments (E)		0.075
Mean prediction low (L)	9.373	Difference due to coefficient (C)		0.044
Raw differential (R, H-W)	0.116	Difference due to interaction (CE)		–0.003
Unexplained difference	0.041	0.044	0.042	0.044	0.040
Explained difference	0.075	0.072	0.074	0.073	0.076
% of unexplained difference	35.5	37.7	36.6	37.5	34.8
% of explained difference	64.5	62.3	63.4	62.5	65.2

Sources: Primary Survey, 2017–2018.

These results are consistent with the findings of earlier studies like Madheswaran and Attewell (2007) and Deshpande et al. (2015), who have also found significant overall earning discrimination against women in India.

The caste-wise wage decomposition result for the ITI-trained workers shows that (see Table 6, panel A) if SC workers are paid according to their own wage structure (i.e., D = 0), then they face about 39 per cent wage discrimination. But, if the non-SC wage structure prevails (i.e., D = 1) in the vocational labour market, then workers belonging to SC would face relatively less discrimination (about 27%). If the average wage structure prevails (i.e., D = 0.5), then SC workers would face a bit more discrimination (about 33%). Moreover, if the wage structure is determined according to their sample proportion as weights in the wage equation, then the discrimination would decrease to 31.4 per cent. However, the pooled wage structure over both groups would reduce the wage discrimination to about 31.9 per cent. This is still very high.

Comparison of the wage decomposition results of the polytechnic degree holder reveals that the earning discrimination components of earning decomposition of a worker belonging to SC would be about 125 per cent (if D = 0), 70.9 per cent (if D = 1), 97.8 per cent (if D = 0.5), 83.4 per cent (as per Cotton, 1988) and 73.3 per cent (as per pooled wage structure), respectively (see Table 6, panel B). This result suggests that like females, SC vocationally trained workers also earn less because of the existing wage discrimination against them in the vocational labour market.

Table 6.

Decomposition of Earning Differentials by Social Group (Oaxaca–Blinder method)

Components of Wage Differential	Oaxaca (1973) D = 0	Oaxaca (1973) D = 1	Reimers (1983) D = 0.5	Cotton (1988) D = 0.776	Neumark (1988) Pooled
Panel A: Summary of decomposition results of ITI pass outs by social group
Mean prediction high	8.879	Difference due to endowments (E)		0.085
Mean prediction low	8.738	Difference due to coefficient (C)		0.038
Raw differential	0.141	Difference due to interaction (CE)		0.018
Unexplained difference	0.055	0.038	0.047	0.044	0.045
Explained difference	0.085	0.103	0.094	0.096	0.096
% of unexplained difference	39.4	27.0	33.2	31.4	31.9
% of explained difference	60.6	73.0	66.8	68.6	68.1
Summary of decomposition results of polytechnic pass outs by social group
Mean prediction high	9.513	Difference due to endowments (E)		–0.036
Mean prediction low	9.365	Difference due to coefficient (C)		0.105
Raw differential	0.148	Difference due to interaction (CE)		0.079
Unexplained difference	0.184	0.105	0.144	0.123	0.108
Explained difference	–0.036	0.043	0.003	0.025	0.039
% of unexplained difference	124.7	70.9	97.8	83.4	73.3
% of explained difference	–24.7	29.1	2.2	16.6	26.7

Source: Primary Survey, 2017–2018.

These results are again consistent with the findings of earlier studies like Madheswaran and Attewell (2007), Deshpande et al. (2015), Duraisamy and Duraisamy (2016), Madheswaran and Singhari (2016), Singhari and Madheswaran (2017), and Padhi et al. (2019), who have found that earning discrimination against workers belonging to the SC category is high in India.

Although the average wage differential reveals that a significant proportion of the earning differentials is due to discrimination component, it fails to explain whether it is due to the ‘sticky floor’ or ‘Glass Ceiling’ phenomena. To examine these phenomena, the quantile wage decomposition results are estimated in the next subsection.

Quantile Decomposition

The quantile decomposition results based on the Machodo–Mata–Melley (MMM) method for both sex and social groups are presented in Figure 3. It is observed that the overall gender–wage gap is highest in the lower quintiles, but it declines as we move up from lower to higher quintiles. This is a clear reflection of the phenomenon of ‘sticky floor’ wages of females (Figure 3, panels A and B). The decomposition (into the explained and unexplained components) of total wage gap shows that the explained component is quite lower across the quantiles but is low in terms of return to skill endowment. But a higher share of unexplained components indicates that gender–wage differential exists in the vocational labour market. This is mainly due to the practice of wage discrimination. But due to the sticky floor, women at the lower end of the wage distribution suffer the most.

However, from the quantile decomposition of wage differentials of workers belonging to SC and non-SC, it reveals that both the phenomena of both ‘sticky floor’ and ‘glass ceiling’ are present. The decomposition (into the explained and unexplained component) of total wage gap shows that the explained component is highest at lower deciles, and it declines as we move up from lower to upper quantiles, but at the top decile, it is again very high (Figure 3, panels C and D). The wage gap curve shows a ‘U-shaped’ pattern, while the return to skill endowment is low at both the bottom and top deciles. This result is slightly different from that Deshpande and Sharma (2013), which claims only sticky floor wage. But it is noticed that SC workers suffer from wage discrimination at both the lower and upper ends of the wage distribution.

Figures 3.

Source: Primary Survey, 2017–2018.

V. Concluding Remarks

To sum up, it can be stated that the workforce participation rates of vocationally trained youth in Punjab and Haryana are very low. Unavailability of government sector jobs along with the skill constraints are among the major distressing factors that determine the employability of vocationally trained youth. Moreover, the poor quality of training along with financial constraints, by and large, restrict a large number of vocationally trained youth to take up wage employment in the private sector or to begin their own business as self-employed. These aspects of employment distress are reflected by higher rates of unemployment, participation in general education (after completion of vocational training) and participation in agriculture and allied sector jobs.

Moreover, it is noted that due to the prevalence of ‘sticky floor’ in the case of female workers and both ‘sticky floor’ and ]glass ceiling’ in the case of workers belonging to SC, there exists a significant wage discrimination against female and SC workers in the vocational labour market. Based on these findings, it is suggested that reducing the labour market demand-supply gap through required job opportunities in industry is necessary. This could be done through pro-labour-intensive industrial growth policies. Particularly, promoting the growth of micro and small industries into medium scales would improve the employability of vocationally trained youth in Punjab and Haryana. On the supply side, improving the quality of pre-vocational training education (at primary and secondary level) is also indispensable as it is a major obstruction in the process of acquiring vocational training for a large number of rural youth. Introduction of vocational education at the school level through ‘The New Education Policy’ of the central government may partly be helpful in this context, but measures for the development ‘English language’ and other soft skills are equally important to improve employability of youth. Moreover, promotion of female enrolment in vocational training by introducing appropriate number of trades and ensuring proper implementation of ‘the minimum wage act’ for vocationally skilled workers would help in improving employability and reduce wage discrimination significantly.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflict of interests with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

District-wise Number of Establishment in Punjab and Haryana

Name of the District	Haryana		Punjab
Name of the District	No. of Establishment	Rank of the District	District Name	No. of Establishment	Rank of the District
Panchkula	24,819	20	Barnala	32,303	20
Ambala	53,055	14	Bathinda	53,829	10
Yamunanagar	54,413	13	Faridkot	37,688	17
Kurukshetra	66,260	5	Firozpur	57,260	9
Kaithal	83,999	2	Fatehgarh	37,712	16
Karnal	60,815	10	Fazilka	44,830	15
Panipat	66,654	4	Ludhiana	213,586	1
Sonipat	56,560	12	Mansa	24,696	22
Jind	62,799	7	Moga	64,824	7
Fatehabad	61,961	8	SAS Nagar	46,650	14
Sirsa	69,606	3	Muktsar	52,030	11
Hissar	96,684	1	Patiala	104,682	5
Bhiwani	63,757	6	Sangrur	126,796	4
Rohtak	42,990	15	Amritsar	134,748	3
Jhajjar	41,188	17	Gurdaspur	61,631	8
Mahendragarh	39,800	18	Pathankot	32,974	19
Rewari	34,688	19	Tarn Taran	50,191	12
Gurgaon	61,914	9	Jalandhar	136,324	2
Mewat	20,139	21	Kapurthala	49,112	13
Faridabad	60,716	11	Hoshiarpur	85,886	6
Palwal	41,969	16	SBS Nagar	31,232	21
			Rupnagar	33,224	18
Total	1,164,786		Total	1,512,208

Source: Compiled from 6th Economic Census (2013–2014).

Sampling Details and Sample Size

Panel A: Detail of vocational education and training providers
Types of Institution	Ludhiana (Punjab)		Hissar (Haryana)
Types of Institution	Total Vocational Institutes*	Sample Institutes	Total Vocational Institutes*	Sample Institutes
Government polytechnic	3	2	2	2
Government polytechnic for women	1	1	0	0
Private/semi government polytechnic	8	2	8	2
Government ITI	7	2	8	2
Government ITI for girls	4	1	3	1
Private ITI	14	2	27	2
Total vocational training providers	37	10	48	9
Panel B: Telephonic or personal interview survey of pass out students
	Punjab		Haryana	Total
Total collected contact information	1458		1296	2754
Not interested to give information	776		644	1420
Not given proper information	165		139	304
Dropouts	27		34	61
Personally meet	34		21	55
Telephonic interview	422		437	859
Total final interviewed	456		458	914
Panel C: Detail of interviewer of pass out graduates
	Punjab		Haryana	Total
Total final interviewed	456		458	914
Polytechnic pass out	257		194	451
ITI pass out	199		264	463
Total female	104		108	212
Total male	352		350	702
Rural pass out	214		308	522
Urban pass out	242		150	392
SCs pass out	173		103	276
OBCs pass out	61		173	234

Source: Compiled from Skill Development and Industrial Training, Haryana and the Department for Technical Education, Haryana. Department of Technical Education (technical education wing) and Industrial Training (IT wing), Punjab; and The Punjab State Board of Technical Education and Industrial Training. Primary Survey, 2017–2018.

Notes

References

Advani

(2013). Factors driving drug abuse in India’s Punjab (ISAS Working Paper No. 177). Institute of South Asian Studies, National University of Singapore.

Agrawal

(2012). Vocational education and training in India: challenges, status and labour market outcomes. Journal of Vocational Education & Training, 64(4), 453–474.

Agrawal

(2014). Gender and caste-based wage discrimination in India: Some recent evidence. Journal for Labour Market Research, 47(4), 329–340.

Ahmed

(2016). Labour market outcome for formal vocational education and training in India: Safety net and beyond. IIMB Management Review, 28(2), 98–110.

Ajithkumar

(2017). A study of the problems faced by Industrial Training Institutes (ITI) in India as perceived by different stakeholders. https://www.igi-global.com/chapter/a-study-of-the-problems-faced-by-industrial-training-institutes-iti-in-india-as-perceived-by-different-stakeholders/176892

Arriagada

A. M.

, Ziderman

, & Mundial

(1992). Vocational secondary schooling, occupational choice and earnings in Brazil (Vol. 1037). Population and Human Resources Department, World Bank.

Ashton

, Green

, James

, & Sung

(1999). Education and training for development in East Asia: The political economy of skill formation in newly industrialised economies. Routledge.

Azariadis

, & Drazen

(1990). Threshold externalities in economic development. The Quarterly Journal of Economics, 105(2), 501–526.

Becker

G. S.

(1993). Human capital: A theoretical and empirical analysis with special reference to education (3rd edition). The University of Chicago Press.

10.

Bellew

, & Moock

(1990). Vocational and technical education in Peru. Economics of Education Review, 9(4), 365–375.

11.

Benabou

(1996). Equity and efficiency in human capital investment: The local connection. The Review of Economic Studies, 63(2), 237–264.

12.

Benhabib

, & Spiegel

M. M.

(1994). The role of human capital in economic development evidence from aggregate cross-country data. Journal of Monetary economics, 34(2), 143–173.

13.

Blaug

(1973). Education and employment problem in developing countries (5th Ed.). International Labour Office.

14.

Blinder

A. S.

(1973). Wage discrimination: Reduced form and structural estimates. Journal of Human Resources, 8(4), 436–455.

15.

Bonnal

, Mendes

, & Sofer

(2002). School-to-work transition: apprenticeship versus vocational school in France. International Journal of Manpower, 23(5), 426–442.

16.

Chernozhukov

, Fernández-Val

, & Melly

(2013). Inference on counterfactual distributions. Econometrica, 81(6), 2205–2268.

17.

Chung

Y. P.

(1990). Educated mis-employment in Hong Kong: Earnings effects of employment in unmatched fields of work. Economics of Education Review, 9(4), 343–350.

18.

Cotton

(1988). On the decomposition of wage differentials. The Review of Economics and Statistics, 70(2), 236–243.

19.

Dasgupta

(1977). India’s green revolution. Economic and Political Weekly, 12(6/8), 241–260.

20.

Deshpande

, & Sharma

(2013). Entrepreneurship or survival? Caste and gender of small business in India. Economic and Political Weekly, 48(28), 38–49.

21.

Deshpande

, Goel

, & Khanna

(2015). Bad karma or discrimination? Male–female wage gaps among salaried workers in India (Discussion Paper No. 9485). Institute for the Study.

22.

Duraisamy

, & Duraisamy

(2005). Regional differences in Wage Premia and returns to education by gender in India. Indian Journal of Labour Economics, 48(2), 335–347.

23.

Duraisamy

, & Duraisamy

(2016). Gender wage gap across the wage distribution in different segments of the Indian labour market, 1983–2012: Exploring the glass ceiling or sticky floor phenomenon. Applied Economics, 48(43), 4098–4111.

24.

Foster

P. J.

(1965). The vocational school fallacy in development planning. Education and Economic Development, 32, 142–166.

25.

Glomm

, & Ravikumar

(1992). Public versus private investment in human capital: Endogenous growth and income inequality. Journal of Political Economy, 100(4), 818–834.

26.

Grubb

W. N.

(1985). The convergence of educational systems and the role of vocationalism. Comparative Education Review, 29(4), 526–548.

27.

Hajela

(2012). Shortage of skilled workers: A paradox of the Indian economy (SKOPE Research Paper no. 111). COMPAS, University of Oxford.

28.

Heckman

(1974). Shadow prices, market wages, and labor supply. Econometrica, 42(4), 679–694.

29.

Hinchliffe

(1990). The returns to vocational training in Botswana—Research note. Economics of Education Review, 9(4), 401–404.

30.

Hoeckel

(2008). Costs and benefits in vocational education and training. Paris: Organisation for Economic Cooperation and Development, 8, 1–17.

31.

IAMR (2011). Skill mapping in two districts: Gurdaspur (Punjab) and Singrauli (Madhya Pradesh). Planning Commission.

32.

IAMR (2013). Rural non-farm employment: A study in Punjab (report 2/2013). Planning Commission.

33.

Kingdon

G. G.

(1998). Does the labour market explain lower female schooling in India? The Journal of Development Studies, 35(1), 39–65.

34.

Kingdon

G. G.

, & Unni

(2001). Education and women’s labour market outcomes in India. Education Economics, 9(2), 173–195.

35.

Lucas

R.E.

(1988). On the mechanics of economic development. Journal of Monetary Economics, 22(1), 3–42.

36.

Machado

J. A.

, & Mata

(2005). Counterfactual decomposition of changes in wage distributions using quantile regression. Journal of Applied Econometrics, 20(4), 445–465.

37.

Madheswaran

, & Attewell

(2007). Caste discrimination in the Indian urban labour market: Evidence from the National Sample Survey. Economic and Political Weekly, 42(41), 4146–4153.

38.

Madheswaran

, & Singhari

(2016). Social exclusion and caste discrimination in public and private sectors in India: A decomposition analysis. The Indian Journal of Labour Economics, 59(2), 175–201.

39.

Mankiw

N. G.

, Romer

, & Weil

D. N.

(1992). A contribution to the empirics of economic growth. The Quarterly Journal of Economics, 107(2), 407–437.

40.

Mehrotra

(2014). India’s skills challenge: Reforming vocational education and training to harness the demographic dividend. Oxford University Press.

41.

Mehrotra

, & Parida

J. K.

(2017). Why is the labour force participation of women declining in India? World Development, 98, 360–380.

42.

Mehrotra

, & Parida

J. K.

(2019). India’s employment crisis: Rising education (CSE Working Paper 2019-04 [pp. 1–23]). Centre for Sustainable Employment, Azim Premji University.

43.

Mehrotra

, Gandhi

, & Kamaladevi

(2015). China’s skill development system: Lessons for India. Economic and Political Weekly, 50(28), 57–65.

44.

Mehrotra

, Gandhi

, & Sahoo

B. K.

(2013). Estimating India’s skill gap on a realistic basis for 2022. Economic and Political Weekly, 48(13), 102–111.

45.

Melly

(2005). Decomposition of differences in distribution using quantile regression. Labour Economics, 12(4), 577–590.

46.

Melly

(2006). Estimation of counterfactual distributions using quantile regression. https://www.alexandria.unisg.ch/22644/2/Counterfactual.pdf

47.

Middleton

, Ziderman

, & Adams

A. V.

(1993). Skills for productivity: Vocational education and training in developing countries. Oxford University Press.

48.

Mincer

(1958). Investment in human capital and personal income distribution. Journal of Political Economy, 66(4), 281–302.

49.

Mincer

(1962). On-the-job training: Costs, returns, and some implications. Journal of political Economy, 70(5, Part 2), 50–79.

50.

Mincer

(1981). Human capital and economic growth. (NBER Working Paper No. 803). NBER.

51.

Mitra

(2013). Can industry be the key to pro-poor growth? An exploratory analysis for India. (Asia-Pacific Working Paper Series). ILO.

52.

Mitra

P. P.

(2018). Skilling and employability: Understanding challenges in India with Special reference to West Bengal. Indian Journal of Public Administration, 64(2), 143–158.

53.

Nelson

R. R.

, & Phelps

E. S.

(1966). Investment in humans, technological diffusion, and economic growth. The American Economic Review, 56(1/2), 69–75.

54.

Neuman

, & Ziderman

(1991). Vocational schooling, occupational matching, and labor market earnings in Israel. Journal of Human Resources, 256–281.

55.

Neumark

(1988). Employers’ discriminatory behavior and the estimation of wage discrimination. Journal of Human Resources, 23(3), 279–295.

56.

NSDC (2013). District and industries wise skill gap studies for the state of Punjab (2012-17, 2017–2022). National Skill Development Council (NSDC). Government of India. https://nsdcindia.org/sites/default/files/files/punjab-sg-report.pdf

57.

Oaxaca

(1973). Male–female wage differentials in urban labor markets. International Economic Review, 693–709.

58.

Padhi

, Mishra

U. S.

, & Pattanayak

(2019). Gender-based wage discrimination in Indian urban labour market: An assessment. The Indian Journal of Labour Economics, 62(3), 361–388.

59.

Pal

, & Birthal

P. S.

(1992). Changes in agriculture of the green revolution states: Implications for agricultural development. Agricultural Economics Research Review, 5(1), 1–21.

60.

Parey

(2016). Vocational schooling versus apprenticeship training. Evidence from vacancy data. Conference Paper (Labor Markets and Institutions, No. F20-V1). ZBW.

61.

Parida

J. K.

, & Raman

K. R.

(2018). India: Rising trends of international and internal migration. In Triandafyllidou

, (Ed.), Handbook of migration and globalisation (pp. 226–248). Edward Elgar Publishing.

62.

Pilz

(2016). India: Preparation for the world of work. Education system and school to work transition. Springer VS.

63.

Psacharopoulos

(1987). To vocationalize or not to vocationalize? That is the curriculum question. International Review of Education, 33(2), 187–211.

64.

Psacharopoulos

, & Chu Ng

(1994). Earnings and education in Latin America. Education Economics, 2(2), 187–207.

65.

Ravallion

, & Datt

(1996). India’s checkered history in fight against poverty: Are there lessons for the future? Economic and Political Weekly, 31(35/37), 2479–2485.

66.

Reimers

C. W.

(1983). Labor market discrimination against Hispanic and black men. The Review of Economics and Statistics, 65(4), 570–579.

67.

Romer

P. M.

(1990). Endogenous technological change. Journal of political Economy, 98(5, Part 2), S71–S102.

68.

Ryan

(1998). Is apprenticeship better? A review of the economic evidence. Journal of Vocational Education & Training, 50(2), 289–325.

69.

Ryan

(2001). The school-to-work transition: a cross-national perspective. Journal of Economic Literature, 39(1), 34–92.

70.

Saha

K. C.

(2012). Irregular migration from India to the EU: Punjab and Haryana Case Study. CARIM-India Research Report.

71.

Schlessman-Frost

(1991). Educational evaluation in a pluralistic democracy: Some policy recommendations. Scope, 91(1), 39–44.

72.

Schultz

T. W.

(1960). Capital formation by education. Journal of Political Economy, 68(6), 571–583.

73.

Shetty

S. L.

(2003). Growth of SDP and structural changes in state economies: Interstate comparisons. Economic and Political Weekly, 38(49), 5189–5200.

74.

Silverberg

, Warner

, Fong

, & Goodwin

(2004). National assessment of vocational education (Final Report to Congress). US Department of Education.

75.

Singh

, Parida

J. K.

, & Pattayat

S. S.

(2020a). Why is the employment outcome of vocationally trained youths so poor? Evidence from selected districts of Punjab and Haryana. IASSI Quarterly: Contributions to Indian Social Science, 39(1), 111–134.

76.

Singh

, Parida

J. K.

, & Awasthi

I. C.

(2020b). Employability and earning differentials among technically and vocationally trained youth in India. The Indian Journal of Labour Economics, 63(2), 1–24.

77.

Singh

, & Kohli

D. S.

(1997). The green revolution in Punjab, India: The economics of technological change. J Punjab Stud, 12(2), 285–306.

78.

Singh

R. B.

(2000). Environmental consequences of agricultural development: A case study from the Green Revolution state of Haryana, India. Agriculture, Ecosystems & Environment, 82(1–3), 97–103.

79.

Singh

(2001). Green Revolution: In light and shade. Economic and Political Weekly, 36(42), 3996–3998

80.

Singhari

, & Madheswaran

(2017). Wage structure and wage differentials in formal and informal sectors in India. The Indian Journal of Labour Economics, 60(3), 389–414.

81.

Smith

(2014). Donkey flights: Illegal immigration from the Punjab to the United Kingdom. Migration and Policy Institute. https://www.migrationpolicy.org/sites/default/files/publications/BadActors-DonkeyFlights-FINALWEB.pdf

82.

Thorat

, Attewell

, & Rizvi

F. F.

(2009). Urban labour market discrimination (Working Paper). Indian Institute of Dalit Studies.

83.

Tilak

J. G.

(1988). Economics of vocationalization: A review of evidence. Canadian and International Education, 17(1), 45–62.

84.

Tilak

J. B.

(2003). Vocational education and training in Asia. In Keeves

J. P.

, Watanabe

(Eds.) International handbook of educational research in the Asia-Pacific region (pp. 673–686). Springer.

85.

World Bank (2007). 2006 Baseline study of India’s industrial training institutes: Institutional performance and employment of ITI graduates. South Asia Region, Human Development Department (SASHD). World Bank.

86.

World Bank (2008). Skill development in India: The vocational education and training system (Report no. 22). Human Development Unit, South Asia Region. World Bank.

87.

Zymelman

, Woodruff

, Horowitz

, & Herrnstadt

(1976). The economic evaluation of vocational training programs. World Bank, Johns Hopkins University Press.