WAGE DIFFERENTIALS BETWEEN LOCAL AND MIGRANT PERMANENT FARM SERVANTS IN PUNJAB(INDIA) Dr. Varinder Sharma Development Studies Unit, Institute for Development and Communication(IDC), Chandigarh, India Email:varinder_10@hotmail.com
Wage Differentials between Local and Migrant Farm Servants in Punjab (India). Permanent ABSTRACT In Punjab, the new agriculture technology increased the production, productivity and the demand and wages of agricultural labrourers. The continuously rising demand and wages of hired farm labourers attracted the labourers from other states of India. The influx of migrant casual agricultural labourers caught the attention of many researchers. A majority of the researchers came on the conclusions that migrant agricultural casual labourers depressed the wage rates of local casual agricultural labourers and the migrants have no interest in agricultural workers unions in Punjab. The scholars paid no attention to the migrant permanent farm servants. Many migrant casual agricultural labourers started working and living as permanent farm servants in the villages of Punjab. Moreover, there is no study on their wage rates etc. The common views of farming community on migrant permanent farm servants are: they are very productive, and essential for farming in Punjab, and work on fewer wage. To probe this view and to fill the gaps in literature, the present study has been carried out. The source of data is primary data of 240 permanent farm servants. On the basis of theories and empirical studies we selected variables like Age, marital status, tractor driving skill, familiarity with electric motor(tubewell)operations, supervisory capacity, farm size and VARINDER SHARMA Development Studies Unit Institute for Development and Communication (IDC), Chandigarh Email: varinder_10@hotmail.com 1
geographical zones like sub mountain zone and central plain zone. The variables like tractor driver s skill, supervisory capacity and farm size positively impact the wage rates. The submountain zone variable negatively affects the wage rate of local and migrant permanent farms servants. By using Blinder s technique of decomposition, we decomposed the wage rates of two groups of permanent farm servant. The results from this came out as: 86 percent difference in the wage rates of local and migrant permanent farm servants is due to their personal endowments representing the productivity. The remaining 14 percent difference emerges due to discrimination towards migrant permanent farm servants in labour market. Keywords: Permanent Farm Servants; Decomposition I. Introduction The new agriculture technology in Punjab increased the demand of casual agricultural labourers for seasonal farm operations and of permanent farm servants for skilled and responsible farm operations. 1 The continuous rise in the demand of farm labourers and higher wages in Punjab attracted the farm labourers from other states of India. 2 The rapid inflow of migrant labourers in rural areas of Punjab raised the attention of many researchers. Many studies discussed the interstate migration of agriculture labourers in Punjab. 3 A majority of the studies discussed the pattern of employment and wage rates of migrant casual agricultural labourers. Over the span of years many migrant agricultural labours started working and living as permanent farm servants in the villages of Punjab. 4 In literature we have not found any single study which studied the factors affecting the wage rates of local and migrant permanent farm 2
servants. Moreover, there is no study which decomposed their wage differentials. In this present article, we have made an attempt to discuss the factors affecting the wage rates of local and migrant permanent farm servants. Secondly, the decomposition of their wage differentials of local and migrant permanent farm servants have divided into two parts: (i) the differences due to the productivity related endowments (ii) the prevalence of differences in the wage rates due to discrimination in the labour market. II. The Data: The main secondary data sources like Census of India and the NSSO (Employment and Unemployment Surveys) do not provide us any information and data on the permanent farm servants. 5 These data sources also do not give any information on the seasonal migrant agricultural labourers. The data collected by individual researchers are not easily accessible. In this study we collected data ourselves from randomly selected 240 permanent farm servants from thirty villages of Punjab during year 200708. The villages from each district of Punjab have selected in proportion to male agricultural labourers in that district. From each selected village we surveyed eight permanent farm servants. Out of 240 permanent farm servants, we surveyed 134 local permanent farm servants and 106 migrant permanent farm servants. III. Determination and Decomposition of Wage Rates of Migrant and Local Permanent Farm Servants: Theoretical Framework: 3
It is generally assumed that hired farm workers get the same wage rate in a particular village/region. On the contrary, vast variations prevail in the wage rates of permanent farm servants not along the villages but within a village. In a village a wide gap exists in the wage rates of permanent farm servants in a village and along the regions needs to be explained. The literature pertaining to wages of agricultural labourers do not give satisfactory answers on this issue. However, there is a vast and rich literature on personal income distribution and determinants of earnings of workers and employers in the urban sector. 6 By exploring this literature we are able to identify several factors that may be responsible for variations in the earnings of different workers. The human capital theory [Becker, 1975 and Joll, 1983] lays emphasis on the role of personal factors and attributes and skills in determining the wage rates of workers. Workers with better personal attributes, higher education and technical skills will get higher wages than others. The institutional economists have emphasized the role of market imperfections, gender and social bias in the determination of wage rates of workers [Watchtel, 1972 and Smith, 1978]. The labour search theory broadly agrees with the human capital theory view but emphasizes the role of market imperfections and search costs [Stigler, 1962]. On the basis of these theories and other clues in the literature the following factors were identified as important in determination of wage rate of permanent farm servants. (i) Age: Most of the theories and empirical studies support a positive relationship between age and wage rate; 4
(ii) (iii) (iv) (v) Marital Status: The pressure of family needs makes a person more responsible and hard working. Training and Skills: Most of the theories and empirical evidence show a positive impact of training and skills of the worker on his earning. Size of the Firm: Bigger firms usually pay higher wage rate for the similar category of workers; Regions: The workers generally get high wages in developed geographical regions. On the basis of above theoretical and empirical studies we selected the following variables in this study: (i) Age (ii) Marital Status (iii) Tractor driving skill (iv) Familiarity with electric motor operations (v) Supervisory capacity (vi) Size of Farm (Acres) (vii) Sub mountain zone where agricultural productivity is low and (viii) Central plain zone where agricultural productivity is high. Further, on the basis of above defined variables we estimated and decomposed the wage equations by using the technique of Blinder (1973) 7 in the following manner: In. WL = L.XL + ul In. WM = M.XM + um (i) (ii) In the above equation (i) and (ii), symbol M stands for migrant permanent farm servants and L for local permanent farm servants. The symbol W stands for wages measured in rupees. The estimated coefficients from such a model measures approximately the proportionate affect on wages of change in the 5
right side variable X which is a vector of measured characteristics of the workers such as Age and of their employing farms such as size of farm and regional location of a farm on the basis of productivity of a region. The vector of regression coefficient (alpha) reflects the return that the market yields to a unit change in endowments such as technical/mechanical skills, Age and regions and size of the farms. The error term u reflects measurement error as well as the effect of unmeasured or unobserved factors. A property of ordinary least square (OLS) regression analysis is that the regression line pass through the mean values of the variables so that: In. WL = ˆ L. XL In. WM = ˆ M. XM (iii) (iv) The hats or caps denote ordinary least squares estimated values of coefficients. If permanent farm servants in group M receive the same returns as permanent farm servants in group L for their endowments of wage determining characteristics i.e. (if permanent farm servants in group M were given the group L wage structure), then their average wage would be: In. W M * = ˆ L. XM (v) This is the average groupm wage that would prevail in the absence of wage discrimination (where wage discrimination is defined as unequal wage for the same endowments of wage determining characteristics). Subtracting (v) from (iii) gives the difference between average group L wages and the average 6
hypothetical group M wages that would prevail if group M were paid according to group L wages. This difference shows their different endowments of wage generating characteristics i.e. ln. WL ln WM * = ˆ L. X L ˆ L. X M = ˆ L (X L X M ) (vi) Subtracting (iv) from (v) yield the difference between the hypothetical nondiscriminating group M wage and theoretical returns to the same wage generating characteristics i.e. ln. WM * ln WM = ˆ L. X M ˆ M. X M = X M ( ˆ L ˆ M) (vii) Adding (vi) & (vii) gives us: = ˆ L (X L X M ) + X M ( ˆ L ˆ M) This is the overall wage gap in group of local permanent farm servants (L) and migrant permanent farm servants (M). The wage gap is divided into two components: one is the portion attributable to differences in the endowments of wage generating characteristics (XLXM) evaluating at the group L returns ( ˆ L). The second portion is attributable to difference in the returns ( ˆ L ˆ M) that group L and group M get for the same endowment of wage generating characteristics (XM). This component is often taken as reflecting discrimination or wage differentials. 7
IV. Estimation of Wage Equations of Local and Migrant Permanent Farm Servants: The wage equations for local and migrant permanent farm servants have estimated separately and given in table1. Equation1, Equation2 having high value of R 2..The variables in both the equations seem well specified. The high values of F represent that both these estimated wage equations are structurally different from each other. 8 The variable Age is negative and not significant even at 10 percent. The negative sign shows the reverse relationship between Age and wage rate. The Agewage relationship may be captured more clearly by taking square of Age. But the squared variable hinders the visibility of other variables and impacts the value of R 2. 9 The variable marital status having right signs as per our expectation but it is insignificant in equation1. In equation 2 it is positive as well as significant from this we may conclude a married migrant permanent farm servant gets more wages. A majority of the local permanent farm servants are married in this study due to this, the variable may be insignificant but positive in equation1. 8
Table1 Wage Determination Model for Local and Migrant Permanent Farm Servants: Used for Decomposition Analysis [Migrant (N) = 106; Local (N) = 134] Dependent variable = Log of Annual Wage Rate (Rs.) Explanatory Variables Local Permanent Farm Servants ( ˆ L) (Equation1) Migrant Permanent Farm Servants ( ˆ M) (Equation2) Age (Years) 0.0009 (0.46) 0.003 (0.63) Marital Status (Dummy) Married=1, Unmarried=0 0.07 (1.30) 0.32 (4.07) a Drives tractor on Farms (Dummy): Yes=1; No=0 0.14 (2.95) a 0.17 (1.96) c Knows all electric motor operations (Dummy): Yes=1, No=0 0.24 (2.00) b 0.003 (0.03) Supervises casual labour on farms (Dummy): Yes = 1, No=0 Farm size (Dummy): (Acres) >10=1, other=0 0.34 0.30 (2.97) a (3.58) a 0.20 0.17 (3.14) a (1.98) c Zone1 (Dummy): Submountain=1,others=0 Zone2 (Dummy): Central plain=1, Others=0 0.26 (2.50) b 0.16 (1.73) c 0.05 0.12 (0.97) (1.33) Intercept 9.36 9.46 R 2 0.69 0.70 F values 21.59 a 16.80 a Note: (i) Figures in parentheses are tvalues. (ii) a, b and c are significant at 1%, 5% and 10%. From the variables relating to technical skills, the tractor driver s skill seems more important in equation 1 and equation2. This variable has right sign according to our expectation and highly significant in both equations. The tractor driver s skill increases the wage rates of both local and migrant permanent farm servants. The second technical variable, familiarity with electric motor (tube well) operations is positive and significant only in equation 1. It seems local permanent farm servants who know 9
how to operate and do minor repair of electric motors (tube wells) get more wages. The migrant permanent farm servants usually lack in this skill and in their case variable is insignificant. On the very big farms, a permanent farm servant not only supervises the casual labourers but arranges labourers also. A permanentfarm servant who does this task acts as a foreman on the farms. Here, in equation 1 and equation2, a local and migrant permanent farm servant who having the quality of a foreman gets more wages. The supervisory capacity variable is positive and highly significant in two equations. A permanent farm servant either local or migrant gets more wages on bigger farms having area more than or equal to ten acres. On such farms a permanent servant remains busy whole year and having more responsibilities. Contrary to technical skill variables and supervision variable, the regional variable i.e. sub mountain zone remains negative and significant in equation 1 and 2. It seems the local and migrant permanent farm servants get fewer wage in this zone. The local permanent farm servants get fewer wages than migrants. Due to the low agriculture productivity and small farms in this zone, the permanent farm servants get fewer wages in this zone. The second regional variable i.e. central plain zone is insignificant in both equations. It seems this variable has no effect as such and having no variation in the wage rates of local and migrant permanent farm servants. From the preceding discussion we may say the technical skills especially tractor driver s, supervisory capacity; size of farm and sub mountain zone affect the wage rates of local and migrant permanent farm servants. 10
V. Decomposition of Wage Differentials: The regression coefficients estimated in equation 1 and equation 2 given in table 1 have reproduced in columns (i) and (ii) in table 2. The mean of different characteristics of local and migrant permanent farm servants have given in columns (iii) and (iv) in table2. These regression coefficients and statistics have used for calculating wage differentials due to personal endowments and structural differences or commonly called discrimination on the basis of Blinder s technique discussed in section (3). The column (v) in table2, which contains the amount of absolute contribution of individual characteristics towards the portion of overall earnings differentials between local and migrant permanent farm servants. The productivity related skills like tractor driver, knowledge of electric motor operations (tubewell), and capacity to supervise the casual labourers, farm size, geographical regions and marital status contribute towards in the favour of local permanent farm servants. The only factor that contributes towards this portion in favour of migrant permanent farm servants is Age. Further, the contribution of different factors behind the portion of wage differentials arises due to differences in coefficients of explanatory variables of two wage equations, we found from column (vi) in table2 that marital status, tractor driving skill and sub mountainous region contribute substantially in favour of migrant permanent farm servants. 11
Table2 Decomposition Of Wage Differentials Between Local And Migrant Permanent Farm Servants [Migrant (N) = 106; Local (N) = 134] Explanatory Variables ˆ L (i) ˆ M (ii) X L (iii) X M (iv) ˆ L(X LX M) Age (Years) 0.0009 (0.46) 0.003 (0.63) 34 25 ()0.01 (3.33) Marital Status (Dummy) 0.07 0.32 0.69 0.49 0.01 Married=1, Unmarried=0 (1.30) (4.07) a (3.33) Drives tractor on Farms 0.14 0.17 0.44 0.18 0.04 (Dummy): Yes=1; No=0 (2.95) a (1.96) c (13.33) Knows all electric motor 0.24 0.003 0.94 0.76 0.04 operations (Dummy): (2.00) b (0.03) (13.33) Yes=1, No=0 Supervises casual labour on farms (Dummy): Yes = 1, No=0 Farm size (Dummy): (Acres) >10=1, other=0 Zone1 (Dummy): Submountain=1,others=0 Zone2 (Dummy): Central plain=1, Others=0 (v) 0.34 0.30 0.91 0.53 0.13 (2.97) a (3.58) a (43.33) 0.20 0.17 0.85 0.76 0.02 (3.14) a (1.98) c (6.67) 0.26 (2.50) b 0.16 (1.73) c 0.11 0.32 0.06 (16.67) 0.05 0.12 0.19 0.52 0.02 (0.97) (1.33) (6.67) Intercept 9.36 9.46 Decomposition of wage Differentials: (i) Differences due to endownments ˆ L (X L X M ) (ii) Difference due to coefficients Note: X M ( ˆ L ˆ M) 0.30 (100) X M ( ˆ L ˆ M) (vi) 0.05 (33.33) ()0.12 (80) () 0.01 (6.67) 0.18 (120) 0.02 (13.33) 0.02 (13.33) ()0.03 (20) 0.04 (26.67) 0.15 (100) (1) Under scripts L and M denotes Local and Migrant Permanent Farm Servants. (2) In column (i) and (ii) are the multivariate coefficients. (3) In column (iii) and (iv) are arithmetic means. (4) In column (i) and (ii) figures in parentheses are t values. a b and c are significant at 1%, 5% and 10% (5) In column (v) and (vi) figures in parentheses are percentages. The most dominant among them is the marital status and the next is sub mountainous zone. The above discussed results have further summarized and analyzed in table3. 12
Table3 Decomposition Of Total Differences In Wage Rates Of Local And Migrant Permanent Farm Servants Sr. No. Summary of Results 1. Wage differentials due to different endowments [ ˆ L (X L X M )] 0.30 (86%) 2. Wage differentials due to difference in the 0.15 coefficient of explanatory variables[x M ( ˆ L ˆ M)] 3. Intercept differentials ()0.10* 4. Wage differentials due to structural differences 0.05 (14%) (2)+ (3) 5. Overall Wage Differentials (1) + (4) 0.35 (100%) 6. Average log wages of local permanent farm 10.09** servants 7. Average log wages of migrant permanent farm 9.73** servants 8. Difference (6) (7) 0.36 9. Geometric mean of wages of local permanent Rs. 24100*** farm servants 10. Geometric mean of wage of migrant permanent Rs. 16815 farm servants 11. Differences in geometric mean of wages (9) Rs. 7285 (10) 12. Due to endowment differences in geometric Rs. 6244 mean 13. Due to structural differences in geometric mean Rs. 1041 Note: * This is unexplained portion of wage differentials ( ˆ OL ˆ OM ) ** Since regression lines pass through the averages of independent and dependent variables: Average logarithm wages (In Y) = X ***Geometric mean of earnings = Antilog [In Y] From table 3, it comes out that main difference almost 86 percent in the wage rate of local and migrant permanent farm servants is due to their different productivity related endowments. Secondly, the migrant permanent farm servants almost get 14 percent less wages in comparison to locals. This 14 percent less wages is due to the discrimination towards migrants. Next, we have also calculated, on the average a migrant permanent farm servant earns Rs. 7285 less than a local permanent farm servant. Out of the total difference of Rs. 7285 in the average wages of local and migrant permanent farm 13
servants, Rs. 6244 is due to the superior endowments of a local permanent farm servant. The remaining difference of Rs. 1041 in wages of local and migrant permanent farm servants exists because the migrant permanent farm servants are paid at a lower rate for the same endowments or characteristics. Conclusions: The new agriculture technology in Punjab increased the demand and wages of farm labourers. The rising wages of agricultural labourers attracted the labourers from other states of India. Many seasonal migrant agricultural labourers started working and living as permanent farm servants in villages of Punjab. Among farmers there are two views on migrant permanent farm servants. One group of farmers consider them very productive and second group think, the migrant permanent farm servants are not skilled in farm operations like locals. Due to this, the migrant permanent farm servants get fewer wages. To probe these views, we first identified the variables which affect the productivity of permanent farm servants. With the help of these variables, we estimated wage equations for local and migrant permanent farm servants. The technical skills like know how to drive tractor, supervisory capacity and farm size positively affect the wage rate of local and migrant permanent farm servants. The sub mountainous zone has negative impact. The decomposition analysis further gives us the answer on the discrimination in labor markets of local and migrant permanent farm servants. Here we found a major difference of 86 percent in the wage rates of local and migrant permanent farm servants is due to their productivity related endowments and 14 percent differences due to the residential status. 14
NOTES 1. The permanent farm servants are hired for performing more responsible farm operations and where there is need of monitoring casual labourers [Aggarwal (1981) and Eswaran (1985) et al]. 2. The migration in rural areas of Punjab continuously increased after green revolution [Johl (1975)]. 3. The interstate and intradistricts migration in Punjab studied by many scholars like [Oberai and Singh (1980) and Khurana (1992)]. 4. For details see Johl (1975). 5. First and Second Agricultural Labour Enquiry Reports during 1950s discussed the attached/permanent farm servants and casual labourers separately but due to the criticism by Raj (1962), on the division of two categories of agricultural labourers i.e. casual labourers and permanent farm servants,the statistical authorities dropped this demarcation from further reports. 6. Many scholars have extensively studied the earnings and income distribution of workers and employees in the urban sector [Soren (1979) and Blaugh (1974)]. 7. The decomposition technique was first developed by Blinder (1973) and Oaxaca (1973). This technique was modified to incorporate selectivity bias [Reimer (1983)] and to overcome the index number problems [Cotton (1988), Newmark (1988), Oaxaca and Ransom (1994)]. Another modification incorporates the occupational distribution earning estimation and was first proposed by [Brown,Moon and Zolath (1980)]. Here, we relied upon Blinder s approach. 15
8. A chow test assumes that variance for subsamples are equal. To test for this assumption, we calculated Fstatistics for each pair as: F = Larger variance / Smaller variance None of the Fstatistics was significant at one percent level. 9. Mincer and Polachek (1974) suggested to drop the square term from regression equation which generally reduces the visibility of life cycle investment profile. REFERENCES Becker, G.S. (1975): Human Capital: A Theoretical and Empirical Analysis, National Bureau of Economic Research, New York. Bina, Aggarwal (1981): Agricultural Mechanization and Labour Use: A Disaggregated Approach International Labour Review, Vol. 120, No.1, pp.115127. Blaugh, Mark (1974): An Economic Analysis of Personal Earnings in Thailand. Economic Development and Cultural Change, Vol. 23, No.1, pp.131. Blinder, A.S. (1973): Wage Discrimination: Reduced Form and Structural Estimates. Journal of Human Resources, Vol. 8, No.4, pp. 436555. Blomquist, N.Soren (1979): Wage Rates and Personal Characteristics. The Scandinavian Journal of Economics, Vol.81, No.4, pp.505520. Brown, R.S., Moon, M. and. Zolath, B.S. (1980): Incorporating Occupational Attainment in Studies of Male/Female Earnings 16
Differentials. Journal of Human Resources, 28. Vol. 15, No.1, pp.3 Cotton, C.J. (1988): On the Decomposition of Wage Differentials. Review of Economics and Statistics, Vol. 70, No.2, pp. 236243. Eswaran, Mukesh and Kotwal, Ashok (1985): A Theory of Two Tier Labour Markets in Agrarian Economies. American Economic Review, Vol. 75, No.1, pp. 162177. Johl, S.S. (1975): Gains of the Green Revolution: How They Have Been Shared in Punjab. Journal of Development Studies, Vol. 11, No.3, pp. 178189. Joll, C. and Mckenna, C et al (1983): Development in Labour Market Analysis, George Allen and Urmin Ltd., London. Khurana, M.R. (1992): Agricultural Development and Employment Patterns in India: A Comparative Analysis of Punjab and Bihar, Concept Publishers, New Delhi. Mincer, Jacob and Polachek, Solomon (1974): Family Investments in Human Capital: Earnings of Women. Journal of Political Economy, Vol. 82, No.2 (PartII), pp. S76S110. New Mark, D. (1988): Employers Discriminatory Behaviour and the Estimation of Wage Discrimination. Journal of Human Resources, Vol. 23, No.3, pp. 279295. Oaxaca, R.L. (1973): MaleFemale Wage Differentials in Urban Labour Market. International Economic Review, Vol. 14, No. 3, pp. 693 709. Oaxaca, R.L. and Ransom, M.R. (1994): On Discrimination and the Decomposition of Wage Differentials. Journal of Econometrics, Vol. 61, No.11, pp. 521. 17
Oberai, A.S. and Singh, H.K, Manmohan (1980): Migration Flow in Punjab s Green Revolution Belt. Economic & Political Weekly, Vol. 15, No.13, pp A2A12. Raj, K.N. (1962): Some Comments on the Second Agricultural Labour Enquiry, in V.K.R.V. Rao (Ed): Agricultural Labour in India, Asia Publishing House, New Delhi, pp. 152169. Stigler, G.J. (1962): Information in the Labour Market. Journal of Political Economy, Vol. 70, No.2 (Part2), pp.94104. Wachtel, H. and C.Betsey (1972): Employment at Low Wages, Review of Economics and Statistics, Vol. 54, No.2, pp.121129. 18