40 PERCENTAGE DIST SOCLRIBUTION Of \L SCIENT]TABLE If TOTAL LAB< COMPONENTS [ST3UR COST Wl •O ITS VARIO us (percen ^ Tear Salaries Iwages, Profit Benefits Old age Other Total allowances sharing in benefits social bonus kind security charges (J) w w w (S) 1963 84.9 6.3 2.4 4.9 1.5 100.0 1964 85.8 5.0 2.3 5.4 1.5 100.0 1967 84.4 5.3 3.1 5.5 1.7 100.0 1968 84.2 5.1 3.1 5.8 1.8 100.0 1969 64.2 5.3 3.0 5.8 1.7 100.0 1970 83.4 5.2 3.4 6.0 2-0 100.0 1971 82.4 6.1 3.5 6.1 1.9 100.0 1973 81.0 6.7 3.7 6.7 1.9 100.0 1974 80.8 6.7 3.8 6.8 1.9 100.0 SOURCE: Indian Labour Tear Book, various issues. NOTES : I) These data are based on labour statistics collected under COS Act. 2) We have taken all years for which data are available 3) Col 3 includes both individual and group benefits. (based on data collected by the COS) gives the break-up of total labour cost into its various components. While column (1) is covered in both POW and COS, the former excludes individual benefits (which would account for about 1 per cent of all benefits in kind) and the latter bonus (6-7 percent). Both exclude approximately 9-10 percent of labour costs under columns (4), (5) and part of (3). The inclusion/exclusion of certain allowances/ benefits would thus affect the average earnings per employee, depending on the source used. Also, the ratio of such allowances/benefits has risen from about 15 percent in 1963 to 20 percent in 1964 so that the rate of growth of earnings in the inclusive series would be higher than if we consider only pay and dearness allowance. Within the narrower definition of earnings as opposed to total labour costs, the POW data are perhaps more inclusive. However, the POW Act covers only earnings upto Rs 400. This tends to give the data a downward bias, since workers whose earnings rise above Rs 400 per month would drop out of the index. On the other hand, COS data, which cover all workers and non-workers, are more inclusive in the coverage of employees. And this gives it a pronounced upward bias since it obviously includes a higher proportion of better paid jobs. Hence the two sets of data are not comparable. However, these data have been used to show