Studies in the small-large issue in the Indian textile industry : relative costs in the handloom, powerloom, and mill sector

7_77 STUDIES IN THE SMALL-LARGE ISSUE IN THE INDIAN TEXTILE INDUSTRY: NO.2. RELATIVE COSTS IN THE HANDLOOM, POWERLOOM AND MILL SECTORS By Dipak Mazumdar Series: Studies in Employment and Rural Development No.77 Division: Employment and Rural Development Department: Development Economics Development Policy Staff International Bank for Reconstruction and Development This is part of the series of papers being produced as part of the World Bank Research Project 671-59 Small-Scale Enterprise De- velopment. The views expressed are those of the author and the report may not be published or quoted as representing the views of the World Bank or its affiliated organizations. Washington, D. C., December 1981. The World Bank Survey of powerlooms and handlooms was conducted in one provincial town - Mau - in Eastern Uttar Pradesh. Mau is essentially a center of production and trade in textiles. These are a large number of handloom and powerioom units, as well as a few large mills. Several of the large mills have been taken over by the NAtional Textile Commissioner's Office after they had become "sick". In Sections I and II we analyze the cost of production data obtained from our survey of handlooms and powerlooms. No data were collected from the large mills. The comparison of costs in mills and powerlooms are made in Section III on the basis of studies conducted by other investigators. In the analysis presented in this paper we consider only average values of key economic ratios. Results from production function analysis of our survey data are given in a separate paper. I The Samle A. Handlooms. Our Survey turned up two separate types of handloom establishments: (i) the small handloom unit which is essentially a family enterprise located in the house of the owner; and (ii) master-weavers' units which typically employ weavers as hired workers who, however, work in their own homes. The distribution of the two types of units in our sample is given in Table I eN -2- Table 1: Distribution of Eandloom Units in the Sample Size No. of looms Mean Employment Mean Capacity No. of units (meters per day) (i) Small 1 3.4 8.2 14 2 4.7 13.6 9 3 5.7 21.1 7 4 6.0 30.0 1 5-6 6.0 20.0 1 (ii) Large 15-20 37.5 310 2 30-35 62.5 363 4 45 98.0 540 1 70 140.0 770 1 The difference in organization between the two types of units needs to be explained a bit more. The weaver (producer) does not typically do his own marketing. He sells his output at a piece rate to the master weaver who generally also advances the yarn needed for his product. This relationship between the producer and the master weaver is as true for the small unit as for the master-weaver's establishment. Are we then merely looking at the same type of firm from two different ends? The answer is no, and the real difference turns on the onwership of looms. In the case of the small establishments, the weaver (producer) himself owns the loom, while in the larger establishments covered in our Survey the weavers are set up in their own residences with equipment paid for by the master weaver. This is revealed by looking at the figures collected in the survey on initial capital required to start the business. They tend to be typically small for the independent units, and large for the master-weavers' establishments (varying with the number of looms owned). The difference between the two types of weaving units is critical to the evaluation of the cost of capital - -3- both fixed and working capital - as well as for value added and earnings per worker - from the data obtained in our survey, and will be considered in detail later in the paper. B. Powerlooms The powerloom units in Mau made use of older or second-hand looms located in small groups. Unlike handTooms, the sheds in whch these powerlooms are located are not always part of residential premises of the owners The raw material i.e. the cotton yarn and staple yarn are first sent to large-scale sizing plants existing in the town for preparing-beans. These beams when prepared are rolled on to bobbins (with the help of winding machine) which in their turn are fixed on to the shuttles of the powerloom for weaving the cloth. At the end of the process the manufactured cloth is again sent for calendering to a larger unit to give a glaze to the product. The size distribution of the powerloom units covered in our sample were as follows: Number of Looms 1 - 5 6 - 10 11 - 15 .16 - 20 21 - 30 31 & above Total Av. No. of Employed: Eired 8.50 14.75 21.70 27.88 43.00 34.00 19.92 Hired plus fa- 8.63 15.25 21.70 27.88 43.00 34.00 20.05 mily No. of Units 8 8 10 8 1 2 37 -4- A pecul-arity of the industry was that all the workers employed were male, and a very large proportion of those employed were hired operators on the looms (weavers). Type of Products The handlooms in our survey produced only sarees. Most of then produced just one quality of saree - medium 60s x 60s count, although there were two or three units producing coarser sarees as well and two or three producing fine sarees as an additional line of business. The powerlooms produced dhoties and chadday as well as sarees. It was not possible to get cost data separately for the three types of output. The comparison between handloom and powerlooms has, therefore, to be in terms of a Rupee of Value Added of the cloth produced. The comparison makes sense because the powerlooms like the handlooms concentrated in the medium 60s x 60s count range in the cloth woven, even if they did not specialize in sarees. -5- II Economics of Handloom Weaving (a) The Earnings of Weavers The master-weaver's establishments makeuse of wage earners. The weavers working under them are paid by the pieces of sarees produced. The independent units use their family labor and sell the sarees produced mostly to the master weavers, although in two of the thirty-two cases of such units we found that they sold their products directly to whose- salers as well. The basic data on labor and production for the two types are given in Table 2. The most important difference between the two types is (a) the type of labor involved in weaving and (b) the output loom ratio. In the small units most of the family'is working at least part time. In the larger units only adult male workers are reported to be working as hired labor. It is conceivable that they are assisted by some family members who were not recorded in the master-weavers' statements to the inter-viewers However, the difference between small and large units as far as adult male workers per establishment, are concerned, seems to be real. Looking at the figures on earnings accruing to the workers, the average monthly wage per adult male-worker is roughly the same in the two types of units (around Rs. 160). The number of looms per adult male worker is twice as much in the independent units than in the master weaver's units (1 as against 0.5). But it is seen that productivity per loom in the small units is a little more than 50 per cent of that in the master-weavers' establishments. Thus Table 2; LABOR, LOOM AND PRODUCTION IN IIANDLOON Type No. of Loom No. of Workers/Unit Pieces Produced Average Wage Earnings per Pieces Pieces per unit Hired Family per month per piece month per produced produced M WF WP C adult male per loom per adult worker (Rs.) per month male worke permonth) Large 34.9 70.4 2,304 161 66.0 32.7 Small 1.9 - 1.94 0.58 0.54 1.42 58.07 5.258 157.2 30.6 29.9 M adult male WF= full=time adult women HP part-time adult women C children in fact, the larger numoer of looms per worker in the independent units merely means that there is a great deal of underutilization of loom capacity in such units leaving the productivity per worker more or less the same in the two types of units (considering that is to say, adult male workers only). This could seen reasonable in terms of the organizational difference noted above between the tw,o types of units. When the looms are owned by the master weavers they can be expected to maintain then in good working conditions, while many of the looms owned by the small-scale weavers will be old and of low efficiency. We thus take the mid-point of the earnings for adult male worker in the two types of unit - approximately Rs. 159 per month - as the wages of an adult male worker in handloom weaving in Mau. (b) The Cost of Capital (i) Fixed Canital The fact that a larger number of looms are available to small units (at a lower level of utilization) might suggest that the capital cost of such units will be higher. But in fact, the low price of loom used in the handloom industry makes the cost of machines a minor part of total fixed capital. Much the larger part of the value of fixed capital is accounted for by land and building used for the production process. The figures from our Survey are given in Table 3. Table 3: Current Value of Fixed Capital per Adult Male Worker (Rs.) Machines Land and Building- Total Small 214 1138 1352 Large 159 396 555 I/ The low value of land and building for the larger master-weaver's firm is misleading. The master-weavers' only gave the value of the premises used for their trading purposes - the space needed for storing raw materials and cloths received from the handloom weavers in their employment. The value of the premises of the weavers actually devoted to production is 9 not included in the questionnaires filled in by the master-weavers. From the point of view of the cost of production in the handloom industry, the figure of the value of land and building given for the small producer is more appropriate. Does one include the cost of land and value of the master weavers units on the ground that it is part of the chain of storing the cloth before it reaches the wholesaler? Probably it should be included, given the peculiar two-stage process of producing cloth in the handloom industry. Thus our figure of fixed capital cost per adult male worker is the value of machines in large firms (to get round the problem of under- utilization of looms in small fi=n4 plus the value of land and building in small firms plus the value of land and building in large firms. This gives a trtal figure of Rs. 1693 per adult mAle workers. (ii) Working Canital We now come to the question of working capital used in the production of handloom sarees. The organization of production in either type of unit discussed earlier is of relevance here. The master weaver advances raw materials (and perhaps wages as well) to the producers who get paid at the stipulated rate per saree produced, as and when they deliver the finished saree to the master weaver. The cost of the working capital involved is borne largely by the master weaver. The amount of such ca7ital required is a function of the period of production, i.e. the time the goods are in the pipeline. A.K. Sen, in his discussion of the problem included all recurrirgcosts - wage bill and material costs - in -9- the requirement of working capital. He commented: "If the average lag is one year, the whole of the year's recurring costs will be locked up as working capital. If the average lag is six months, with a working capital equal to half the years' recurrirgcosts we can meet all the recurring cost requirements."- In the accounting data obtained from a typical industrial firm we do not have to consider the period of production, because the statistics on raw materials and semi-finished products in stock, as well as cash in hand together add up to the working capital required for the rate of production actually observed. For the handloom weavers'units surveyed by us, however, no such statistics for working capital can be clearly identified. The small independent units did not provide any data on stocks of raw material or products carried. They, however, did indicate the amount of yarn obtained from the master-weaver which was processed in their homes. The lag involved in pro- ducing the cloth was stated to be 1-2 months. The larger master-weavers' establish- ments did have stock data, but it is not clear if the stocks of materials advanced b: them to their workers (and presumably retained in the latter's homes) were included in the figures given to the interviewers. We try to see how figures on working capital derived from the different approaches compare. Table 4 gives three sets of estimates - two for the large unit4 ane based on reported stock data and the other on the basis of recurring costs. A third estimate is for small firms based on materials and wage costs. The three estimates are not too widely apart and we have already mentioned that the stock data for larger firms probably underestimate working capital. 1/ A.K. Sen Choice of Techniques, Oxford, 1968, p. 101. Table 4: ALTERNATIVE MEASURES OF WORKING CAPITAL Material Cost Wage Cost Total 1/6 of (3) Working Capital Current Current per Adult Male Asset Assets per Worker Adult Male Worker (2) (3) (4) (5) (6) (7) Large firms (per unit) 306,000 135,000 441,000 73,500 1,044 53,250 756 Small firms (per unit) 6,646 3,660 10,306 1,718 886 I- O - 11 - An acceptable figure for working capital requirement in handloom weaving would probably be Rs. 900 per adult male worker. (c) Value Added Der Worker Having derived approximate value for wages and capital per worker in handloom weaving industry, it remains to estimate value added per worker. The large units do provide data on value added in the statement of accounts given to the interviewers. From these data the value added per worker was found to be Rs. 251 as against average wage of Rs. 161 per worker per month. The difference between the two - wages according to these data are 64 per cent of value added - account for payments to capital and land, and for profits of the master weavers. There is some difficulty in coming to a conclusion on the appropriate figure for value added per worker for the small units. The difficulty arises from the fact that the ownership of capital is split between the small producers and the master weavers - the former own ing the fixed capital and the latter the working capital. Furthermore, both groups enjoy a share of entrepreneurial profits - though it might be concluded that the more strategic business position occupied by the master-weavers allows them to enjoy a larger share of such profits. The problem of separating out factor payments accruing to different sets of people is an almost insuperable one when the business is,in fact, being carried out by both in an interlinked way. The best course open to us would seem to be to take the earnings of the small weavers as pure wage payment and arrive at a figure on value added by applying the same wage/value added ratio as observed for the master-weavers' firms. The fact that the earnings of the small weavers and the wages of the workers hired by the master weavers are more or less at the same level lends support to this procedure. - 12 - Thus the value added per adult male worker in handloom weaving - taking both types of firms together - is calculated at Rs. 248 per month. (Rs. 2,976 pet adult). (d) Conclusion The cost data for handloom weaving is summarized in Table 5. Table 5: Inputs Recuired to Produce Value Added of Rs. 10,000 per Annum in Handloom Weaving, CaDital (Rs) Adult Male Fixed Working Total Wage Per Workers Machinery Land & Building Total Worker (Rs) 3.36 534 5,155 5,689 3,024 8,713 1,908 -13- III Relative Costs in Handlooms and Powerlooms The cost data available from our survey of Powerlooms in Mau needed much less doctoring. The relevant statistics are available directly from the balance sheets of the firms collected during the survey. Table 6 sumarizes the data (Because of the large inter firm variability geometric means are used). Table 6: Inputs Required to Produce Value Added of Rs. 10,000 per Annum Powerlooms Wage Adult Male Fixed Capital Working Total per Workers Machinery Land & Building Total CaDital CaDital Worker (Rs. 000) Rs. (000) Rs.(000) Rs. 1.76 3,426 4,480 7,906 2,600 10,506 2,136 One point of adjustment about the number of workers required for production should be noted. Many of the powerlooms in the town of the survey sufferred from power shortage. Thus although the number of shifts the plants were planning to work was generally 1.5 - involving 12 hours of work - the actual number of hours worked were smaller. The stock of labor employed reported by each firm gives the expected number of man-hours worked if each of the workers were employed for an average of 1.5 shifts. The actual uumber of man-hours fell short of this amount. Since labor was paid on piece- rates, based on the actual amount produced, the flow dlabor time involved in the production of the actual output observed, has to be adjusted downwards from the expected flow. Two sets of wage data available in our survey enable us to do this adjustment. We have, first,. the average wage per worker actually paid based on the wage bill for the year given in the balance sheet of the - 14 - establishment. Secondly, we have a figure of monthly wage for the workers employed in the establishment (appropriately weighted by major occupational categories based on the expected level of output. The ratio of the former to the latter is 0.88 and would be used as an index of underemployment. The reported stock of labor per unit is multiplied by this factor to give the f"effective" rate of man-years of labor used by the unit. Private Profitabilit7 The private profitability of powerloom weaving compared to handlooms might be assessed if we ask the question: does it pay the master weaver who is employing handloom workers on a "putting out" system to switch to production on powerloom? Data given in Tables 5 and 6 show that the two processes are points on a well-behaved isoquant - production in powerlooms require more capital per unit of value added, but less labor. It is also seen that wages per adult male workers are very close together in the two industries - the mean could be taken as a rounded figure of Rs. 2,000 per annum with only about 5-6 per cent difference in either direction. Thus whether or not the switch to powerlooms will be profitable fog the master weaver depends on the rate of interest entering his calculation. The question can be posed in cost-benefit terms: what is the internal rate of return which equates the increase in capital cost to the present value of the saving - in wage bill? The master weaver, as discussed earlier, bears the cost of the machinery and of working capital counting to Rs. 3,558 (per Rs. 10,000 of Value Added), but not of the land and buildings of the weaver's residence in which the looms are located. Thus for the same amount of value added, a switch to powerlocms would mean an additional capital cost of Rs. 6,948. The saving in wage bill could be 1.6 times Rs. 2,000 per annum. If we work with an expected life 6f the capital stock of 25 years, the internal rate of return which makes the master weaver indifferent as between the two processes is - 15 - 46.05 per cent. An effective interest rate higher than this would mean the master weaver has no incentive to shift to powerlooms. Since widespread co-existence of powerlooms and handlooms is observed in Mau, the interest rate implicitly accepted by entrepreneurs can be expected to hover round this figure. Social Profitability To assess social profitability two additional considerations become important (i) Some cost must be attached to the land and building devoted by the handloom weaver to the production process, although it is doubtful if the full value at market rates (as given in Table 5), is appropriate. Some of the area in the owner's residence assigned to production must have a joint function. (ii) The social cost of labor used in the two industries is an even more complicated issue. We saw that the average wage per adult male worker was roughly the same in the two industries. But employment in handloom weaving offers two important advatanges. First, the variable hours, and the ability on the part of the workers to fit in work on the looms with their domestic activities (and incidentally, save on travel to and from work). Work on the powerlooms, in addition, has the additional disadvantage of being affected by interruptions in the supply of power, so that workers have to adjust their work time to a pattern dictated by the availability of power. Secondly, the difference between the two sectors in the employment of other family members is an important issue to be taken into consideration. Employment in handloom weaving is forthe whole family. It will be recalled that in our sample the mean employment in handloom unit consisted of 1.9 adult males, 1.12 adult females (of which half worked part time), and 1.42 children - 16 - (part time helpers). The total income accrues to the family as a whole. In the data presented in Table 5 the wage has been calculated by dividing this total income by the number of adult male workers only.. The rationale for this procedure might be that in a regression analysis of earnings among a cross :section of weaving units, the contribution of auxilliary workers (including women and children) was found to be small.-/ But the presence of auxilliary workers does increase the productivity of the adult male workers. Thus to attribute the whole of the earnings to male workers is an over estimate of their marginal productivity. There is a further point about male worker from a handloom weaver's family if he chooses to seek work as a wage earner in the powerloom industry. As a family worker the weaver is assured of employment in the household enterprise. The probability of his getting work in powerlooms will be less than unity depending on the rate of unemployment among powerloom weavers. We can conclude, therefore, that in equilibriam the earnings of an adult male worker in handlooms should be below the level in powerlooms. How then do we explair. the (more or less) observed equality in average earnings per male worker in the two sectors. These seems to be two answers: (a) The skill mix of workers in powerlooms is slightly different, since some semi-skilled (e.g. winders) workers are used along with weavers. The earning of powerloom weavers by itself is about 15 per cent higher. (b) More importantly the earnings of handloom weavers include the return to land and building owned by the household and which, as we have seen, account for the major part of the fixed capital stock of the enterprise. Thus to calculate the relative social profitability of handlooms and powerlooms, an appropriate part of the cost of land and building has to be added to the capital cost of production in handlooms, but at the same time the average wage of handloom weavers should be reduced by some fraction. 1/ The results are discussed in another paper. See Study No. 3 in this series on the Indian Textile Industry. - 17 - Unfortunately, we do not have any basis for making quantitative assessments, for either type of adjustment which is needed. The best we can do is to get a faelfor the orders of magnitude of the internal rate of return involved in a switch from handlooms to powerlooms under alternative assumptions about the extant of the adjustment to capital and wage costs. Adding part of the cost of land and building to the capital cost of handloom production means that the saving in capital cost by switching to powerlooms becomes smaller, so that the same saving in the stream of wage bills over the lifetime of the machine has to be discounted at a higher rate of interest to achieve equality of costs and benefits. On the other hand, Table 7: Internal Rates of Return in Switching from Handloomsto Powerlooms Percentage of Land Percentage of Annual Wage for Male Worker Subtracted and Building Costs from Average Handloom Wage Added to Handloom Capital Costs 25 33.3 50 25 40.45 35.92 26.90 50 52.40 46.54 34.67 75 74.30 65.99 49.53 valuing the wage of handloom weavers at a lower rate than their observed earnings implies a lower internal rate of return at which the two processes of weaving are valued equally. The message of the numerical example of - 18 - Table 7 is that from the social cost-benefit point of view the switchover interest rate is generally higher than when we are concerned only with private profitability for a wide range of adjustments to the capital and labor costs. That is to say, while a rate of interest of more than 40 percent is needed to make handlooms profitable for a private entrepreneur we probably need a significantly higher interest rate to make it socially profitable. Even if only a small part of land and building costs are added to handlooms and the handloom shadow wage is substantially lower than the earnings per male worker, the "switchover" interest rates, although lower than the private profitability rate, does. not fall to very low levels (see the figures marked out by a solid line in the North-East part of the Table). The Significance of Shadow Wage. It should be emphasized that the considerations involved in the valuation of the social cost of an adult male worker in handloom weaving do not include an assessemnt of the general rate of un-or underemployment in the labor market in question. We may or may not want to promote mere labor intensiv-imethods of production based on judgments about the potential rate of unemployment. The calculations undertaken above were based on the much more narrow argument that the social cost of labor in handloom weaving should take account of help received from auxilliary household workers, and of the lower disutility,cost of labor in a household setting. - 19 - Relative Costs in Mills and Powerlooms Our borehole work did not attempt to collect cost - data from mills. Any such attempt would be time consuming and difficult. It will have to face up to the basic problem of separating cost data fro= the composite (vertically integrated) mills which cover both spinning and weaving operations. It was decided to make broad comparisons from studies which have recently been undertaken. Three sets of data could be identified: (i) There is a study of relative costs by M. Padmanabhan for I/ 1974 for weaving coarse medium grey cloth ( 34s x 34s counts)-. The cost data were derived from relevant published and unpublished sources available with organizations connected with the Textile trade, rather than from a sample survey. (ii) The Planning Comission of the Goverrment of India put together a study under the direction of Professor Raj Krishna in 1979. Again the basic data were in the nature of "most informed" estimates culled by the staff of the project from a variety of sources including interviews with selected establishments.2 (iii) An attempt to quantify the relative costs of producirLg coarse cloth was made by a working group of the National Industrial Development Corporation Ltd. The data seemed to be mainly derived from the technical 1/ M. Padmanabhan, "The Sick Mill Problem in the Indian .Cotton Textile Industry, Ph. D. Thesis, Bombay, 1974, Appendix I. 2/ Planning Commission, Government of India, Project Appraisal Division: "Choice of Technology in the Textile Industry," prepared by Miss Ritu Anand under the guides of Nitin Desai and Raj Krishna. 3/ Paper presented on the "Appropriate Technology for the Production of Textiles" (Working Group No. 3) to the International Forum on Appropriate Technology sponsored by UNIDO in New Delhi in November 1978. - 20 - experience of Textile engineers, particularly those working in the Ahmedabad Textile Industry's Research Association (ATIRA). There is unfortunately no systematic survey of the economics of Cotton Textile Weaving in the large scale sector in India, even confined to a small region of the country. The attempt at quantification of costs of production which follozmakes use of all the three sources cited above, and also refers to the statistics reported in the Annual Survey of Industry of the Indian Government (although these refer to registered spinning and weaving mills together). The estimate of relative costs of producing grey cloth of medium count (no finer than 40s x 40s count) will be built up step by step. A. Capital Cost per unit of Value Added (a) Cost of Loom We will first consider the non-automatic looms which are most usually used by Indian weaving mills. (A reference to automatic looms which are not so common will be made towards the end of this section). The looms are essentially the same as used by powerlooms, although these is a great deal of difference in the conditions of the looms used in the two sectors. The powerlooms used in the small-sector are older, often bought second-hand, and less well maintained. There is then a substantial difference in the price of a non-automatic loom used in the two sectors. Padmanabhan's study gives the ratio of the price of looms in mills to that in the decentralized units as 3:1. The Planning Commission study gives a higher ratio of 5:1. We will take the average of the twr, for our calculations - 4:1. - 21 - (b) Productivity per loom Productivity in each sector depends on the number of shifts worked per day, and secondly, the output per loom shift of the type of cloth being woven. The relevant data from the two studies cited are as follows: Powerlooms Mills 1. Padmanabhan No. of shifts 1.5 2.25 Output in meters per loom shift 27 34 Total output in meters per day 40.5 76.5 2. Planning Commission Production, Annual Output in meters 10,000 25,000 The two sets of data are fairly close together noting that the Planning Commission figures include difference in shifts per day. Taking into consideration the, fact that shifts in the mill sector have been increasing continuously, we adopt the Planning Comission figures as representing overall difference in physical productivity per loom. (c) Value Added Mill cloth is priced somewhat higher. The Planning Commission study gives the value added per meter as 1.018 per powerloom and 1.201 for mills - i.e. 20 percent higher for mills. Note that this price difference represents intrinsic difference in the quality (e.g. durability) of the cloth produced, and not the effect of excise duties imposed on mill made cloth. -'22 - Thus the value added per loom is 2.7 times in the mill sector compared to powerlooms. The loom capital - output ratio in the mill sector is thus (4 - 2.7) or 1.48 times the level of the ratio in the powerloom sector. (d) Non-loom capital costs So far we have concentrated on the costs of loom relative to value added. But there are other elements of fixed capital costs in both sectors consisting of the value of preparatory weaving machinery and of land and building. Secondly, there is working capital to be taken into account. In the World Bank Survey of powerlooms, loom cost accounted for no more than 30% of fixed capital (i.e. including land and building) and 23% of total capital (i.e. including working capital). It is appropriate to assume that the proportion of fixed capital other than loom would be a larger proportion of the total in the mill sector than in powerloom (Nothing very much can be said about the relative importance of working capital without much more intensive enquiry). Thus capital value added ratio would be higher in the mill sector by a factor significantly higher than 1.48. This is borne out by figures available from two other studies. In our own survey the capital - value added ratio in powerlooms in Mau was 1.05 - taking geometric means The Government of India Annual Survey of Industries for 1977-78.gives the capital - value added ratio in the factory sector at 1.91 (although admittedly it includes all cotton textile firms including spinning and weaving). In the study by the Working Group on Textile Technology, figures are given on capital costs - including machines other than looms. The capital - output ratio for mills in this study comes to 2.05 times the ratio for powerlooms.1/ 1/ UNESCO, op. cit., Table 9, p. 36. - 23 - In all probability, therefore, we shall not be far wrong if we take the ASI estimate that Rs. 19,000 of the value of capital is required to produce Rs. 10,000 value added in the mill sector. 3. Employment - Value Added Ratio We next need to calculate the employment per unit of value added in the two sectors. This is a function of two terms: value added per loom as has already been discussed; and the number of workers per loom shift. From Padmanabhan's data the value added per loom shift ratio between mills and 34 powerlooms is equal to (I-x 1.36) or 1.71. The number of looms looked after 27 by a worker in a weaving mill varies between 2 and 4, so that we can take the average as 3. In powerloom units, according to the Bank Survey the number of looms is.1 - 2, with a mean value of around 1.5. We assume that the ratio of other workers to weavers are the same in the two industries. Thus the ratio of employment per unit of value added in mills to that in powerlooms = 0.5 i- 1.71 - 0.29. C. Relative Wage Levels For the analysis of relative profitability in mills and powerlooms, it remains to consider relative wage levels in the two sectors. The skill mix of the workers in the two sectors would be different. But since we are working with the average number of workers per unit of Value Added needed in the two sectors, the relevant concept is average wage per worker. The most comprehensive data on wages are available (for the country as a whole) from the Annual Survey of Industry. The relevant data for 1977-8 are given in Table 8. Table 8: Average Earnings per Annum, 1977-8 Wage Earners Salaried Workers All Workers Powerlooms 2,545 3;136 2,647 'Mills 5,854 16,713 7,027 -24- The figure shows very substantial difference in the levels of earnings in the two sectors. Table 9 brings together estimates of the differential in earnings per worker reported in other studies. Table 9: Difference in Average Earnings Between Mills and Powerlooms (Powerloom Earnings = 100) ASI, 1977-8 Planning Commission Padmanabhan SITrA'l 1979 1974 1978 Wage Earners Only All Workers Weavers Only Weavers only 230 265 371 195 200 It is seen that the ASI statistics on the earnings differential between the two sectors lie somewhere in the middle of the range reported in the various studies. For our purposes we shall select the ASI figure for wage earners only (230) as a measure of the differential. The World Bank Survey reported that the average earnings per worker in powerlooms was Rs. 2,136. Applying the ASI ratio the average earnings per worker in mills comes to Rs. 4,913 per worker. D. Private Profitability Consider an entrepreneur in the mill sector weighing the benefits of switching to powerloom production. The data given above gives the saving in capital cost of producing value added of Rs. 10,000 in powerlooms rather than mills as,(19,000 - 10,506) or Rs. 8,494. But there is an increase in wage cost, because although wage rates are lower in powerlooms, the number of workers needed to produce the given value added of Rs. 10,000 is 1/ South Indian Textile Research Association: A Study of the Handloom Sector to Increase Productivity, Coimbatore, August 1978. - 25 - sufficiently larger to lead to a higher wage bill. The increase in labor cost has to be compared to the saving in capital cost to judge the net profitability of the two processes. As in the case of the comparison of handlooms and powerlooms we calculate the rate of interest which will make the present value of the increase in wage cost equal to the saving in capital cost. (The assumption about the life of the machines is the same as before - viz 25 years). The increase in,wage .bill = Rs1.7 x 2,136 - 0.5 x 4,913) - Rs. 1,175 The rate of interest which equates this wage stream for 25 years to the saving in capital cost of Rs. 8,494 in 13.2 percent. At rate of interest higher than this it will be profitable for an entrepreneur to switch to production with powerlooms rather.than non-automatic looms in the mill sector. E. Importance of the Wage Differential and the Apropriate Shadow Wage The above calculation is clearly very significantly affected by the wage differential. For private profitability calculations the differential can be taken as given. But for social cost benefit analysis it is impossible to avoid coming to judgments about the shadow wage to be applied to the calculation of relative costs in .the two sectors. For example, if we make the judgment that the lower wages in the powerloom sector really represent the appropriate social cost of using labor in either sector, the switch value of the rate of interest becomes quite different. At the powerloom wage the increase in wage cost involved in a switch from mills to powerloom = Rs. (1.7 - 0.5) x 2,136 = Rs. 2,563. - 26 - The rate of interest which equate this stream over a period of 25 years to the saving in capital cost (Rs. 8,494) is 30.13 per cent. V Conclusions We can now draw together the overall picture for the three sectors in cotton weaving. The "best estimates" of capital and labor requirements in mills, powerlooms and handlooms for producing value added of Rs. 10,000 are qualified in Figure I. The data show a well-behaved isoquant with the variations in capital intensity being in the generally expected directions. None of the three processes are technically dominant in the sense of making more of both less capital and less labor er unit of value added. Note that two alternative points are plotted for handlooms, point H representing only the cost oE machinery plus working capital,while point H' adds 75% of the value of land and building to capital costs. The former as explained in Section II, is relevant for private profitability calculations, but for social profitability the relevant point will most likely be somewhere between H and-H'. It was seen that the wage per adult male worker was very close together for handlooms and powerlooms. Hence the cost per unit of labor faced by a private entrepreneur will be similar for these two processes (Auxilliary workers helping the adult male worker are excluded from the labor units measured along the X - axis in figure 1). But wages in mills are nearly 2 times that in powerlooms. Labor units required in powerloom for the sawe value added per worker are higher than in mills b., a factor larzer than this, so that labor costs ar still higher in powerlooms at the same time that capital costs are lower. Assuming a time horizon of 25 years the boundary (or "switchover") rate of interest could be calculated at which the savings in capital costs were just balanced by the present value of the increase in wage bill. For the choice 盯 27 between mills and powerlocms this switchover interest rate is 13'.2 percent, and for the choice between powerlocms and handlocms it is 46.03 percent. Thus, given the wage levels prevailing in the three sectors, powerlooms are profitable within this rather wide range of interest rates. Of course, private entrepreneurs face different levels of interest rates in the three sectors as well as different wage levels. It.is arguable that the interest rate in the informal capital market will be not -vrery far apart for handlooms and powerlooms, but will be considerably lower for the formal capital market to which only the large-scale mill sector has access. In other words, capital market segmentation follows much the same line on labor market segmentation with a large price difference between mills and powerlooms, and very small differential between the powerloom and handloom sectors. If, in fact, the interest rate in the fo-rmal capital market is lower than 13.2 per cent, and the rate in the informal capital market is higher than 46.05 percent, the powerlocms will not be privately profitable. Entrepreneurs with access to the formal capital market would prefer mill production, and those operating in the informal capital market would find handlooms profitable. But the very large difference in interest rates needed to produce this result should be emphasized. The magnitude involved suggest that the high growth rate of production in powerloom in India in recent years may be largely due to pricing factors rather than quantitative restrictions of the type described in the earlier study. The calculation of relative social profitability of the three techniques is crucially dependent on the level of the shadow wage used. Th is topic will be treated more enctensively in a separate study. For the moment, we assume that the powerloom wage is the appropriate shadow wage for mills and powerlooms, but the shadow wage for handlooms is lower (to an unknown degree) due to the advantages of wcrking in a household industr7 - 28 - and the help received from auxilliary workers. The other point we must take into account for social profitability analysis is, as already mentioned, the addition of some of the land and building cost to the total capital cost of handlooms. As an example, if 50 percent of land and building cost is added to the cost of handlooms in this way, the "'switchover" interest rates between the three techniques will be as in Table 10 for various values of the shadow waee in handlooms. Table 10: "Switchover" Interest Rates for Various Values of Shadow Wage Handloom Shadow Wage as a Percentage of Market Handloom Wage 75 percent 66.6 percent 50 percent Eandlooms to Powerlooms 52.03 46.18 34.67 Powerlooms to Mills 31.6j 31.63 31.63 Note: The shadow wage for mills and powerlooms is the same and equal to the powerloom market wage. The rate of interest at which production in the mills becomes profitable is increased enormously if we adopt the (lower) powerloom wage as the shadow wage of labor used in both the mills and the powerloom sectors. The result might appear counter intuitive because the effect of the use of a shadow wage lower than the market wage in the mill sector leads to the prediction that the more capital intensive sector is more profitable. This is because we are considering movement between sectors with a large wage differential, and the use of the same shadow wage to both sectors is tantamount to renoving the effective tax on labor in the mill sector. Two 929 - 29 - points should be noted. First, the variability of capital intensity within the mill sector is not considered. Secondly, the shadow wage might indeed be different between mills and powerlooms depending on the causes of the wage differential. Indeed, taking alternative opportunities and distributional considerations into account the shadow wage in powerlooms might even be lower than the market wage in this sector (which will have the effect of reducing the switchover interest rate between powerlooms and mills). Another result to note in Table 10 is that the range of interest rates within which production in the powerloom sector is socially profitable is very much narrower when we are using social price of labor than in the case when market wage rates were used. In fact, the last column of Table 10 shows that if the shadow wage in handlooms were 50 percent of the market wage, the switchover interest rates are nearly the same. With an even lower handloom shadow wage, or if a smaller proportion of land and building cost were added to the capital costs of handlooms, the interest rate relevant for the switchover from handlooms to powerlooms drops below that for the switchover from powerloom to mills - and powerlooms drop out as an economically viable alternative from the social point of view.