Measurement of productivity growth, efficiency change and technical progress of selected capital-intensive and labour-intensive industries during reform period in India.

Abstract

This study examines the total factor productivity growth (TFPG) and its components TE (technological progress) and TEC (technical efficiency change) in ten manufacturing industries, classified them into capital-intensive and labour-intensive industries (five in each segment) using annual time series data for the period 1994 to 2004. The TFP growth is estimated by applying Malmquist Productivity Index (MPI) on the panel data of aforesaid segments separately. The study finds that the average TFP growth in the capital-intensive industry segment grew at a moderate rate of 1.7 per cent per annum during the entire study period, whereas, its counterpart, selected labour-intensive industries have shown a productivity regress, it is -0.9per cent. The decomposition of TFP improvement into technical efficiency change (catching-up effect) and technological progress (frontier shift) reveals that the TFP growth is primarily contributed by technological progress rather than by technical efficiency change in capital-intensive industries whereas in labour-intensive industries low growth of technical efficiency (0.5 per cent) has been offset by a higher rate of decline in technological progress. The results are, by and large, useful for policy makers in designing industrial policies.

I. INTRODUCTION

Industrial performance has been a subject of debate in India since the early 1950s. The Assessment of industrial performance after, adoption of new liberalized policies since 1991, in view of linkages between trade liberalization and productivity growth has gained important among academicians and policy makers. Theoretically, trade liberalization could have both a positive as well as negative impact on productivity (Tybout, 2000) (1), but recent surveys by Tybout (2000) and Epifani (2003) enlighten that the empirical literature generally support a positive effect of trade liberalization on productivity. Thus the effect of trade liberalization on productivity is an empirical question. Therefore, the present paper explores the impact of trade liberalization on Total Factor Productivity (TFP) and it sources in selected Indian manufacturing industries.

The extent of impact of liberalization may vary across different industries. In fact, different forms of industries could demonstrate different reactions to environmental changes. Hence, the impact of trade liberalization could vary across industries. Therefore, whether all industries benefited or suffered equally from the new economic environment is an important issue to be investigated. In this study we made an attempt to find out sources of productivity growth by using Data Envelop Analysis (DEA) based Malmquist Productivity Index (MPI) to estimate TFP growth and its decomposition into technical efficiency change and technological progress.

The productivity growth is considered as an indicator of sustained economic growth and improvement in standard of living. A reasonable standard of living, typically, defined as real GDP per capita, can be influenced by a number of factors including; changes in employment/population ratio, changes in terms of trade and or changes in productivity. While improvement in either of these factors result in a higher standard of living, employment/population ratio and, to a lesser extent, terms of trade have upper limits and therefore can impact living standard only in the short run. In the long run, the only sustained manner to increase per capita GDP is by increasing the amount of output produced by a given quantity of inputs, that is raising Total Factor Productivity (TFP). Higher levels of output relative to given inputs generally translate into higher returns to factors of production. An improvement in living standard is the most well-kwon benefit of productivity gain and productivity growth that can also provide to an economy as well. More competitive business and higher employment level can result from increasing productivity.

At the industry and firm level, productivity gains relative to it competitors enhances an industry's or a firm's competitive position allowing it to increase profit margins or sell products cheaper. Lower products prices relative to its competitors would allow it to expand production and gain market share. For industries producing products in highly competitive market conditions, productivity gains are often crucial just to survive. Industries are all striving to obtain a competitive edge and those industries do not pursue productivity gains are unlikely to survive in the long run. Another potential benefit of increasing productivity is expanding economic activity and employment growth. If a company or industry has higher rates of productivity growth relative to other competitors, it is likely that the industry or company develops a competitive advantage over time. Such a competitive advantage may likely attract new investment as industry expands output to take the advantage.

Technological progress and technical efficiency change are the two sources of productivity growth. A study of these sources are crucial for identifying the factor that are responsible for productivity stagnation and for adopting appropriate measures at company, industry or government levels to improve productivity.

This study is organised as follows. Section II briefly discusses productivity measurement approaches and section III describes the data and methodology. The empirical results are interpreted and discussed in section IV. Concluding remarks are made in section V.

II. PRODUCTIVITY MEASUREMENT APPROACHES

There are two measures of productivity, namely, partial or single factor productivity (SFP) and total factor productivity [TFP]. Partial factor productivity is calculated by dividing the total output by the quantity of an input. The main problem of this measurement of productivity is that it ignores the fact that the productivity of an input also depends upon the levels of other inputs used. For example, a higher dose of capital application may increase the productivity of labour even when other inputs including labour remain constant. The TFP approach over come from this problem by take into account the level of all inputs used in the production of output. TFP is defined as the ratio of weighted sum of output to the weighted some of inputs. In other words, the TFP approach measures the amount of aggregate output produced by a unit of aggregate inputs. MFP is deemed to be the broadest measure of productivity and efficiency in resource use. It aims at decomposing changes in production due to changes in quantity of inputs used and changes in all the residual factors such as change in technology, capacity utilisation, quality of factors of production, learning by doing, etc. An increase in TFP, therefore, implies a decrease in unit cost of production. Since TFP incorporates all the residual factors, it has also been dubbed as an 'index of ignorance' [Abramovitz, 1956]. However, the concept of TFP scores over SFP. It is observed that all factors affecting the production process are captured in the former concept unlike that in the latter. Therefore, the study used TFP as a measure of productivity.

Over the last three decades, researchers have developed several theories and method of TFP measurement. The two main approaches applied for the estimation of growth in TFP are the Production Function Approach (PFA) and the Growth Accounting Approach (GAA). Before the mid-1990's most studies estimated TFP growth by growth accounting approach (Ahluwalia, I.J. 1991, Balakrishnan, P. and K. Pushpangadan 1994, ICICI Limited 1994, Rao, J. M. 1996a, Rao, J. M. 1996b and Pradhan G. and K. Barik 1998) in India. But both approaches have some limitations. One of the major disadvantages of using PFA is the problem of identification of production function due to the simultaneity in determination of input intensities and output levels. The problems of autocorrelation and multicollinearity encountered in the use of PFA vitiate the empirical estimates obtained by this approach. To massaging the data in order to take care of these statistical problems render it difficult to interpret the empirical results. The assumption of 'well-behaved' production function takes away flexibility and the ability of Translog production function to approximate a non-homothetic production structure. On the other hand, the limitation of GAA is that, if the share of capital is treated as a residual, it implies the assumption of constant returns to scale. Moreover, if output elasticities are proxies by the observed factor shares, it implies the assumption of a competitive market structure. It is also assumes that an industry operates on its production frontier, implying that it has 100 per cent technical efficiency. Thus, TFP growth measured through this approach is due to technical change, not due to technical efficiency change (Mawson et al., 2003). It has been well documented in the literature [Rao, 1996a] that both PFA and GAA assumed a well-behaved production function, stability of the production function over time and cost minimisation, which is a sub-goal of profit maximisation. In recent years, stochastic frontier analysis and Data Envelop Analysis (DEA) -based Malmquist Productivity Index (MPI), which uses panel data, have become popular approaches for estimation of TFP growth. These approaches do not assume that all production units operate at 100 per cent technical efficiency.

Among these, the most popular approach to measuring productivity changes is based on using Malmquist Productivity Indexes-a method originated by Caves et al. (1982). According to MPI approach, TFP can increase not only due to technical progress (shifting of production frontier) but also due to improvement in technical efficiency (catching-up). According to Grifell-Tatje Lovell [1996], the Malmquist index has three main advantages. First, it does not require the profit maximization, or cost minimization, assumption. Second, it does not require information on the input and output prices. Finally, if the researcher has panel data, it allows the decomposition of productivity changes into two components (technical efficiency change, or catching up, and technical change, or changes in the best practice). Its main disadvantage is the necessity to compute distance functions. However, the Data Envelop Enalysis (DEA) technique can be used to solve this problem.

Another important advantage of Malmquist index is that, it allows us to distinguish between shifts in the frontier (technology change, TC) and improvements in efficiency relative to the frontier (efficiency change, TEC), which are two mutually exclusive and exhaustive sources of total factor productivity change (TFPC). It is also possible to decompose efficiency change into its distinct components with Malmquist index; changes in management practices (pure efficiency change, PTEC) and changes in production scales (scale efficiency change, SEC). This treatment ideally improves analytical efforts while tracing the underlying sources of productivity developments.

MPI is Data Envelop Approach (DEA)--based approach but unlike DEA that is static in nature as it assesses the productivity of a firm or an industry in ration to the best practice industries in the given year, it also accounts for the shift of production frontier over time. Since it is capable of decomposing productivity growth into technical efficiency change and technological progress, it is able to shed light on the mechanism of productivity change (Ma, et al., 2002). The DEA-based MPI method was initially introduced by Caves, Christensen and Diewert (CCD) in 1982 and was empirically used by Fare, Grosskopf, Lindgren and Roos (FGLR) in 1992 and Fare, Grosskopf, Norris and Zhing (FGNZ) in 1994. Since then several versions of MPI have been developed.

III. DATA AND METHODOLOGY

The study is based on panel data collected from various issues of Annual Survey of Industries (ASI), Central Statistical Organisation (CSO), Ministry of Statistics and Programme Implementation, Government of India, New Delhi, for the period 1994 to 2004. In this study output is measured in gross value added, labour is measured in terms of the total number of persons employed and capital is measured in gross fixed capital.

Malmquist Productivity Index is explained by distance or technical efficiency functions. One feature of distance functions is that these allow description of multi-input, multi-output production technology without the need to specifying a behavioural objective, such as profit maximization or cost minimization. Distance functions are two types; the input distance functions and the output distance functions. Input distance functions look for 'by how much can input quantities be proportionally reduced without changing the output quantities produced?' On the other hand output distance functions addresses 'by how much can output quantities be proportionally expanded without altering the input quantities used?' The Malmquist method is most commonly used for output comparisons. Hence, in this paper we adopt an output-oriented distance function approach.

In order to specify the Malmquist index using output-oriented distance functions, the paper first define the relevant technology. Let, [y.sub.t][epsilon][R.sub.+.sup.m] and [x.sub.t] [epsilon][R.sub.+.sup.n] denote an (Mx1) output vector and an (Nxl) input vector, respectively. Then the diagram of the production technology in period t is the set of all feasible input-output vectors given as, [GR.sub.t] = {([y.sub.t], [x.sub.t]): [x.sub.t], can produce [y.sub.t]} (1)

Where, the technology is assumed to have the standard set of properties, such as convexity and strong disposability of outputs.

The output set, [P.sub.t]([x.sub.t]), which represents the set of all output vectors, y, are defined in terms of [GR.sub.t] as follows,

[P.sub.t]([x.sub.t]) = {[y.sub.t]: ([y.sub.t], [x.sub.t]{([y.sub.t], [x.sub.t])[epsilon][GR.sub.t]} (2)

The output distance functions for period t technology is defined on the output set [P.sub.t]([x.sub.t]) as,

[d.sup.t.sub.o]([y.sub.t], [x.sub.t]) = inf {[[partial derivative].sub.t]: [[y.sub.t] / [[partial derivative].sub.t]][epsilon][p(x)} (3)

where the subscript '0' indicates 'output-oriented' measure. The notation [d.sup.t.sub.o]([y.sub.t], [x.sub.t]) stands for the distance from period t observation to the period t technology. In other words, this efficiency function represents the smallest factor, [[partial derivative].sub.t], by which an output vector ([y.sub.t]) is deflated so that it can be produced with a given input vector ([x.sub.t]) under period t technology. Similarly, [d.sub.o.sup.t+1]([y.sup.t+1], [x.sup.t+1]) would indicate distance from period t observation to the t + 1 technology.

Using these definitions, the following Malmquist TFP index can be constructed to measure productivity change between periods t and t + 1, based on t technology,

[m.sup.t.sub.o]([y.sub.t], [x.sub.t][y.sup.t+1],[x.sup.t+1] = [d.sup.t.sub.o]([y.sup.t+1], [x.sup.t+1]) / [d.sup.t.sub.o]([y.sub.t], [x.sub.t])(4)

A similar output oriented Malmquist index can be obtained, based on the t + 1 technology as follows,

[m.sup.t+1.sub.o]([y.sub.t], [x.sub.t][y.sup.t+1], [x.sup.t+1]) = [d.sup.t+1.sub.o]([y.sup.t+1], [x.sup.t+1]) / [d.sup.t+1.sub.o]([y.sub.t], [x.sub.t]) (5)

Equations (4) and (5) imply that estimation of TFP change between the two periods could depend on the choice of technology. In order to avoid the effect of any arbitrarily chosen technology, one has to measure change between two data points relative to a common technology. Thus, following Fare et al. (1994) the output oriented TFP could be estimated as the geometric mean of the indices based on period t and t + 1 technologies as given by equations (4) and (5) respectively. Hence we have,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

When the value of [m.sub.o] exceeds one indicates a positive total productivity growth from period t to t + 1 and a value of the index less than unity indicates a decline in TFP growth. If the value is equal to unity indicates no change in TFP.

To measure source of TFP growth, technical efficiency chance and technological change, the equation (6) can be written as,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

The ratio outside the square brackets measures the change in the technical efficiency change (TEC) between periods, t and t + 1, i.e., moving closer to frontier or 'catching-up'. In other words, efficiency change is equivalent to the ratio of Farell technical efficiency in period t + 1 to the Farell technical efficiency in time t. the remaining part of the index that appears inside the square brackets indicates the technological change (TC) or shift in frontier between two periods, evaluated at [x.sub.t] and [x.sub.t+1]. Hence the Malmquist Index given by equation (7) shows that productivity growth is the product of technical efficiency change (catching-up) and technological change. Hence, TFP change can be written as:

TFPC = TEC x TC

Both the above indices can be interpreted as progress, no change and regress when their values are greater than one, equal to one and less than one, respectively.

Farell et al. (1994) illustrated how the distance function can be estimated using DEA based on Malmquist Productivity Index can decomposed into technical efficiency change and technological change. TEC is further decomposed into (PTEC) and (SEC). To measure Malmquist TFP Growth between any two periods as defined in equation (7), four distance functions have to calculated, which would involved four Linear Programms (LPs).They are as follows:

[[[d.sup.t.sub.o]([y.sub.it], [x.sub.it])].sup.-1] = max [phi], [lambda][phi]

subject to - [phi][y.sub.it] + [Y.sub.t][lambda][greater than or equal to] 0, - [x.sub.it] + [X.sub.t][lambda][greater than or equal to] 0, [lambda] [greater than or equal to] 0, (8)

[[[d.sup.t+1.sub.o]([y.sub.it+1], [x.sub.it+1])].sup.-1] = max [phi], [lambda][phi]

subject to - [phi][y.sub.it+1] + [Y.sub.t+1][lambda][greater than or equal to] 0, - [x.sub.it+1] + [X.sub.t+1][lambda][greater than or equal to] 0, [lambda][greater than or equal to] 0, (9)

[[[d.sup.t.sub.o]([y.sub.t+1], [x.sub.t+1])].sup.-1] = max [phi], [lambda][phi]

subject to - [phi][y.sub.it+1] + [Y.sub.t+1][lambda][greater than or equal to] 0, - [x.sub.it+1] + [X.sub.t][lambda][greater than or equal to] 0, [lambda][greater than or equal to] 0, (10)

[[[d.sup.t+1.sub.o]([y.sub.it], [x.sub.it])].sup.-1] = max [phi], [lambda][phi]

subject to - [phi][y.sub.it] + [Y.sub.t+1][lambda][greater than or equal to] 0, - [x.sub.it] + [X.sub.t+1][lambda][greater than or equal to] 0, [lambda][greater than or equal to] 0, (11)

where [y.sub.it] and [x.sub.it] represent an (Mx1) vector of output and an (Nx1) vector of input, respectively, of the i-th industry in the t-th period (t = 1, 2,.... T), and [Y.sub.t] and [X.sub.t] represent an (M x K)output matrix and (N x K) input matrix in period t, respectively for all K industries in the t-th period. Lastly, [lambda] = [[lambda].sub.1][[lambda].sub.2]....... [[lambda].sub.k] is a (K x 1) vector of constants representing weights and [phi] is a scalar.

Among the above four PLs, equations (8) and (9) are standard DEA LPs, which measure technical efficiency of the i-th firm in the t-th and t + 1 -th year, respectively. In equations (10) and (11) production points are compared to technologies from different time periods. Therefore, in these two LPs the value of [phi] parameter need not be greater than or equal to one, if technical regress or progress has occurred.

The above four LPs are required for each pair of adjacent years. Thus, if there are T time periods, then a total of (3T-2) LPs must be calculated for each industry. Since the sample used in this study consists of 11 years (from 1993-94 to 2003-2004), 31 LPs must be solved for each industry. The above LPs can be extended by decomposing technical efficiency change into pure technical efficiency change and scale efficiency change components. This requires the calculation of the distance functions with Variable Returns to Scale (VRS) technology, which could be done repeating LPs (8) and(9) and adding the convexity constraint KI' [lambda] = i to each of these LPs. The K1 is a K x 1 vector of ones and that the convexity constraint essentially ensures that an inefficient firm is only benchmarked against firms of a similar size (Colli, Rao and Battese, 1998). The values obtained with Constant Return to Scale (CRS) and VRS technology ear be used to calculate the scale efficiency residually. This would increase the total number of LPs to (4T-2), for our sample it became 42 for each industry.

Classification of Industries

Since we are examining the TFP growth in selected capital-intensive and labour-intensive industries of India during reform period, we have classified manufacturing industries as capital-intensive and labour-intensive by the following procedure. We have calculated capital-labour ratio (K/L) for all manufacturing industries and K]L ratio for whole manufacturing sector. Then industries are ranked on the basis of their K/L values. We were chosen five industries above the mean K/L of whole manufacturing sector and classified them as capital-intensive, and five from below mean value called labour-intensive industries. Selecting five industries from each segment, in addition to K/L ratio, we also considered the share of value added of selected industries to total manufacturing and their export intensity (export/sales), which shows competitiveness of these industries in international market. Since the competitiveness of an industry primarily depends on productivity in the long period.

Industries Selected for Analysis

On the basis of above criteria the following industries have been selected for empirical analysis: (A) Capital-intensive industries--(1) Chemicals, (2) Drugs and pharmaceuticals, (3) Dyes and Pigments, (4) Metal and Metal Products, and (5) Passenger Car and Multi Utility Vehicles. (B) Labour-intensive industries--(1) Readymade Garments, (2) Gems and Jewelry, (3) Leather Products, (4) Coffee and Tea, and (5) Cotton Textiles.

IV. EMPIRICAL FINDING

As mentioned earlier, this study applies Malmquist Productivity Indices to measure TFP change and its sources using Data Envelope Approach. We used the computer software DEAP (Coelli, 1996) to compute these indices. Table 1 demonstrates average estimates (geometric mean) of Malmquist indices of total factor productivity change (TFPC), decomposed into technical efficiency change (TEC) and technological change (TC) in capital-intensive industries. TEC is further decomposed into pure technical efficiency change (PTEC) and scale efficiency change (SEC). The industries are arranged in descending order of their Malmquist productivity indices (TFPC). The value of TFPC greater than one reveals productivity growth, value equal to one indicates no change and lower than unity indicates regress in productivity growth. To estimate percentage change in productivity, one in subtracted from the TFPC index and then value is multiplied by 100, [(TFPC-1) x 100]. The rule applies to the other indices presented in the table.

Our results reveal that four out of five industries have recorded productivity growth over the years. The highest productivity growth is recorded by passenger cars and multi utility vehicle industry (5.3 per cent) followed by drugs and pharmaceuticals (4.6 per cent), chemical (4.3 per cent), and dyes and pigments (1.2per cent). The metal and metal products (-0.3 per cent) industry recorded productivity regress. In the former, technological progress is not offset by a decline in technical efficiency whereas in the later; negative growth in technology is accompanied by no change in technical efficiency.

Technological progress is seems to be major driver of TFP improvement in all the four industries where TFP showed a positive growth over the period. However, somewhat contrasted picture is observed for metal and metal products industry where a regress in technological progress and no change in technical efficiency. Technological progress enhanced by 5.7 per cent, 4.4 per cent, 2.5 per cent and 1.4 per cent in passenger cars and multi utility vehicles, chemical, drugs and pharmaceuticals and dyes and pigments, respectively. Contrasted result is observed in case of technical efficiency for the above industries. It was -2.5 per cent, -0.9 percent and -1.3 per cent respectively. The reason for decline in technical efficiency of these industries was a regress in scale efficiency (SEC) without any change in pure technical efficiency. The reason for downward trend in TFP change for metal and metal products was due to decline in technological progress by -1.1 per accompanied by no change in technical efficiency and PTEC and SEC contribute equally to TEC.

The mean indices reported in the end of the last row of Table- 1 are the averages of industry specific indices. We may note that the TFP growth in the aggregate of all five capital-intensive industries group has been 1.7 per cent per annum over the period of investigation. Technology has improved by 1.8 per cent but technical efficiency has declined by 0.9 per cent per annum.

The results for labour-intensive industries are presented in the Table 2. Estimated result reveals that only one out of five industries has recorded productivity improvement over the year, which is readymade garments (0.5 per cent). The cotton textiles (-2.1) gems and jewelry (-1.7 per cent), leather products (-0.9) and coffee and tea (-0.2 per cent) industries have recorded productivity regress. In the former, technical efficiency is not offset by a decline in technological progress whereas in the later; negative growth in technology is accompanied by an improvement in technical efficiency. It implies that all industries are used inputs in hand more efficiently, but this not outweighed the decline in technological progress.

Technical efficiency (catching-up) has shown an improvement in all the labour-intensive industries over the period except in cotton textiles where it has shown a declining trend, but Technical progress (i.e., shifting of frontier) in all industries recorded a declining trend during the study period. Technical efficiency was improved by 1.1 per cent in readymade garments, 0.6 per cent in gems and jewelry, 2.7 per cent in leather products and 1.1 per cent in coffee and tea. This improvement in technical efficiency is attributed by both PTEC and SEC in two industries and this is due PTEC growth in readymade garments and due to SEC in gems and jewelry. But cotton textile industry recorded negative trend in both PTEC and SEC.

The mean indices reported in the end of the last row of Table- 2 are the averages of industry specific indices. We observed that aggregate of all five industries in labour-intensive segment obtained negative TFP growth (-0.9per cent). Technical efficiency is improved by 0.5 per cent but Technology has declined by 2.2 per cent per annum.

The year-wise estimates of TFP indices and its components are reported in table 3 which do not show a steady improvement in productivity for the selected capital-intensive industries as a whole. Productivity growth has been observed in 6 out of 10 years and the rest 4 years exhibits productivity revert. Among the years showing productivity growth, the TFP growth ranged from 2 to 9 per cent and productivity regress is ranged from 3 to 7 per cent. Productivity regress is seen largely in the early years of reform. However, it seems to be some signs of recovery in productivity growth during later period of the study.

In most cases, the growth in productivity seems to be caused by an improvement either in technological progress or technical efficiency. Technological improvement recorded in 6 years ranging from 8.3 to 9.5 per cent and technical efficiency improved in 4 years ranging from 0.4 per cent to 9.3 per cent. Growth in both technical efficiency and technology is observed in only two years. Improvement in PTEC is found in some early years of the study, where as SEC is improved in latter years.

The annual changes in TFP indices and its components are presented in table-4 also do not show a steady progress in productivity in the selected labour-intensive industries as a whole. TFP growth has been observed in 5 out of 10 years and the rest 5 years exhibits productivity regress. Among the years showing productivity growth, the TFP growth ranged from 0.9 to 8 per cent and productivity regress is ranged from 1 to 12 per cent. Productivity regress is seen largely in the early years of reform. However, it seems to be some signs of recovery in productivity improvement during latter period of the study.

In the majority cases, the growth in productivity seems to be caused by an improvement either in technological progress or technical efficiency. Technological progress has been recorded in 4 years ranging from 0.9 to 2.6 per cent and technical efficiency has been improved in 6 years ranging from 4 per cent to 7.2 per cent. Growth in both technical efficiency and technology is observed in only two years. Improvement in PTEC is found scattered in all the years of the study, where as SEC is improved in latter years.

CONCLUSION

By using the non-parametric technique of DEA-type Malmquist index in this paper, we measure the productivity changes for selected capital-intensive and labour-intensive industries in Indian manufacturing sector from 1993-94 to 2003-04. This model helped us to isolate the contributions of technological change, efficiency change and scale change to productivity change in the industries. The empirical results show that the capital-intensive industries have recorded a positive TFP growth by 1.7 per cent per annum and the technological progress (shifting of frontier) is the sole contributor to the TFP growth, but their counterpart, labour-intensive industries recorded a productivity regress at the rate of-0.9 per cent per annum during the study period, although technical efficiency has shown an improvement but it is offset by high rate of regress in technological change. It is observed that relatively capital-intensive industries are improving productivity as compare to their counterpart, viz. labour-intensive industries. Relatively TFP growth achieved by capital-intensive industries during the study period provides some indication that policy-induced factors, such as de-licensing, flow of foreign direct investment and import of advanced technology have made positive impact on the TFP growth, but labour-intensive industries were failed to utilize these benefits. Hence, labour-intensive industries should attract foreign direct investment and, import and adopt advanced technology to survive during era of globalization.

Further, large scale production in the industries may be encouraged to take the advantage of the economies of scale, which would lead to greater efficiency in the industries, and consequently force the production points closer to the frontier. Economies of scale in the industries coupled with advanced technology acquisition would further develop downstream activities in the related industries.

Reference

Ahluwalia, I. J. (1991), Productivity and Growth in Indian Manufacturing, Oxford University Press, Delhi.

Balakrishnan, P. and IC Pushpangadan (1994), 'Total Factor Productivity Growth in Manufacturing Industry: A Fresh Look', Economic and Political Weekly, July 30, 2028-2035.

Balakrishnan, P., K. Pushpangadan and M. Suresh Babu (2000), 'Trade Liberalisation and Productivity Growth in Manufacturing: Evidence from Firm Level Panel Data', Economic and Political Weekly, October 7, 3679-82.

Balk, B. M. (2001), 'Scale Efficiency and Productivity Change', Journal of Productivity Analysis, Vol. 15, pp. 159-183.

Bauer, P. W. (1990), "Decomposing TFP Growth in the Presence of Cost Inefficiency, Non-Constant Returns to Scale, and Technological Progress", Journal of Productivity Analysis, Vol. 1, pp. 287-300.

Coelli, T. (1996), "A Guide to DEAP Version 2.1: A Data Envelopment Analysis (Computer) Programme", CEPA Working Paper 96/08, CEPA, Department of Econometrics, University of New England, Armidale.

Fare, R., S. Grosskopf, M. Norris, and Z. Zhang (1994), "Productivity Growth, Technical Progress and Efficiency Changes in Industrialised Countries", American Economic Review, Vol. 84 (1), pp. 66-83.

Industrial Credit and Investment Corporation of India Limited (1994), Productivity is Indian Manufacturing: Private Corporate Sector 1972-73 to 1991-92, ICICI, Mumbai.

Kumar, S. and R. R. Russell (2002), "Technological Change, Technological Catch-up and Capital Deepening: Relative Contributions to Growth and Convergence", American Economic Review, Vol. 92 (3), pp. 527-548.

Mawson, P., K. I. Carlaw and N. Mclellan (2003), "Productivity Measurement: Alternative Approaches and Estimates", New Zealand Treasury, Working Paper 03/12, www.treasury.govt.nz.

Mahedevan, R. (1992), "Productivity Growth in Australian Manufacturing Sector: Some New Evidence", Applied Economic Letters, Vol. 92, pp. 920-936.

Mokhtarul Wadud, I. K. M. and Satya Paul: "productivity Growth, Efficiency Change and Technical Progress- A Case of Australian Private Sector Industries", The Indian Economic Journal, Vol. 54 (2), July-September, 2006, pp. 145-165.

Pradhan, G. and K. Barik (1998), 'Fluctuating Total Factor Productivity Growth in India: Evidence from Selected Polluting Industries', Economic and Political Weekly, February 28, M25-M30.

Singh, S. P. (2006), "Technical and Scale Efficiencies in Indian Sugar Mills: An Inter-state Comparison", The Asian Economic Review, Vol. 48(1), pp. 87-100.

Singh, S. P. and Shivi Agarwal (2006), "The Total Factor Productivity Growth, Technical Progress and Efficiency Change in Sugar Industry ofUttar Pradesh, The Indian Economic Journal, Vol. 54 (2), July-September, 2006, pp. 59-80.

D. H. MANJAPPA & M. MAHESHA

University of Mysore, Mysore

Note

(1.) Refer survey by Tybout (2000) for a detailed look at the contrasting theoretical arguments on the effects of trade liberalization on productivity.

Table 1
Estimates of Malmquist Indices of Total Productivity Growth and its
Components for Capital-intensive industries in India, 1994 to 2004

Industry                   TEC      TC    PTEC     SEC    TFPC

Passenger Cars & Multi    0.975   1.057   1.000   0.975   1.053
  Utility Vehicles
Drugs & Pharmaceuticals   1.021   1.025   1.001   1.020   1.046
Chemicals                 0.991   1.044   1.000   0.991   1.043
Dyes and Pigments         0.987   1.014   1.000   0.987   1.012
Metal & Metal Products    1.000   0.989   1.000   1.000   0.997
Mean                      0.991   1.018   1.000   0.991   1.017

Table 2
Estimates of Malmquist Indices of Total Productivity Growth and its
Components for Labour-intensive industries in India, 1994 to 2004.

Industry              TEC      TC    PTEC     SEC    TFPC

Readymade Garments   1.011   0.995   1.014   0.997   1.005
Cotton Textiles      0.977   0.979   0.985   0.992   0.979
Gems & Jewelry       1.006   0.964   0.998   1.009   0.983
Leather products     1.027   0.965   1.005   1.022   0.991
Coffee & Tea         1.011   0.991   1.003   1.008   0.998
Mean                 1.005   0.978   0.999   1.006   0.991

Table 3
Annual Changes in Total Factor Productivity, Technical Efficiency and
Technology in Capital-intensive Industries in India, 1994-95 to 2003-04

Years      TEC      TC    PTEC     SEC    TFPC

1994-95   0.987   0.976   1.000   0.985   0.961
1995-96   0.881   1.095   1.000   0.881   0.965
1996-97   0.976   0.976   1.000   0.976   0.952
1997-98   0.961   1.092    0.85   0.447   1.049
1998-99   1.093   0.992   1.176   2.672   1.034
1999-00   0.963   1.049   1.000   0.963   1.011
2000-01   1.031   0.898   1.000   1.031   0.927
2001-02   0.997   1.083   0.172   1.050   1.080
2002-03   1.004   1.083    0.66   1.195   1.097
2003-04   1.009   1.085   1.000   1.292   1.094
Mean      0.991    1.18   1.000   0.991   1.017

Note: the estimates of any two years indicate change over the
preceding to the following year. For example, the years 1994-1995
refer to the changes over the year 1994 to 1995, and so on.

Table 4
Annual Changes in Total Factor Productivity, Technical Efficiency
and Technology in Labour-intensive Industries in India, 1994 to 2004

Years      TEC      TC    PTEC     SEC    TFPC

1994-95   0.942   0.971   1.000   0.942   0.914
1995-96   0.979   1.009   0.871   0.906   0.989
1996-97   1.072   0.917   1.049   1.022   0.983
1997-98   0.912   0.985   0.892   0.922   0.898
1998-99   1.053   0.981   0.772   1.084   1.033
1999-00   1.057   1.012   0.999   1.059   1.070
2000-01   1.073   0.981   1.029   0.979   1.053
2001-02   0.865   1.021   0.867   0.698   0.883
2002-03   1.054   1.026   0.997   1.098   1.081
2003-04   1.040   0.971   1.000   1.040   1.009
Mean      1.005   0.987   0.947   1.005   0.991

Note: the estimates of any two years indicate change over the
preceding to the following year. For example, the years 1994-1995
refer to the changes over the year 1994 to 1995, and so on.