Chapter-VI TECHNOLOGY TRANSFER, INTERNATIONAL TRADE AND INDUSTRIAL DEVELOPMENT

Transcription

1 Chapter-VI TECHNOLOGY TRANSFER, INTERNATIONAL TRADE AND INDUSTRIAL DEVELOPMENT 6.1 INTRODUCTION Determining the factors that triggers the sustainable industrial growth is an issue of great debate amongst the scholars since at least the early years of industrial revolution, which emerged with greater intensity during the last half of the twentieth century (Ruttan, 2001). The whirlwind revelation in the seminal paper by Solow (1956) of an exogenous output determining factor other than the tangible capital and labour inputs which was regarded as the measure of our ignorance by Abramovitz (1956); broke the earlier presumptions of growth of labour force (Smith, 1776; Ricardo, 1817) and investment in capital equipments (Harrod, 1939; Domar, 1946, 1947) for sustaining growth. Several attempts were made since then to understand this Solow-residual which determines technological change (Verspagen, 1992). Prominent among them were the learning by doing model by Arrow (1962); Uzawa s (1965) model of investment in human capital and Shell s (1967) model of inventive activity. But these nascent endogenous technological change models died out soon (Verspagen, 1992) until their dynamic counterparts emerged during the mid 1980s. Pioneered by Romer (1986) and Lucas (1988), these New Growth Theory regarded investment in Research and Development (R&D), human capital accumulation and externalities as the factors determining the long run economic growth Endogenous Technological Change Since technological change is the intentional outcome in the endogenous growth models which requires the rate of investment to be internalized. Although it is ultimately the factor accumulation that accounts for growth, yet for factor accumulation to grow, the return to capital stock should not diminish. The new knowledge, which prevents diminishing returns to capital stock, is produced by investment in research technology which prohibits diminishing returns (Singh,

2 a). Technology, a form of knowledge, has some peculiar properties unlike the other tangible factors of production. First, technology is non-rival good, the use of which does not preclude others from using it even simultaneously. Secondly, it is partially non-excludable, in the sense that the creators or owners of technical information often have difficulty in preventing others from making an unauthorized use of it, at least in some applications; albeit on policy choice with respect to intellectual property rights (Grossman and Helpman, 1991). Basically these features may generate technological spillover which means that (1) firms can acquire information created by others without paying for that information in the market transaction, and (2) the creators or owners of the information have no effective recourse, under prevailing laws, if other firms utilize information so acquired (ibid, 1991). Moreover, Griliches (1979) distinguished spillovers in to two types. Besides the above mentioned knowledge spillover, he found that spillover may occur because downstream users do not pay the full values of the input, which he called as pure rent spillover. Further, trade is regarded as a prominent source of transmission of technology as it can lead to knowledge spillover which is embodied in goods which are imported internally or internationally or technology may be acquired through direct trade in disembodied knowledge - through contracts supported by policies that protect intellectual property (Hoekman and Jevorcik, 2006) Technology, Trade and Growth Theory and empirics of recent economic history suggests two important trends in the world economy. First, the technology is acclaimed as the source of economic growth. Second, the nations in the world economy are becoming increasingly open and interdependent (as discussed in Chapter 2). Thus, studying the relationship between the technology, trade and growth becomes imperative. On the one hand, the relationship between technology and growth is the centerpiece of the endogenous growth models where the industrial success depends on how each country learns and organizes itself to use constantly changing industrial 40 The different modes of technology transfer are discussed in the chapter

3 technologies on its technological competence (Lall, 2000). On the other hand, describing the relationship between the technology and trade Krugman (1990) regarded that there is a natural alliance between the new trade theory, with its emphasis on increasing returns and imperfect competition, and the view that technological change is a key factor driving international specialization. Thus, these developments in the theories of endogenous technological change, in particular the work by Romer (1986; 1990) and Lucas (1988) has stimulated a new analysis of the relationship between technology, trade and development in open economies (Grossman and Helpman, 1991; Rivera-Batiz and Romer, 1991). In the New Growth Theory, (Grossman and Helpman, 1991) it was analyzed how trade affects long run growth as technology got diffused by being embodied in inputs which in turn increase the importing country s productivity through the research and development (R&D) efforts of the exporting country (Keller, 2000). The empirical justification to these theoretical models was initiated by Coe and Helpman (1995) wherein they found a significant positive correlation between the TFP level and the R&D expenditure of the trade partners for the 22 OECD countries and Israel for the period But using the same dataset, Keller (1998) found that international spillovers are an important source of productivity growth in the host countries. Coe et al., 1997 extended their analysis to 77 developing countries where again trade emerged as the important vehicle of technology spillover to the developing countries from the industrially advanced countries of the North. These divergent results raised the debate regarding the role of trade as a transmitter of knowledge spillovers that remain unresolved till date. A huge literature is available on the issue with respect to different parts of the world, but the overall results remain inconclusive. Eaton and Kortum (1996), Helpman (1997), Xu and Wang (1999), to name few found a trade to be a transmitter of technology spillovers. But, Evenson and Singh (1997) using the sample of 11 Asian countries for the period found that the higher elasticity of indigenous knowledge stock as compared to foreign technology spillover in determining growth. 142

4 Similar results were accrued in case of Korea (Singh, 2004 a) wherein it was found that when the technology becomes tacit, trade does not play a role in technology diffusion. So, in the condition of the present inconclusive results, the present chapter is an attempt to find the justification for the new growth theory by empirically analyzing the scenario in case of the Indian organized manufacturing industries. Thus, the hypothesis of the present exercise is that imports leads to technology spillover from the advanced countries which would trigger the TFPG in the sector. This chapter is organized as follows. Besides the present section on introduction, in Section 6.2 the model is specified and the results are presented and discussed. Section 6.3 then presents the results of various studies on the subject while building the comparative analysis with the current study. The last section concludes the chapter. 6.2 PRODUCTIVITY AND ITS SOURCE Model Specification Measuring technology, an intangible factor of production is a difficult task. However, there are two widely used indicators for its measurement. One, being the input factor (investment in R&D) and other being the rate of productivity growth (Keller, 2004) 41. Thus, taking these variables the model is formed (Chapter 3) wherein the relationship is built between the TFPG and the two sources of R&D investments, that is indigenous and through spillovers. Debate on investment in R&D as a source of growth started with the theoretical breakthrough of growth models in which technology having certain peculiar features became endogenous in late 1980s; and more so with the empirical findings of Coe and Helpman (1995) wherein it was much acclaimed that unlike the other tangible factor, investment in R&D generates technology which generates externalities. These studies further claims that the benefits of R&D done by others can 41 However, he also cited patents as another output indicator of technology measurement. But he regarded that patent counts may not measure technology output well as decision to get patent is an act of choice on the part of the firm and thus there is a great variation in the propensity to patent. 143

5 be accrued easily by establishing close integration with technology generating partners. But these claims are refuted by Evanson and Singh (1997) and Singh (2004 a). Moreover, taking the recourse to these sources, investment in indigenous R&D not only generates new information but it also develops the ability to identify, assimilate and exploit knowledge from the environment in terms of enhancing the absorptive capacity (Cohen and Levinthal, 1989). Further, spillover can occur due to exports, imports and/or FDI (Lall, 2001) which are closely interlinked (Singh, 2004). But these two important sources of R&D indigenous R&D and spillovers can either be complementary (as shown by studies like Cohen and Liventhal, 1989) or substitutable (Coe and Helpman; 1995; Coe et al., 1997). Thus, the model takes the form in which it analyze the relative elasticity of indigenous R&D stock and import weighted R&D spillover on the TFP index (as calculated in Chapter 5, using the growth accounting method). The descriptive statistics of these three variables are presented in Appendix V.II followed by the results presented in Table 6.1. The descriptive statistics (Appendix V.II) for the whole data composed of 1326 observations for industries and 26 years show that negatively skewed and leptokurtic log of TFP Index which signifies that the dependent variable was higher in the earlier years of the analysis. Further, the standard deviation statistic for the log of indigenous R&D stock for the period is 3.03 which show higher variability around the mean. The variability was also found in case of the other independent variable (spillover) of the analysis for the period The overall descriptive statistics signifies that the log values of all the variables in the present analysis are normally distributed for most of the cases except for the few discrepancies (TFP index is negatively skewed and leptokurtic in pre-reform period for all industries and specifically for the HT and LT industrial sub-groups) industries with NIC 04 codes 182, 223, 233, 243, 273, 315, 319, 342 and 343 have been dropped out of 60 industries due to the non availability of the data. 144

6 Table 6.1 Impact of Indigenous R&D and Technology Spillovers on TFPG Dependent Variable: log Total Factor Productivity Growth Index Pre- Post Pre- Post Pre- Post Pre- Post Pre- Post All Organized Manufacturing HT MHT MLT LT GLS GLS GLS Random^ Fixed # TSCS Random Fixed # TSCS Random # Fixed TSCS TSCS Rando Fixed # m log R&D *** capital (0.63) (0.32) (1.72) (-0.82) (0.39) (-0.55) (-0.02) (0.86) (0.11) (-0.47) (3.68) (-0.01) (1.08) (1.09) (0.51) stock log Spillover * (-2.58) (0.25) * (-2.73) *** (-3.4) 0.43 (1.25) *** (-3.44) (-1.22) (0.67) (-1.72) (-2.43) 0.06 (0.61) (-1.86) (-0.96) (-2.32) (0.49) Constant 4.27 *** (70.13) 3.42 *** (36.05) 3.96 * (45.32) 4.42 *** (11.4) -1.2 (-2.3) 4.28 *** (13.2) 4.33 *** (20.57) 2.9 *** (20.8) 4.24 *** (38.2) 3.88 *** (8.33) (-1.21) 3.88 *** (16.02) 4.28 *** (51.79) 3.37 *** (11.24) 3.01 *** (146.8) B-P Test P-value Hausman Test P-value DW Test AR(1) Wald *** 12.2 ** chi2(2) F-test *** *** 5.21 *** Observ * * Notes: 1. Generalised Least Square Method (GLS) and Time-Series and Cross-Section (TSCS) is estimated after controlling for auto-correlation and heteroscedasticity in the panel dataset. 2. # and ^ means the estimated is done after controlling for auto-correlation and heteroscedasticity only, respectively. 3. Figures in brackets are the z-value, but for Fixed effect models it is the t-values. 4. ***, ** and * means the values are significant at 1 percent, 5 percent and 10 percent level. 5. B-P test means the Breusch-Pagan/Godfrey Test for heteroscedasticity and DW test means Durban Waston test for autocorrelation and AR(1) is first order autocorrelation coefficient. 145

7 Estimation of parameters using ordinary least squares (OLS) assumes that the error term to be homoescastic and serially independent. The violation of these assumptions could results in unbiased estimates. In the present exercise, Breusch- Pagan (B-P) test is used to deduct the problem of heteroescasticity. The significant p- values of B-P test rejects the null hypothesis of homoescasticity for most of the cases. Secondly, Durban-Waston (DW) test is used to test the problem of autocorrelation. The value of DW statistics is high than 2 only in four cases (pre-reform period model for HT and MHT industries and post-reform model for MLT and LT industries). Thus, after controlling for the discrepancies, where ever arises, the various models were estimated for the Indian manufacturing industries for the four technology intensive industrial sub-groups for the period , pre- and post-reform period (Table 6.1). The Wald Chi2 test with degree of freedom two and F-test were used to test the goodness of fit of the models in case of GLS/TSCS models and Random/Fixed models, respectively. Further, the coefficient of first order autocorrelation was lying between -1 and 1 for all the cases signifying that the problem of serial correlation is controlled during the estimation of parameters. It was found (Table 6.1) that the coefficient of indigenous R&D stock remained an important factor in determining TFPG index for the pre-, post- reform and also for the whole period under study. The magnitude of the import weighted technology spillover stock remained insignificant at (Table 6.1) for the period This shows that the indigenous R&D investment plays a dominant role in determining the productivity growth during the period under study. In order to understand the relationship between the technology gap and spillovers, an analysis is made by classifying the industries in to four technology intensive industries according to OECD (2007) classification. This exercise put light on the fact that whether the industries with high technical intensity (HT and MHT industries) attract greater spillover or it is the phenomenon witness in case for the low 146

8 technical intensive industries (MLT and LT). During the period to , the coefficients for indigenous R&D stock and technology spillover stock emerge negative for HT and MLT industrial sub-groups. The magnitude of indigenous R&D stock were although positive for MHT (0.002) and LT (0.06) during the period ( ), but the magnitude of technology spillover remained low in both these industrial sub-groups at and 0.05, respectively. This clearly show that the indigenous investment in R&D stock acts as a source for TFPG in the organized manufacturing industries, irrespective for the four technology intensive industries form high to low technological industries rather than the import weighted technology spillover Productivity Gap and Spillovers In the previous section, the analysis showed that the technology spillovers remain an insignificant factor in generating productivity irrespective of technology gap. In conjecture, the present section aims to analyze whether the technology spillovers varies with productivity gap. By productivity gap, we mean the difference in productivity level. But in order to find the productivity gap within the manufacturing industries, total factor productivity ratio is estimated (TFPG MAX / TFPG i ) as the ratio of the industry with the maximum trend growth of TFP over the years to the trend growth of TFPG for the individual industries i during the period. For the categorization, three groups were identified (1/3 of the sample each) 43 : group 1 (small gap), with the total factor productivity ratio between 1 and 0.46 (composed of 1/3 industries of the sample with high trend growth of TFP); group 2 with medium sized gap (TFPG MAX / TFPG i is between 0.46 and -0.15); and group 3 (large gap) where the industries have a very low trend growth rate of TFP during the period (total factor productivity ratio of less than -0.15). 43 Group 1 include industries with NIC codes 232, 231, 333, 341, 332, 311,331, 342, 261, 281, 202, 359, 271, 210, 361 and 252; group 2 include 269, 222, 300, 313, 351, 172, 292, 289,221,2423, 321, 353, 171, 369, 323, 352 and 251; rest composed group

9 Table 6.2 Productivity Gap and Spillovers ( ) Dependent Variable: log Total Factor Productivity Growth Index Small Gap Medium Gap Large Gap TSCS # TSCS Fixed^ log Indigenous R&D stock (2.01) (0.19) (-1.79) log Spillover (-1.96) Constant 2.36 * (4.9) * (-3.46) * (34.68) (-2.42) 4.31 * (67.97) B-P Test P value DW Test AR(1) F test 4.25 Wald Chi ** Notes: As Table 6.1. The regression estimation for the sample of small, medium and large productivity gap industries shows that the coefficient of spillovers remains insignificant for the aforesaid three samples. But the coefficient of indigenous R&D stock is positive and relatively high in magnitude for the small gap and followed by medium gap. But the magnitude is negative for the large productivity gap industries. These results signifies that the amount of indigenous R&D investment plays an important role in determining the total factor productivity growth in the Indian organized manufacturing industries. This type of phenomenon gave an important implication from the policy prescription point of view wherein the need is to divert the resources towards the development of technology so as to narrow the technology gap within the different industries; for which the investment in R&D is a paramount factor. 6.3 ABSORPTIVE CAPACITY After the empirics of the previous section wherein it was found that import weighted technological spillovers failed to became an important determinant of total factor productivity growth of the organized manufacturing industries irrespective of 148

10 the technology gap and/or productivity gap within the industries. These results put in skepticism the whole literature on the effects on technology spillovers which emerged with the theoretical revelation in the paper by Romer (1986) and got the empirical backing from the empirical evidences presented by Coe and Helpman (1995), Coe et al. (1997). The neoclassical economists regard the public good character to technology which leads to its free and effortless spillover with various embodied and disembodied means. But the recent empirical literature based on the evolutionary theories pioneered by Nelson and Winter (1982) emphasised that spillovers are not automatic and required specific strategies for its absorption. Thus, the assimilationist theories regards absorption of spillovers as an important process in which the cost and effort is required for acquiring the technology. Thus, technological learning is a real and significant process, which is conscious and purposive rather than automatic and passive (Lall, 2001). So, in the world of imperfect knowledge, capabilities are required to build as technological developments are needed at different depths. The attainment of minimum level of capability (know-how) is essential at all activities and for attaining deeper capabilities, an understanding of the principles of technology (know-why) is essential. In this regard the evolutionary theories stress the centrality of developing technological capabilities. But building technological capability is a much complex task that requires broader range of effort to access, implement, absorb and build upon the technologies required in production (ibid). Research and Development (R&D) is often regarded as an important indicator of such absorptive capacity. Cohen and Levinthal (1989) while underlining the strength of investment in R&D regarded that there are two faces of R&D- one, being the generation of technology and other is developing the absorptive capacity so as to realize technological spillovers. Thus, to explore this aspect of absorptive capability, the manufacturing industries were classified into three groups: the first being the one that invest more in R&D 44 ; second with the medium R&D investment and the third with the low investment in R&D with one-third industries in each group. 44 This classification should not be confused with the one done for the technology gaps as the latter was based on the OECD (2007) classification while the former was done according to the quantum of R&D stock investment in the Indian organized manufacturing industries. 149

11 This division (Table 6.3) shows that from the medium to low R&D stock industries, as the level of investment in indigenous R&D falls in the latter group, the magnitude of spillovers also falls as they became negative for the low R&D stock industries. But the case for the high R&D stock industries requires more scrutiny to understand the underlying trend. The group (high R&D stock industries) includes industries like pharmaceuticals (2423), office equipments (300), basic chemicals (241), and special purpose machinery (292). These industries are relatively high technology intensive in nature (Appendix I). Thus, the quanta of R&D investment in these industries were not sufficient to attract more spillovers in these industries. These results were similar to the one found for Korea by Singh (2004 a) wherein it was found that as the technology became complex, trade no longer acts as a medium of technology spillovers. Table 6.3 Absorptive Capacity and Spillovers ( ) Dependent Variable: log total factor productivity growth index R&D stock Capital-Labour Ratio HIGH MEDIUM LOW HIGH MEDIUM LOW Random^ TSCS # TSCS TSCS TSCS # TSCS log Indigenous R&D stock (-1.71) 0.42 * (4.2) 0.03 (0.46) 0.02 (0.58) (-0.57) 0.09 (2.26) log Spillovers * (-6.99) (0.24) (-0.5) (-2.47) * (-3.79) (-1.04) Constant 5.76 * 1.62 *** 3.97 * 3.79 * 4.05 * 3.93 * (4.03) (2.72) (0.000) (0.000) (0.000) (32.12) B-P P-value Test DW Test AR(1) Wald Chi * * * 5.84 Notes: As Table

12 Apart from investment in indigenous R&D, alternative factors like investment in physical capital, human capital along with the institutional structure, industryuniversity linkages etc. all individually and more often in tandem with each other have an impact in building the absorptive capacity. But the analysis of all these factors would be beyond the scope of the chapter and thus, capital intensity as a proxy for the absorptive capacity is analyzed as a factor determining spillovers. Again, the classification of industries was done on the extent of capital intensity into high, medium and low capital intensity (Table 6.3) with one-third of industries in each group. Here again, the extent of spillovers were too low (negative for the three groups). This signifies that the attracting the spillovers is not an easy task. It is the investment in indigenous R&D which is a paramount factor determining the productivity. But if we consider the R&D expenditure as a percentage of sales turnover, it was 0.62 percent on an average during the pre-reform period which fell to 0.46 percent during the post-reform period. It is pertinent to mention here that R&D expenditure as a percentage of sales turnover for most of the advanced countries ranges between 2 to 4 percent (DST, New Delhi, 2003). Further, the low elasticity for the import weighted spillovers are quiet contrary to the claims pioneered by the neoclassical economics that were backing the new growth and trade theories since late 1980s. The present results refute the recent theoretical models which highlight the importance of trade as a vehicle for technological spillovers that allow less developed countries to reap the benefits of R&D invested in the industrial countries. However, from to the imports from OECD countries to India has increased by 3.1 percent at the rate of 0.3 percent per annum and from to , by 6.6 percent at the rate of 0.5 percent per annum 45, respectively. Thus, despite seeing a surge in the imports from OECD countries, the rate of technology spillovers remains insignificant; which could also be the result of the strict intellectual property rights regime as imposed by the World Trade Organization and the uneven rules of the game (Stiglitz, 2006). 45 Calculated using WITS imports data to India from OECD countries. 151

13 6.4 DIFFUSION OF TECHNOLOGY THROUGH TRADE: COMPARISON WITH SIMILAR STUDIES The literature has found several modes of technology transfer but Pack (1994) and Singh (2004 a) regarded that trade is the most widely accepted channel of indigenous and international technology spillovers across industries. Thus, Table 6.4 presents various studies from the literature which shows the impact of trade on the extent of technology spillovers. Table 6.4 Diffusion of Technology through Trade Author Sample Period Results Coe and Helpman (1995) Eaton and Kortum (1996) 20 OECD countries and Isreal 19 OECD countries Coe et al. (1997) 77 Developing countries Evenson and Singh (1997) Lichtenberg and Potterie (1998) 11 Asian Countries 20 OECD countries and Isreal Keller (1998) 20 OECD countries and Isreal Edwards (1998) 93 advanced and developing countries Xu and Wang (1999) 21 OECD countries Keller (2000) 8 OECD countries Singh (2004 a) South Korean 28 Manufacturing industries Present Study Indian Manufacturing Industries 1971 to 1990 Substantial benefits accrue from the R&D done by trade partners. Mid 1980 s to mid 1990 s More than 50 percent growth in the sampled countries derived from the innovations in the US, Germany and Japan to 1990 Substantial technological flows occur from North to South Higher elasticity of indigenous knowledge stock than foreign on productivity to 1990 More open to trade a country is, the more likely it is to benefit from foreign R&D to 1990 Claimed that trade are important source of R&D spillovers Total factor productivity is faster in more open economics 1983 to 1990 About half of the returns on R&D investment in G7 countries spilled over to the OECD countries 1970 to 1991 Technological spillovers will be higher if the import share from high knowledge economies is higher to 2000 When the technological complexity increases, trade does not lead to technology spillovers to 2006 Trade does not automatically lead to technology spillovers. 152

14 In the Table 6.4, some selective studies were presented in which the relationship between technology, trade and growth was undertaken. The pioneer study on this relationship was done by Coe and Helpman (1995) wherein empirical justification to the role of trade in transmitting the technology to the host countries were presented while undertaking the case study of the 22 OECD countries and Israel. A year later, Eaton and Kortum (1996) found the similar results for the 19 OECD countries. Again Coe at al. (1997) took 77 developing countries as the sample to reveal the validity of trade as the transmitter of knowledge and again found substantial technology spillovers from North to South. But in the same year Evenson and Singh (1997) does not find the evidence for the technology spillover for the 11 Asian countries. Also Young (1991) found that the impact of trade varies with the developmental aspects of the economy. Whereas, it raises the rate of GDP growth in the developed countries, however, it has an adverse impact on the LDC. Moreover, from the literature it was found that two important studies have done a similar analysis as the present one. One was by Keller (2000) where he tested two hypotheses: first being that importing largely from the high-knowledge countries could lead to higher productivity as compared to importing from the low-knowledge economies. Second, the effect of spillovers will be strong if the country s overall import share is higher. Based on the eight OECD countries and using the import of intermediate capital goods as a transmitter of technology spillovers, the study suggests that imports help in explaining productivity growth. The other study was done by Singh (2004) on Korea wherein using the panel data for twenty-eight manufacturing industries over the period , the study found that that trade is an important channel for domestic and international spillovers, but when the knowledge become more complex and tacit, trade no longer serves as a means of technology spillovers. The present analysis is a modest attempt to find that whether imports acts as a means of technology spillovers by taking in to preview the Indian organized manufacturing industries. The attempt was made to find empirically the extent of technology spillovers from the advanced OECD countries. So confirming the results 153

15 as has been found by Evenson and Singh (1997) and Singh (2004 a), the present analysis also found that investment in indigenous R&D is an important factor in determining productivity and despite having an increased trade relations with the OECD countries 46 (that are on the frontiers of technology generation) the extent of technology spillovers remains very low. 6.5 CONCLUSION The recent development of endogenous theories of technological change has stimulated a plethora of debate regarding the relationship between technology, trade and growth in open economies. The present chapter is an attempt in this direction where the source of growth of the manufacturing industries in terms of TFPG is analyzed with respect to either R&D done by indigenous industries or the trade acts as a transmitter of technology spillovers from the major R&D investment countries; which in present case are the major OECD countries. The panel regression analyzes for the different technology intensive industries are done. The pre- and post reform periods are compared to specifically address the neo-liberal claims of increasing integration boost growth hypothesis. The overall results shows that investment in indigenous R&D remained an important factor in determining productivity growth for the pre-, post- reform and also for the whole period under study while the magnitude of technology spillover stock remained insignificant. Further, the analysis was also done by classifying the industries in to four technology intensive industries (OECD, 2007) for understanding the relationship between the technology gap and technology spillovers. The result shows that during the period to , technology spillover remains low irrespective of the technology gaps. The productivity gap and the extent of spillovers were also estimated which also showed insignificant amount of spillovers irrespective of productivity gaps. 46 In imports from OECD countries were 35 percent out of the total imports, which rose to 50.1 percent, 54.6 percent and 51.2 percent in , and , respectively (calculated using WITS imports data). 154

16 Thereafter, it was felt to find out the major requisites which could lead the technology to spillover. This exercise was aimed to find out the causes where the Indian manufacturing is lacking despite having trade relations with the major OECD countries. Based on the recent development in evolutionary theories that emphasized on developing the absorptive capacity, the analysis was done to analyze the impact of indigenous investment in R&D and the quantum of capital intensity in generating spillovers. With regard to the former factor, it was found that the investment in R&D is an important factor in generating spillovers but for the relatively high technology industries, the quanta of R&D investment were not sufficient to attract spillovers in these industries. Thus, the results points to the fact that to have high productivity growth is a continuous and path dependent phenomenon in which the investment in indigenous R&D is an important factor. The investment in R&D not only generates new technology but it also enhances the absorptive capacity of the country to reap the benefits from greater trade relations with major R&D investment countries who are in the frontier of technology generation. But to regard, the neoclassical paradigm of technology of possessing a public good character and its spillover being an automatic process is refuted by the present analysis. Thus, the analysis lead to two major policy implications. One, there should be strategic investment in indigenous R&D. Second, the quantum of R&D as a percent of sales turnover and also as a percent of GDP should be increased. 155