THE EFFECTS OF TRAINING, EDUCATION AND REGIONAL CHARACTERISTICS ON TECHNOLOGY DIFFUSION:

THE EFFECTS OF TRAINING, EDUCATION AND

REGIONAL CHARACTERISTICS ON TECHNOLOGY

DIFFUSION:

MSc. Thesis

Author: Daniel Nyári

University of Groningen

Faculty of Economics and Business

Student number: S2997843

Email: d.nyari@stud.rug.nl

A

BSTRACT

This thesis analyzes the effects of the firm’s main market, training programs and regional characteristics on technology diffusion. I use data from the EBRD’s Business Environment and Enterprise Survey, the European Values Survey and EUROSTAT. The results show that firms with in-house training programs are more, and that firms operating only on local markets are less likely to adopt technologies. Regarding the regional characteristics, the analysis showed that a larger share of the population with a desire to learn has a beneficial effect on technology diffusion. Regional IT penetration and tertiary education, however, are not found to be significant factors.

K

EYWORDS

T

ABLE OF CONTENTS ABSTRACT ... 2 1. INTRODUCTION ... 4 2. LITERATURE REVIEW AND HYPOTHESES ... 6 2.1 LITERATURE REVIEW ... 6 2.2 THEORETICAL FRAMEWORK AND HYPOTHESES ... 10 3. METHODOLOGY ... 15 3.1 DATASET ... 15 3.2 VARIABLES ... 16 3.3 MODEL SPECIFICATION ... 18 4. ANALYSIS AND RESULTS ... 18 4.1 SUMMARY STATISTICS ... 18 4.2 RESULTS ... 21 4.3 ROBUSTNESS CHECKS ... 23 4.4 DISCUSSION ... 23

4.5 LIMITATIONS AND FURTHER RESEARCH ... 26

5. CONCLUSION ... 27

APPENDIX ... 28

1.

I

NTRODUCTION

The diffusion of technologies is a well-established factor of productivity growth (Comin & Hobijn, 2010) and therefore, it also serves as a driver of catching-up among countries (Madsen, 2010). The possible benefits of technology diffusion are more pronounced in developing countries, as in these economies, 82% of the potential productivity growth could come from catching-up to the technological frontier (McKinsey Global Institute, 2015). As the convergence of developing countries ultimately depends on the firms operating there, it is essential to understand which technological, organizational and external factors influence individual firms in their decision to adopt technologies relevant to their production processes. The issue of catching-up is especially interesting in the case of the European Union for two reasons. First, Eastern European countries that joined the EU in 2004 and 2007 have had various levels of success in catching up to the old member states. Naturally, understanding technology diffusion would provide further support to the progress of these countries. Second, as these countries – and Europe as a whole – are regionally fragmented, the issue of regional cohesion is on the top of the EU’s agenda. According to Aiginger & Leitner (2002), with the rise of services, economies of scale is becoming less important, which causes regional concentration to decline. If it is indeed the case, the time for achieving regional cohesion is as right as it ever will be, and regional policies aimed at promoting technology diffusion might just be the perfect tool to achieve this goal. However, to design regional policies, a more thorough understanding of the effect of regional characteristics on technology diffusion is necessary.

regional level – remains a question as well, despite the fact that their importance is well established in the National (Nelson, 1993) and Regional Innovation Systems (Cooke et al., 1997) and also in the Smart Specialization concept (Foray et al. 2009).

Therefore, the aim of this thesis is to provide new insights on the effect of the environmental-regional level as well as not yet explored firm-level factors that influence the technology adoption of firms. The analyzed factors include: (1) tertiary education and training, the on regional and firm levels, respectively, (2) personal characteristics and values on the regional level, (3) regional IT penetration (4) participation of the firm in (inter)national markets. As these factors directly influence the technology adoption patterns of firms, information on them could prove to be useful when designing policies that aim to boost productivity as well as achieve regional cohesion. Regarding the effects of regional characteristics, this thesis draws upon insights from the Smart Specialization and RIS concepts.

I use data on firms located in Eastern European member states of the European Union provided by the European Bank for Reconstruction and Development (EBRD). The EBRD’s Business Environment and Enterprise Performance Survey (BEEPS) provides data on most of the firm-level characteristics that are proven to be influencing technology diffusion, as well as location data, which makes it possible to link the BEEPS dataset with the EUROSTAT database to test regional effects and with the European Value Survey compiled by the GESIS to test the effect of human values on technology diffusion. The case of Eastern European member states is particularly interesting, as this group of countries is heterogeneous in terms of regulatory environment, competitiveness, regional development and institutional obstacles - such as corruption -, thus providing a sufficiently diverse sample.

literature and introduce the theoretical framework and hypotheses, Section 3 gives information on the datasets and the variables, Section 4 presents the summary statistics, the results of the regression and robustness checks, provides a discussion of the results as well as the limitations of this thesis. Section 5 concludes.

2.

L

ITERATURE

R

EVIEW AND

H

YPOTHESES 2.1LITERATURE REVIEW

Two aspects of technology diffusion can be differentiated: inter-firm diffusion is the ratio of the firms using a new technology that is relevant to their production processes, while intra-firm diffusion is “measured by the proportion of the firm’s capital stock that incorporates the new technology” (Battisti & Stoneman, 2003, p. 1643.) Technology diffusion is therefore, a term for the spread of a given technology. Technology adoption, however, is a subset of the diffusion process, because it refers to the decision of individual firms to obtain a certain technology. It is defined as the acquisition of a technology that is relevant to the firms’ production process and other activities. This thesis will analyze the technology adoption of individual firms, but it will often use the term technology diffusion, when referring to the general process or a group of firms.

Despite the fact that the mainstream theoretical model of technology diffusion is the epidemic model (David, 2015), I will instead use the threshold model, which is considered to be the main alternative (Geroski, 2000). The threshold model of technology diffusion is appropriate for firm-level analyses such as this one for several reasons. Threshold models account for firm heterogeneity by allowing actors to differ in some key characteristics that affect the profitability of the adoption of a given technology. Firms will only adopt a technology if said characteristics exceed a certain threshold (David, 1969), for example: firms will only adopt ICT if a certain percent of employees knows how to use computers. These characteristics are as diverse as the innovations available. Additionally, in the case of cross-sectional data, epidemic models are not usable due to the fact that in such models, the passage of time is a proxy for the dissemination of information, which is a crucial part of the theory (Griliches, 1969).

1963; Schnepp et al., 1990; Roberts et al. 2003; Zhu et al. 2006; Vivarelli, 2013), firm culture and values (Lee and Palmer, 1999; Miron et al, 2004; Battista and Stoneman, 2005), and the level of competition (Porter, 1990; Viswasrao & Bosshardt, 1998).

The reason behind the positive relationship between firm size and technology adoption is that large firms face less financial constraints (Davies, 1979; Geroski, 2000), which makes adoption quicker because large firms are able to: pay higher wages to attract skilled labor, incur investments with high fixed costs and are generally less risk averse, thus are more likely to invest in complex and/or high-risk high-reward technologies (Haller & Siedschlag, 2011). As mentioned above, management practices also play a key role in the ability of firms to successfully adopt new technologies. The training and abilities of the middle- and top management are important factors that determine how quickly a firm can introduce an innovation (Mansfield, 1963). The lack of effective management systems as a blocking factor of technology adoption is confirmed by empirical studies as well (Roberts et al. 2003; Zhu et al. 2006; Vivarelli, 2013).

license foreign technologies and that the licensing creates additional spillovers as the number of technology purchases by other firms in the same industry tends to increase as a result of the initial adoption.

Innovation and technology diffusion are closely related. There are two underlying reasons for this. First, innovation activities might behave as magnets for additional knowledge within the firm, meaning that by developing new products and processes, firms might realize the potential of their innovation’s complementarity with know-how, process or any type of innovation outside the firm, thus giving incentives to adopt such technologies. These incentives might be remarkably strong, as only the largest firms are able to compile all the necessary knowledge to realize complex innovations. Second, innovative activity might as well be a proxy for the open-mindedness of the management, meaning that firms that develop new products and processes value novelties and are more likely to adopt new technologies to further increase their productivity. While this relationship between technology adoption and innovative activities, such as the introduction of new products and processes is straightforward, the case of R&D activity and technology adoption is ambiguous. On the one hand, Battista and Stoneman (2005) report that firms investing in R&D are more likely to adopt technologies and Raut (1988) also finds that R&D and foreign technology licensing are complementary. On the other hand, Katrak (1997) finds that R&D activities performed by the firm do not influence technology purchases, but as a result of licensing, firms are likely to have higher R&D expenditures later. The findings of Katrak (1997) thus present a different type of causality between R&D and technology adoption, which in the case of analyses that use cross-sectional data - such as this one- means that R&D as a control variable would be statistically insignificant.

Information flows due to increased competition on international markets also speed up technology diffusion. Haller & Siedschlag (2011), for example, report that firms exposed to international competition in export markets are more inclined to innovate and adopt new technologies. While competition has beneficial effects, too much competition lowers the returns to investments thus hindering technology adoption (Viswasrao & Bosshardt, 1998).

(2005), US multinationals respond to improvements of intellectual property rights protection. They find that royalty payments tend to increase between subsidiaries and the parent companies as a result of said improvements.

Contrary to firm- and country-level characteristics, the technology diffusion literature is scarce on regional and other close environmental effects, except for the positive impacts of industry concentration and competition (Gatington & Robertson, 1989; Haller & Siedschlag 2011; Zhu, Kramer & Xu, 2006). Therefore, to construct a comprehensive theoretical framework including regional effects, insights from related fields are useful. The already mentioned Regional Innovation Systems literature (RIS) is appropriate to draw ideas from regarding the role of regional characteristics for several reasons. First, it mainly focuses on innovation networks and innovation programs (Cooke et al. 1997), that contribute to regional development mainly through positive externalities and agglomeration effects, such as thick labor markets, which are important according to the technology diffusion literature as well. Second, the relationship also holds on the firm-level as the connection between innovation and technology adoption is well established in the literature (Battisti & Stoneman, 2005; Rogers, 1995; Cohen and Levinthal 1989). Besides RIS there is another, even more relevant theory, the Smart Specialization concept.

While there are some RIS papers like Lawton-Smith & Clark (2003) and De Brujin & Lagendijk (2005) that argue that goals of the EU2020 and Lisbon strategy – based on RIS concepts – namely, regional cohesion and increased competitiveness are contradictory, because the former aims to promote the development of the weaker regions, while the latter is about the promotion of high-tech and knowledge-intensive sectors (located in the most developed regions). In a framework based on technology diffusion and the Smart Specialization concept (Foray et at., 2009) they are actually self-reinforcing instead of being contradictory.

(2) adoption of said technologies, (3) co-invention of applications for the GPTs. Therefore, based on the literature, the role of technology diffusion is twofold: the adoption of technologies is a source of productivity growth in itself and it also serves as a basis for future innovations. As it is apparent, once the main goals of the EU based on RIS are interpreted in a technology diffusion framework, there is no contradiction or burden on policymakers to either pick winners or support regional cohesion, as both can be achieved at the same time. For this reason, to formulate hypotheses regarding regional and other external effects, I will use additional support from the RIS and Smart Specialization literatures.

2.2THEORETICAL FRAMEWORK AND HYPOTHESES

The technology-organization-environment (TOE) framework used in Zhu et al (2006) - originally devised by Tornatzky & Fleischer (1990) – is suitable for the purposes of this thesis because it provides a clear structure and is able to separate the regional effects from firm-specific effects.

In the TOE framework, there are three separate contexts that influence technology diffusion. The technological context captures the technical knowledge that is relevant to adoption process. The organizational context refers to the internal structure of the firm, while the environmental context contains a wide scope of external factors that have an effect on technology diffusion. Using the TOE framework, I develop the conceptual model as shown in Figure 1.

TECHNOLOGICAL CONTEXT

In the last few decades, information and communication technologies (ICT) have become significant drivers of productivity, even causing the divergence in competitiveness between the EU and the US (Timmer & Van Ark, 2005). The reason why ICT is important is that most industry 4.0 process and even product innovations are internet based. ICT facilitates the flow of information, making coordination of value chains easier (Porter, 2001), thus providing a basis for process innovations like ‘just-in-time’ logistics. According to Chatterjee et al. (2002) ICT provides significant organizational improvements, such as management support and cross-department coordination by increasing information flows. For this reason, Regional IT penetration is a significant factor of technology diffusion for several reasons. First, IT penetration is a proxy for the underlying ICT infrastructure of the region. Logically, if firms are to profit from the many advantages brought about by ICT, a supporting regional infrastructure is essential. Second, regional IT penetration is a proxy for the supply of labor capable of using ICTs. Employees with capabilities of using such technologies are needed because if the supply of such labor falls short, region-wide technology diffusion could be blocked. In fact, McCann and Acs (2011) find that ICT adaptation exacerbated the differences between core and none-core regions, which seems to confirm this blocking effect. Therefore,

Hypothesis I: In regions with higher IT penetration, the rate of firm-level technology adoption

will be higher.

ORGANIZATIONAL CONTEXT

aspects of a new technology, with human capital being one type of software to new technologies. The effect of skilled labor is also confirmed by many empirical studies: Bugamelli & Pagano (2004), Zhu et al. (2006), Lin & Lin (2007), Haller & Siedschlag (2011) report that the rate of diffusion of ICT and e-business is influenced by the availability of skilled labor. The importance of skilled labor is tremendous for several reasons. The first is straightforward: educated labor has a more diverse skillset, and is able to operate more complex systems, providing higher value added. Second, educated labor has better capabilities to grasp the concept of new technologies, thus ensuring successful adaptations and returns on physical capital investments. Based on the above, firms that invest in training their employees should adopt new technologies faster for three reasons. First, training programs are able to provide the technical knowledge or “software” necessary to adopt a certain innovation. Second, training programs improve the skills of the employees. Third, investing in training programs also signals the values of the firm, in this case, the fact that they value knowledge and novelties, which are the same qualities that support technology diffusion according to the literature.

Hypothesis II: A firm that invests in training its employees is more likely to adopt technologies.

ENVIRONMENTAL CONTEXT

The importance of skilled labor has been outlined in the organizational context. However, the issue of skilled labor supply is interesting in the environmental context as well, due to the fact that labor supply and demand heavily depends on regional characteristics. The thesis presents a hypothesis on the question of regional labor supply and firm-level technology diffusion. It is related to the supply of skilled labor through university education, which might be crucial for technology diffusion for two reasons:

regional development. In this case, if the supply of high skilled labor is not sufficient in the region, it might block the adoption of complex technologies, as there would be no personnel to operate such technologies. Based on the prominent role of skilled labor in technological change and regional development, I propose the following hypotheses:

Hypothesis III: In regions where the share of the population with tertiary education is higher,

firm-level technology adoption will be higher.

draw on codified, publicly shared knowledge” (Foray et al. 2009, p 2.). Naturally, as the GPTs are key in this concept, licensing, as a method of acquiring such technologies is closely connected to the matter. Therefore:

Hypothesis IV.: In regions where the ‘learning culture’ is stronger, firm-level technology

adoption will be higher.

market. Naturally, this means that locally operating firms are at the bottom of the productivity and therefore of the technology adoption ladder as well.

Hypothesis V.: Firms that only operate locally are less likely to adopt technologies.

3.

M

ETHODOLOGY

3.1DATASET

The data used in this thesis is obtained by combining three datasets. The first is the 5th round of Business Environment and Enterprise Performance Survey (BEEPS) which is a firm level survey conducted by the European Bank for Reconstruction and Development (EBRD) in cooperation with the World Bank between 2012 and 2013. The BEEPS provides an extensive cross-sectional dataset on firm characteristics and performance as well as the firms’ perception of the environment they operate in, such as the level of competition, regulatory obstacles or problems originating from political instability. The dataset is a representative sample of the firms in the economies of Central and Eastern European as well as Central Asian countries. The second dataset is obtained from EUROSTAT and contains regional data on the NUTS2 level that are used to test the hypotheses, such as regional IT penetration and tertiary education. The EUROSTAT database is a compilation of data collected by the national statistical authorities of the member states using the same methodology to provide data consistency. The third dataset is obtained form the European Value Survey (EVS), 2008 conducted by GESIS. The European Value Survey is the most comprehensive research on human values in European countries. It is a large scale project with the goal of creating a representative dataset on what Europeans think about work, politics religion and other issues. It also contains regional data, making it possible to link it with the BEEPS. I calculated the averages for each region from the responses of the total of 8,502 individuals.

countries as well as incomplete questionnaires from the BEEPS database 1,254 observations were left.

3.2VARIABLES

Dependent variable:

Technology adoption is defined as the acquisition of a technology that is relevant to the firms’ production process and other activities. To measure the individual firm’s decision to adopt technology, I will use the following indicator:

- foreigntech: is an indicator variable that takes the value of 1 if the firm is using technology

licensed from a foreign-owned company (excluding office software) and takes the value of 0 otherwise.

Independent variables:

To test Hypothesis I and to measure the interaction between technology diffusion and the IT penetration in the region, I will use the variable broadgrowth, which is growth in the ratio of households that have broadband internet connection in the firm’s region over the last 2 years. To test Hypothesis II I will use the indicator variable training, that takes the value of 1 if the firm had formal training for its full time employees in the last year, and 0 otherwise.

To test Hypothesis III and to measure the effect of higher education, I will use the variable

tertiarygrowth, which is the growth in the share of persons in the region with tertiary education.

To test Hypothesis IV and measure the benefits of a learning culture in the region, I will use

jobnewskills, the share of the population that thinks learning new skills is an important aspect

of a job, calculated from the European Value Survey.

To test Hypothesis V, I will use the indicator variable onlylocal, that takes the value of 1 if the firm’s main product was sold mostly in the same municipality where it is located.

Control Variables

The most important control variable is size, since it reflects the financial constraints as well as the firm’s efficiency. Size will be measured as a natural logarithm of the number of full-time employees. Foreign ownership is measured by an indicator variable that takes the value of 1 if more than 50% of the firm is owned by foreign entities and 0 otherwise.

Table 1: Variables Indepenent Variables

broad: the share of households with broadband internet connection in the firm's region

training dummy variable =1 if the firm invested in training programs for its full-time _{employees in the last fiscal year} tertiarygrowth the overall growth in the share of the region's population who has tertiary

education over the period between 2003 and 2008. (Source: EUROSTAT)

jobnewskills: the share of people in the region who think that acquiring new skills _{is an important aspect of a job. (Source: EVS)}

onlylocal: dummy variable =1 if the firm's main market is the municipality it is located in.

Control Variables

lnsize: the natural logarithm of the number of full-time employees in the firm

newproduct: dummy variable = 1 if the firm has introduced a significantly improved product in _{the last fiscal year} exporter: dummy variable = 1 if the firm's main product's market is foreign

foreign: dummy variable = 1 if more than 50% of the firm is owned by _{foreign entities}

orgopen: dummy variable = 1 if the employees are given time to try out and develop new approaches methods and ideas about products or services, business process, firm management or marketing

compusage: share of employees that use computers for their work

expcent: the experience of the firm's top manager minus the average top manager

experience in the sample in years

competition: a score between 1 and 5 calculated from the number of competitors

novelties. ICT usage is measured by the number of employees using computers. Finally, due to the ambiguous relationship between R&D and technology adoption, I will instead use the introduction of a new product as a measurement of innovation. A more detailed description of the control and independent variables is presented in Table 1.

3.3MODEL SPECIFICATION

I use a probit model to measure technology diffusion as indicated by the licensing of foreign technology by individual firms.

𝑓𝑜𝑟𝑒𝑖𝑔𝑛𝑡𝑒𝑐ℎ₊, = 𝛽_/+ 𝛽₁Σ₁₃₄5 _{𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 + 𝛽}

<𝑡𝑟𝑎𝑖𝑛𝑖𝑛𝑔 + 𝛽=𝑜𝑛𝑙𝑦𝑙𝑜𝑐𝑎𝑙+

+ 𝛾₄𝑏𝑟𝑜𝑎𝑑𝑔𝑟𝑜𝑤𝑡ℎ + 𝛾_B𝑡𝑒𝑟𝑡𝑖𝑎𝑟𝑦𝑔𝑟𝑜𝑤𝑡ℎ_C+ 𝛾_D𝑗𝑜𝑏𝑛𝑒𝑤𝑠𝑘𝑖𝑙𝑙_C + +𝜇₁Σ₁₃₄DD _{𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿}

1Σ1344/ 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 + 𝜀+

Where 𝛽_<𝑡𝑟𝑎𝑖𝑛𝑖𝑛𝑔 is the indicator of training programs in firm i, and 𝛽₌𝑜𝑛𝑙𝑦𝑙𝑜𝑐𝑎𝑙 is an indicator variable that takes the value of 1 if the firm operates only in its local market. Additionally, 𝛾₄𝑏𝑟𝑜𝑎𝑑𝑔𝑟𝑜𝑤𝑡ℎ_C is the growth of IT penetration in region r in the 2 year time period prior to the observation, 𝛾_B𝑡𝑒𝑟𝑡𝑖𝑎𝑟𝑦𝑔𝑟𝑜𝑤𝑡ℎ_Cdenotes the growth in the share of persons with tertiary education in region r. The variable 𝛾_M𝑗𝑜𝑏𝑛𝑒𝑤𝑠𝑘𝑖𝑙𝑙_C captures the share of people in the region that think learning new skills is an essential aspect of a job.

4.

A

NALYSIS AND

R

ESULTS 4.1SUMMARY STATISTICS

Table 2 contains summary statistics on the dependent variable based on firm characteristics. A few important observations can be made from Table 2:

First, size has a clear impact on the technology adoption of firms. While in the case of firms with less than 20 full-time employees (the smallest size category), the share of technology adopters is 18,3%, in the largest size category, 37,1% of the firms have adopted foreign technology.

Table 2:Technology licensed from a foreign-owned company (Yes/No)

No Yes Total

Obs. % Obs. % Obs. %

Third, ownership type seems to have a significant effect on technology adoption, which is expected given the fact that the dependent variable captures the licensing of foreign technologies.

Fourth, both investing in training programs and having open-minded management seems to provide significant benefits, as firms that have either (or both) are more likely to be adopters of technology.

Fifth, there is some cross-country variance, which is for most cases only marginal, except for Slovakia, where almost 40% of the firms have adopted foreign technology.

To further investigate the relationship between the dependent variable and the independent variables and also among the independent variables themselves, Table 3 contains pairwise correlations. Table 3: Means and pairwise correlations Mean 1. 2. 3. 4. 5. 6. 7. 1. foreigntech 0.1968 1.0000 2. newprod 0.5370 0.1170* 1.0000 3. onlylocal 0.3704 -0.1103* 0.0360 1.0000 4. compusage 46.717 0.1142* 0.0885* -0.0148 1.0000 5. foreign 0.1304 0.1657* 0.0195 -0.1879* 0.0289 1.0000 6. orgopen 0.4884 0.0880* 0.2288* -0.0285 0.1540* 0.0685* 1.0000 7. training 0.4473 0.1524* 0.0923* -0.0101 0.0731* 0.1343* 0.2650* 1.0000 8. size 84.826 0.0515 -0.0236 -0.0917* -0.0460 0.1559* 0.0390 0.1114* (*p<.05)

Several interesting relationships are apparent from Table 3. First, technology adoption seems to be connected with all the variables, excluding size. Innovation, proxied by newprod is also correlated with computer usage, organizational openness and training programs, meaning that innovating firms tend to: (i) have a more flexible organizational structure, (ii) invest in training programs, (iii) be more ICT intense.

thus less productive. The other is that local firms are not part of a concentrated industry, because such industries serve a large, often international market. According to this logic, foreign firms would establish subsidiaries in the concentrated region to benefit from positive externalities. The fact that local firms are less likely to be owned by foreign entities supports this argument. Third, foreign-owned firms are more likely to have qualities that support the adoption of new technologies, which is in line with the literature, as these firms are likely to experience technology and knowledge spillovers from their parent companies.

Summary statistics on regional technology diffusion is included in the appendix. 4.2RESULTS

Table 4 presents the results of the probit model. Model (0) only contains the control variables. In the models, all but one control variables is significant. As suggested by the literature, firm size, innovation, ownership type, ICT intensity are significant across all countries and industries in the sample. The only control variable that is not significant is orgopen, which stands for the organizational openness of the firm. The insignificance might be due to the fact that organizational openness and innovation capture the same effects. This is supported by the significant correlation between the two variables, as shown in Table 3 above.

Models (1) to (5) test the hypotheses with the same number and Model (6) contains all independent variables. Hypothesis I is rejected as the coefficient broadgrow has the wrong sign and is statistically insignificant. In Model (2) the coefficient of training is significant and positive, thus confirming the importance of improving the employees’ skills by investing in training programs, as proposed by Hypothesis II. The results of Model (3) are not in line

Hypothesis III. The growth rate in the share of population with tertiary education in the region

Table 3: Foreign Technology License (0/1)

(1) (2) (3) (4) (5) (6) (7)

foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech lnsize 0.117*** 0.117*** 0.088** 0.117*** 0.115*** 0.107*** 0.073** (0.035) (0.035) (0.036) (0.035) (0.035) (0.036) (0.037) newproduct 0.334*** 0.334*** 0.316*** 0.335*** 0.310*** 0.341*** 0.296*** (0.093) (0.094) (0.094) (0.094) (0.094) (0.094) (0.095) foreign 0.467*** 0.472*** 0.442*** 0.469*** 0.445*** 0.446*** 0.417*** (0.124) (0.124) (0.124) (0.124) (0.124) (0.124) (0.126) compusage 0.008*** 0.008*** 0.007*** 0.008*** 0.008*** 0.008*** 0.007*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) orgopen 0.129 0.130 0.074 0.130 0.138 0.127 0.086 (0.094) (0.094) (0.096) (0.094) (0.095) (0.094) (0.097) broadgrow -0.005 -0.018 (0.012) (0.012) training 0.302*** 0.306*** (0.095) (0.096) tertiarygrow -0.015 -0.062 (0.045) (0.048) jobnewskills 1.991*** 2.425*** (0.703) (0.767) onlylocal -0.189* -0.194* (0.107) (0.109) _cons -2.431*** -2.382*** -2.420*** -2.369*** -3.558*** -2.318*** -3.249*** (0.303) (0.324) (0.305) (0.355) (0.504) (0.311) (0.524)

Industry FE Yes Yes Yes Yes Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes Yes

N 1254 1254 1254 1254 1254 1254 1254

pseudo R-sq 0.121 0.121 0.129 0.121 0.128 0.124 0.141

4.3ROBUSTNESS CHECKS

To test if the results hold with additional control variables, three different robustness checks were conducted, the results of which are included in the appendix. First the export status of the firm was included in the model, using the variable exporter, which is an indicator variable that takes the value of 1 if the firm’s main product was sold in international markets and 0 otherwise. Including the export dummy only resulted in marginal changes in the coefficients and no negative change in their level of significance, while the dummy variable itself was statistically insignificant in all models. Second, the experience of the top manager was included in the model using the variable expcent, which is the top manager’s years of experience minus the average years of experience of top managers in the sample. Including said variable did not change the results, however, it was statistically significant on the 5% level and negative, meaning that having a top manager with above average experience negatively affects technology adoption. Discussion about why this might be indeed the case, is included in the following chapter. Third, the effect of competition was incorporated into the original model. The variable competition is a score between 1 and 5 and was constructed by dividing the firms into quintiles based on the number of competitors they faced. The value of 1 was assigned to the first quintile, the value of 2 to the second and so on. The coefficient of competition was statistically insignificant in all models. However, as a result of including it in the regression, the variable onlylocal became statistically insignificant. This is in-line with the theory and Hypothesis V as onlylocal also captures competition effects.

Overall we can establish that the results of the original models are robust as including several control variables only marginally impacted the results and the changes are in line with the theory.

4.4DISCUSSION

programs increase technology adoption, is perfectly in line with the SBTC hypothesis. Despite the fact that creating training programs for employees do not require substantial physical investments, it is one of the changes that are most difficult to see through, mainly because such programs were most likely not absent due to lack of resources in the first place. It is more probable that structural inertia and obsolete management practices block the technology adoption processes and thus the development of firms. This is also supported by the fact that when expcent, the centered variable of top manager experience was included in the robustness tests, it was statistically significant on the 5% level and negative. According to the coefficient (Table IV of the appendix) having plus one year over the average of 19 years of experience decreases the probability of technology adoption.

The results also confirm Hypothesis V as they indicate that firms that operate only on local markets are less likely to adopt new technologies by 4.7%. The findings are in line with the literature as the imitative and competitive pressures are low in the case of these firms. Without network effects, they neither get the necessary information about innovations relevant to their production process, nor do they face increased competition due to the lack of peers. However, it would not make much sense to recommend national expansion by all means necessary. Following the logic of the arguments of Hypothesis IV, being constrained to local markets is not a problem to be solved by firms, it is rather a symptom of other ones, like low productivity and poor management. For this reason, firms should rather focus on creating more efficient organizational structures and promote the personal growth of both managers and employees. Additionally, concentration effects are likely to become less relevant (Aiginger & Leitner, 2002) due to the fact that services benefit significantly less from them than manufacturing. From the policymaker’s perspective, the most effective way of overcoming this problem is to follow the ideas of the Smart Specialization concept as this problem is likely to occur more often in less developed regions. Predetermining the specialization of the region in a certain type of industry and designing comprehensive top-down policies are not recommended. Policymakers should rather focus on promoting the networking of firms, collecting information about the technologies that are relevant to the region’s activities, innovation opportunities, exploitable niche markets related to GPTs (co-invention of applications for GPTs) and above all, providing the necessary infrastructure.

the ‘learning culture’ as described by the Regional Innovation Systems concept (Cooke et al., 1997), or to the result of an entrepreneurial process of discovery as proposed by the Smart Specialization concept (Foray et al., 2009), or possibly to both.

The effect of regional IT penetration is insignificant, which is surprising, given the fact that higher firm computer usage significantly increases the propensity to acquire foreign technology in all models – confirming the special role of ICT as a GPT -, even after controlling for industry-specific effects. Therefore, increasing IT penetration in the region should also be significant. The reason why this is not the case might be the existence of lags between increasing IT penetration of regions and the ICT usage of firms. According to this logic, while regional IT penetration does indeed have a significant effect, it only appears with a considerable lag, as it takes time for firms to benefit from it. This would mean that in regions where IT penetration heavily increased 5-10 years ago, firm-level technology diffusion is higher now. Unfortunately, it is impossible to test this, due to the unavailability of data on regional IT penetration.

4.5LIMITATIONS AND FURTHER RESEARCH

5.

C

ONCLUSION

This thesis analyzed the effects of training and education as well as other regional characteristics on firm-level technology diffusion. A substantial part of the findings is in line with the literature as the positive effects firm training programs and the benefits of regional ‘learning cultures’ are also confirmed. While there is evidence on the diverging productivity growth caused by ICT, both on the international (Timmer & Van Ark, 2005) and regional levels (McCann & Acs, 2011), the results of this thesis indicate that regional IT penetration growth in the past 2 years is not accountable for the differences in technology diffusion across regions. The results also indicate that firms operating only locally are exempt from beneficial effects of competition and imitative pressures. Slower technology adoption and productivity growth of isolated firms are in line with the RIS literature. Finally, the effect of regional tertiary education growth is not significant in this model, possibly due to the different mode of knowledge accumulation and educational needs of less developed regions.

APPENDIX

Table I: Main variables regional breakdown

foreigntech _(yes) broadgrowth training(yes) tertiarygrowth jobnewskills _(yes) onlylocal _(yes)

Table II: Foreign Technology Licensing (0/1) Average Marginal Effects

(1) (2) (3) (4) (5) (6) (7)

foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech lnsize 0.029*** 0.029*** 0.022** 0.029*** 0.029*** 0.027*** 0.018** (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) newproduct 0.083*** 0.083*** 0.079*** 0.084*** 0.077*** 0.085*** 0.072*** (0.023) (0.023) (0.023) (0.023) (0.023) (0.023) (0.023) foreign 0.117*** 0.118*** 0.110*** 0.117*** 0.111*** 0.111*** 0.102*** (0.031) (0.031) (0.031) (0.031) (0.031) (0.031) (0.031) compusage 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) orgopen 0.032 0.033 0.018 0.033 0.034 0.031 0.021 (0.023) (0.023) (0.024) (0.023) (0.023) (0.023) (0.024) broadgrow -0.001 -0.004 (0.003) (0.003) training 0.075*** 0.075*** (0.024) (0.024) tertiarygrow -0.004 -0.015 (0.011) (0.012) jobnewskills 0.494*** 0.594*** (0.174) (0.187) onlylocal -0.047* -0.048* (0.027) (0.027)

Industry FE Yes Yes Yes Yes Yes Yes Yes

Table III: Foreign Technology License (0/1): Export Status

(1) (2) (3) (4) (5) (6) (7)

foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech lnsize 0.120*** 0.119*** 0.090** 0.119*** 0.117*** 0.110*** 0.075** (0.035) (0.035) (0.037) (0.035) (0.035) (0.036) (0.037) newproduc t 0.332*** 0.332*** 0.314*** 0.332*** 0.308*** 0.337*** 0.292*** (0.094) (0.094) (0.094) (0.094) (0.094) (0.094) (0.095) foreign 0.479*** 0.484*** 0.454*** 0.481*** 0.457*** 0.465*** 0.435*** (0.127) (0.127) (0.128) (0.127) (0.127) (0.127) (0.129) compusage 0.008*** 0.008*** 0.007*** 0.008*** 0.008*** 0.007*** 0.007*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) orgopen 0.131 0.132 0.075 0.132 0.139 0.129 0.088 (0.094) (0.094) (0.096) (0.094) (0.095) (0.094) (0.097) exporter -0.060 -0.061 -0.059 -0.059 -0.057 -0.104 -0.096 (0.136) (0.136) (0.136) (0.136) (0.136) (0.137) (0.138) broadgrow -0.005 -0.018 (0.012) (0.012) training 0.301*** 0.306*** (0.095) (0.096) tertiarygrow -0.014 -0.061 (0.045) (0.048) jobnewskills 1.988*** 2.411*** (0.703) (0.768) onlylocal -0.202* -0.207* (0.108) (0.110) _cons -2.436*** -2.387*** -2.425*** -2.377*** -3.561*** -2.319*** -3.244*** (0.304) (0.324) (0.306) (0.356) (0.504) (0.311) (0.524)

Industry FE Yes Yes Yes Yes Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes Yes

N 1254 1254 1254 1254 1254 1254 1254

Table IV: Foreign Technology License (0/1) Top Manager Experience

(1) (2) (3) (4) (5) (6) (7)

foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech lnsize 0.118*** 0.118*** 0.087** 0.118*** 0.116*** 0.109*** 0.072* (0.035) (0.035) (0.036) (0.035) (0.035) (0.036) (0.037) newproduct 0.330*** 0.330*** 0.310*** 0.330*** 0.306*** 0.336*** 0.288*** (0.094) (0.094) (0.094) (0.094) (0.094) (0.094) (0.095) foreign 0.443*** 0.448*** 0.413*** 0.444*** 0.419*** 0.422*** 0.388*** (0.124) (0.125) (0.125) (0.124) (0.125) (0.125) (0.127) compusage 0.008*** 0.008*** 0.007*** 0.008*** 0.008*** 0.007*** 0.007*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) orgopen 0.136 0.137 0.078 0.137 0.144 0.133 0.090 (0.095) (0.095) (0.097) (0.095) (0.095) (0.095) (0.098) expcent -0.010** -0.010** -0.012** -0.010** -0.011** -0.010** -0.012** (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) broadgrow -0.005 -0.019 (0.012) (0.012) training 0.321*** 0.325*** (0.096) (0.097) tertiarygow -0.013 -0.061 (0.045) (0.048) jobnewskills 2.024*** 2.488*** (0.703) (0.770) onlylocal -0.180* -0.183* (0.107) (0.109) _cons -2.438*** -2.386*** -2.427*** -2.384*** -3.584*** -2.330*** -3.293*** (0.304) (0.325) (0.307) (0.357) (0.504) (0.312) (0.525)

Industry FE Yes Yes Yes Yes Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes Yes

N 1254 1254 1254 1254 1254 1254 1254

pseudo R-sq 0.125 0.126 0.134 0.125 0.132 0.128 0.146

Standard errors in

Table V.: Foreign Technology License (0/1) Competition

(1) (2) (3) (4) (5) (6) (7)

foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech foreigntech lnsize 0.128*** 0.127*** 0.102** 0.128*** 0.127*** 0.117*** 0.087** (0.041) (0.041) (0.042) (0.041) (0.041) (0.041) (0.043) newproduc 0.315*** 0.314*** 0.304*** 0.316*** 0.291*** 0.319*** 0.276** (0.108) (0.108) (0.108) (0.108) (0.109) (0.108) (0.109) foreign 0.404** 0.414** 0.364** 0.410** 0.382** 0.379** 0.362** (0.165) (0.167) (0.167) (0.165) (0.166) (0.166) (0.170) compusage 0.009*** 0.009*** 0.008*** 0.009*** 0.009*** 0.008*** 0.009*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) orgopen 0.157 0.158 0.109 0.159 0.166 0.155 0.120 (0.105) (0.105) (0.108) (0.105) (0.106) (0.106) (0.109) competitio -0.031 -0.030 -0.029 -0.029 -0.033 -0.033 -0.029 (0.036) (0.036) (0.036) (0.036) (0.036) (0.036) (0.037) broadgrow -0.006 -0.020 (0.013) (0.014) training 0.273** 0.281*** (0.107) (0.108) tertiarygrow -0.027 -0.076 (0.051) (0.054) jobnewskill 1.916** 2.440*** (0.774) (0.854) onlylocal -0.173 -0.175 (0.112) (0.114) _cons -2.473*** -2.422*** -2.470*** -2.365*** -3.552*** -2.348*** -3.238*** (0.362) (0.381) (0.363) (0.413) (0.570) (0.371) (0.593)

Industry FE Yes Yes Yes Yes Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes Yes N 1016 1016 1016 1016 1016 1016 1016

pseudo R-sq 0.133 0.133 0.139 0.133 0.139 0.135 0.151 Standard errors in parentheses

R

EFERENCES

AIGINGER, K., & LEITNER, W. (2002). “Regional concentration in the United States and Europe: Who follows

whom?” Weltwirtschaftliches Archiv, 138(4), 652-679.

ARBO, P., & BENNEWORTH, P. (2007). “Understanding the regional contribution of higher education institutions.”

BATTISTI, G.; STONEMAN, P. (2003) “Inter-and intra-firm effects in the diffusion of new process technology” Research Policy 32 (9), 1641-1655

BATTISTI, G.; STONEMAN, P. (2005) “The intra-firm diffusion of new process technologies” International Journal of Industrial Organization 23 (1), 1-22

BRANSTETTER, L., FISMAN, R., & Foley, C. F. (2005). Do stronger intellectual property rights increase

international technology transfer? Empirical evidence from US firm-level data (No. w11516). National Bureau of

Economic Research.

BUGAMELLI, M.; PAGANO, P. (2004). “Barriers to Investment in ICT” Applied Economics, 36(20), 2275-2286.

DAVID, P., (1969). “A Contribution to the Theory of Diffusion” Mimeo. Stanford University and forthcoming as Chapters 1–3 in his Behind the Diffusion Curve, Westview Press, Boulder, CO.

DAVID, P. A. (2015). “Zvi Griliches and the Economics of Technology Diffusion: Linking innovation adoption,

lagged investments, and productivity growth “(No. 15-005).

DAVID P., AGHION, P., & FORAY, D. (2008). “Science, Technology and Innovation for Economic Growth: Linking Policy Research and Practice” in STIG Systems, Paper No. MPRA Paper, 12096.

DAVIES, S. (1979), “The Diffusion of Process Innovations” Cambridge University Press, Cambridge.

DAVIS, F. D. (1989). “Perceived usefulness, perceived ease of use, and user acceptance of information technology.” MIS quarterly, 319-340.

DE BRUJIN, P., & LAGENDIJK, A. (2005). “Regional innovation systems in the Lisbon strategy” European

Planning Studies, 13(8), 1153-1172.

COHEN, W. M., & LEVINTHAL, D. A. (1989).” Innovation and learning: the two faces of R & D.” The economic

journal, 99(397), 569-596.

COMIN, D.; HOBIJN, B. (2010), “An Exploration of Technology Diffusion.” The American Economic Review COMIN, D.; MESTIERI, M. (2013), “Technology Diffusion: Measurement, Causes and Consequences” NBER Working Paper Series

COOKE, P., URANGA, M. G., & ETXEBARRIA, G. (1997).” Regional innovation systems: Institutional and Organisational dimensions” Research policy, 26(4), 475-491.

EBRD Business Environment and Enterprise Survey, 5th round, Data available at: http://ebrd-beeps.com/data/beeps-v-and-mena-es-2012-2015/

EUROSTAT, General and Regional Statistics, available at http://ec.europa.eu/eurostat/data/database?p_p_id=NavTreeportletprod_WAR_NavTreeportletprod_INSTANCE

_nPqeVbPXRmWQ&p_p_lifecycle=0&p_p_state=normal&p_p_mode=view&p_p_col_id=column-2&p_p_col_count=1

FAGGIAN, A., & MCCAN P. (2006). “Human capital flows and regional knowledge assets: a simultaneous equation approach” Oxford Economic Papers, 58(3), 475-500.

FORAY D., DAVID P. and HALL B. H. (2009) Smart Specialisation – The Concept. Knowledge Economists Policy Brief Number 9, June. European Commission, DG Research, Brussels.

GESIS, Leibniz Institute for the Social Sciences, European Value Survey 2008, available at: http://zacat.gesis.org/webview/index.jsp?object=http://zacat.gesis.org/obj/fCatalog/Catalog5

GATINGTON, Hubert, and Thomas S. ROBERTSON. 1989. “Technology Diffusion: An Empirical Test of Competitive Effects”. Journal of Marketing 53 (1). American Marketing Association: 35–49. doi:10.2307/1251523.

GEROSKI, P. A, (2000), “Models of technology diffusion”, Research Policy 29 (2000) 603-625.

GRILICHES, Z. (1957), “Hybrid Corn: An Exploration in the Economics of Technological Change” Econometrica vol. 25 (October).

GRILICHES, Z. (1969), “Capital-Skill Complementarity” The Review of Economics and Statistics, Vol. 51, No. 4 (Nov., 1969), pp. 465-468

HENDRICKS, K.B.; SINGHALl, V.R., (2008) “The effect of product introduction delays on operating

performance.” Management Science 54 (5), 878–892.

HOCKMANN, H., & PIENIADZ, A. (2008). “Farm heterogeneity and efficiency in Polish agriculture: a stochastic frontier analysis.” In Vortrag präsentiert auf 12th Congress of the European Association of Agricultural

Economists–EAAE (pp. 26-29).

KRUGMAN, P. (1998). “What's new about the new economic geography?” Oxford review of economic

policy, 14(2), 7-17.

LEE, K.; and PALMER, K. (1999). “An Empirical Examination of ISO 9000- Registered Companies in New

Zealand,” Total Quality Management 10(6), 887–899.

LAWTON SMITH, H., TRACEY, P., & CLARK, G. L. (2003). “European policy and the regions: A review and

analysis of tensions.” European Planning Studies, 11(7), 859-873.

LIN, H. F.; LIN, S. M. (2008). “Determinants of e-business diffusion: A test of the technology diffusion

perspective” Technovation, 28(3), 135-145.

KARSHESAS, M.; STONEMAN, P., (1995) Technological diffusion. In:Stoneman, P.: Handbook of the Economics of Innovation and Technical Change. Basil Blackwell, Oxford.

KATRAK, H. (1997).” Developing countries' imports of technology, in-house technological capabilities and efforts: an analysis of the Indian experience.” Journal of Development Economics, 53(1), 67-83.

NELSON, R. R. (Ed.). (1993). ”National innovation systems: a comparative analysis”. Oxford university press. MADSEN, J. B. (2010). The anatomy of growth in the OECD since 1870. Journal of Monetary Economics,57(6), 753-767.

MANSFIELD, E. (1961), “Technical change and the rate of imitation” Econometrica 29 (4), 741–766.


MANSFIELD, E., (1963a), “Intrafirm rates of diffusion of an innovation.” The Review of Economics and Statistics 45 (4), 348–359.

MANUELLI, R. E., & SESHADRI, A. (2014). “Frictionless technology diffusion: The case of tractors.” The

American Economic Review, 104(4), 1368-1391.

MCCANN P. and ACS Z. J. (2011) “Globalisation: countries, cities and multinationals,” Regional Studies 45(1), 17–32.


MCCAN, P., & ORTEGA-ARGILÉS, R. (2013). “Modern regional innovation policy” Cambridge Journal of

MCCAN P., & ORTEGA-ARGILÉS, R. (2015). “Smart specialization, regional growth and applications to European Union cohesion policy” Regional Studies, 49(8), 1291-1302.

McKinsey Global Institute (2015), Global Growth: Can Productivity Save the Day in an Aging World?(http://www.mckinsey.com/~/media/mckinsey/global%20themes/employment%20and%20growth/can%2 0long%20term%20global%20growth%20be%20saved/mgi%20global%20growth_executive%20summary_janua ry%202015.ashx )

METCALFE, S. (1988), “The diffusion of innovation: an interpretative survey” In: Dosi, G., et al. (Eds), Technical Change and Economic Theory. Frances Pinter, London.

MIRON, E., EREZ, M., & NAVEH, E. (2004). Do personal characteristics and cultural values that promote innovation, quality, and efficiency compete or complement each other?. Journal of organizational behavior, 25(2), 175-199.

MIZRUCHI, M. S., & FEIN, L. C. (1999). “The social construction of organizational knowledge: A study of the uses of coercive, mimetic, and normative isomorphism.” Administrative science quarterly, 44(4), 653-683. MELITZ, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695-1725.

MONTALVO, J. G., & YAFEH, Y. (1994).” A microeconometric analysis of technology transfer: The case of licensing agreements of Japanese firms.” International Journal of Industrial Organization, 12(2), 227-244. NELSON, R. R. "National innovation systems." Regional Innovation, Knowledge and Global Change (London:

Pinter, 2000) (1998): 11-26.

PEREZ-ALEMAN, P. (2011). “Collective learning in global diffusion: Spreading quality standards in a developing country cluster.” Organization Science, 22(1), 173-189.

PORTER, M. (1990) “The Competitive Advantage of Nations”, Free Press, New York. PORTER, M., 2001.” Strategy and the Internet.” Harvard Business Review 79 (3), 63–78.

POWELL, W. W., & DIMAGGIO, P. J. (Eds.). (2012). The new institutionalism in organizational analysis. University of Chicago Press.

RAUT, L. K. (1988). “R & D Behaviour of Indian Firms: A Stochastic Control Model.” Indian Economic Review, 207-229.

ROBERTS, B., S. JARVENPPA, C. BAXLEY. (2003) “Evolving at the speed of change: Managing change

ROGERS, E. (1995), Diffusion of Innovations, 4th ed. Free Press, New York.

ROMEO, A. (1975), “Inter-industry and inter-firm differences in the rate of diffusion of an innovation” Review of Economics and Statistics 57, 311–319.

SCHNEPP, O., VON GLINOW, M. A. Y., & BHAMBRI, A. (1990). United States-China Technology Transfer. Prentice Hall.

SCHUMPETER, J. A.,1934,” The Theory of Economic Development: An Inquiry into Profits, Capital, Credit, Interest and the Business Cycle” Trans. Redvers Opie. Cambridge, MA: Harvard University Press.

TIMMER, M. P., & VAN ARK, B. (2005). “Does information and communication technology drive EU-US productivity growth differentials?”, Oxford Economic Papers, 57(4), 693-716.

TORNATZKY, L. G., FLEISCHER, M., & CHAKRABARTI, A. K. (1990). Processes of technological

innovation. Lexington Books.

VISHWASHRAO, S., & BOSSHARDT, W. (2001). Foreign ownership and technology adoption: evidence from Indian firms. Journal of development economics,65(2), 367-387.

VIVARELLI, M. (2013). “Technology, Employment and Skills: An Interpretative Framework” Eurasian Business Review, 3(1), 2013, 66-89

WAN, D., ONG, C. H., & LEE, F. (2005). “Determinants of firm innovation in Singapore”. Technovation, 25(3), 261-268.

WANG, C. L. (2008). “Entrepreneurial orientation, learning orientation, and firm performance.” Entrepreneurship

theory and practice, 32(4), 635-657

WELCH, F. (1970). “Education in Production” Journal of Political Economy, 78(1), 35–59.

ZHU, K.; KRAEMER, K.L.; XU, S. (2006) “The Process of Innovation Assimilation by Firms in Different