WPS3985 Openness and Technological Innovations in Developing Countries Evidence from Firm-Level Surveys Rita Almeida The World Bank and Ana Margarida Fernandes The World Bank Abstract This paper analyzes the role of international technological diffusion for firm-level technological innovations in several developing countries. Our findings show that, after controlling for firm, industry, and country characteristics, exporting and importing activities are important channels for the diffusion of technology. We also find evidence that majority foreign-owned firms are significantly less likely to engage in technological innovations than minority foreign-owned firms or domestic-owned firms. We interpret this finding as evidence that the technology transferred from multinational parents to majority-owned subsidiaries is more mature than that transferred to minority-owned subsidiaries. This finding supports the idea that equity joint ventures maximize technology transfers to local firms. Keywords: Innovation, Technology Adoption, Exports, Imports, Foreign Ownership, Firm Level Data. JEL Classification codes: F1, F2, O3. World Bank Policy Research Working Paper 3985, August 2006 The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Policy Research Working Papers are available online at http://econ.worldbank.org. We thank Robin Burgess, Alberto Salvo, Luis Serven and Jim Tybout and participants at the 2006 International Industrial Organization Conference for their suggestions. Please address correspondence to Ana M. Fernandes, The World Bank, 1818 H Street, NW, Washington DC 20433. Phone: 202-473-3983. Fax: 202-522-3518. E-mail: afernandes@worldbank.org. 1. Introduction Growth theory has for long established that improvements in technology have an effect on long-run growth (Romer, 1990; Aghion and Howitt, 1998). Moreover, differences in technology have been found to be an important determinant of differences in total factor productivity across countries (Hall and Jones, 1999; Klenow and Rodriguez-Clare, 1997) and across firms (Griliches, 1998; Parisi et al., 2006). While most firms improve their technology by simply imitating or adapting existing production techniques to local conditions, other firms are truly engaged in the creation of new technologies. Even though this is true for firms worldwide, this is particularly important in developing countries. There, the adoption of existing technologies from abroad can occur through the contacts with foreign partners, foreign suppliers and/or foreign clients or through the direct trade in technologies (e.g., licensing). Most empirical studies on the international diffusion of knowledge provide indirect evidence on the importance of different channels by documenting a link between trade or FDI and firm-level productivity. In this paper, we use firm-level data on the adoption of new technology in several developing countries to provide new evidence on the incentives of firms to adopt new technology and on the importance that trade and foreign direct investment (FDI) have as channels for the diffusion of technology.1 New technologies can be transmitted across countries through different activities. For- eign direct investment is one possible channel. Multinational parents, endowed with a more advanced technology, usually transfer their knowledge and technology to their subsidiaries. However, there is some debate on the quality of the technology that is actually transferred from multinational parents. Throughout the world, many policies to attract FDI are based on the premise that joint ventures between foreign and domestic-owned firms are more fruit- ful than fully-owned foreign subsidiaries because they induce a greater technology transfer. This view is not consensual, though, as there is some evidence that multinational parents 1The international diffusion of technology to firms integrated into global markets may subsequently spill over to other domestic firms in the same region or industry not directly engaged in international activities. The latter could happen through demonstration effects, labor turnover, or reverse engineering. This paper will focus only on technology transfers to firms integrated into global markets, not on spillovers. 2 have an incentive to transfer older technologies to their subsidiaries in developing countries than to those in developed countries to avoid the risk of expropriation (Mansfield and Romeo, 1980). Others find that technology transfers from parent firms to fully-owned subsidiaries are higher than those to minority-owned subsidiaries (Ramachandran, 1993). Alternatively, the international transmission of technology can occur through trade. If new knowledge is embodied in imported capital goods or intermediate inputs, we would expect importers to be more innovative than firms that source only in the domestic market. Importers can improve their technology by incorporating in their production processes these state-of-the- art inputs or machinery, which may not be available domestically (Grossman and Helpman, 1991). Similarly, exporters can learn about new technologies or products through their inter- action with more knowledgeable foreign buyers in external markets. Alternatively, they may be exposed to more competitive markets and hence be forced to improve their technology frequently. If the exposure to export markets is indeed a channel promoting innovations, we would expect exporters to be more likely to adopt new technologies than firms selling exclusively to the domestic market. Finally, firms can engage in the direct trade of knowl- edge, through technological licensing agreements, to acquire new technology. These channels are not necessarily used in isolation. For example, the use of licensing and FDI are often complementary channels for the international transmission of technology. Perhaps the biggest difficulty in the analysis of the international transmission of technol- ogy is to gather good indicators on the adoption of new technology. How one exactly defines a technological innovation depends on the specificities of each firm and its production process. Most of the empirical work using micro data uses information on firm-level research and development (R&D) expenditures or the number of patented technologies.2 However, the R&D activities of the firm are only one type of research inputs in the process of generating new technologies and they do not necessarily lead to a successful and applicable new tech- nology. Moreover, the propensity to patent varies widely across industries and is likely to be more important for the creation of new knowledge than for the adoption and adaptation 2One exception is Comin et al. (2006) that use a cross-country dataset on the adoption of specific technologies. 3 of existing knowledge. The firm-level dataset we use allows us to construct a broad measure of technological innovation, which includes the creation of new production processes but also the adoption and adaptation of existing technologies to local conditions. In particular, the survey collects information on whether firms introduced recently new technology that substantially changed their main process of production. This measure of technological inno- vation is more likely to reflect advances of the firm towards the country's knowledge frontier, rather than movements of the technological frontier itself. This is arguably a better measure to study technological innovations and their international diffusion, especially in the context of developing countries. Also important to our analysis, the survey collects very detailed information on firm characteristics, including their export and import activities and foreign ownership. Using this firm-level dataset covering 43 developing countries, we will analyze the fol- lowing questions. First, which firms have a larger incentive to engage in technological inno- vations? Second, is the firm's integration into global markets an important channel for the acquisition of new technologies? Which specific channels for the transfer of new technology are explored by those firms that engage in trade or have some foreign ownership? Our paper relates to two strands of literature. First, it relates to the large micro literature that studies the determinants of innovation, which originated in the work by Schumpeter (1942) (e.g., Cohen and Levin, 1989; Cohen, 1995; Aghion et al., 2005). Particularly related to our work is Criscuolo et al. (2005). Like us they explore firm-level data to compute a broad measure of innovation and relate the firm's propensity to innovate with its integration into global markets. Their findings show that those UK. firms that are more integrated into global markets are more likely to innovate, but that most of the difference is explained by the number of scientists and researchers used. Vishwasrao and Bosshardt (2001) also find that foreign-owned firms in India are more likely to adopt new technology than domestic-owned firms.3 Second, our paper relates to the vast literature on the effects of integration into global 3The cross-country evidence also points to a positive correlation between trade openness and the speed at which countries adopt new technologies (Comin and Hobijn, 2004) or invest in R&D (Lederman and Maloney, 2003). 4 markets on productivity (Tybout, 2000; Keller, 2005). One possible channel through which the integration into global markets results in higher productivity is through technological transfers and the resulting access to a wider knowledge base (e.g., Grossman and Helpman, 1991). While most empirical studies look at the effect of the participation in international activities on firm-level productivity, there is little evidence on the actual effect that these activities have on firm-level technological innovations. The evidence available suggests that firm level productivity is higher for those firms integrated into global markets through ex- ports, FDI, or imports of intermediate inputs. In the literature studying the link between firm performance and exports there is strong evidence of self-selection of the most produc- tive firms into exporting (Clerides et al., 1998; Bernard and Jensen, 1999) but there is also evidence of a learning-by-exporting effect (Alvarez and Lopez, 2005; Fernandes and Isgut, 2006). The latter is consistent with the contacts of exporting firms with more knowledgeable foreign buyers generating an increased access to (or demand for) better technologies. Also, firms may be forced to frequently improve their technological capabilities in order to face strong competition in export markets. There is also evidence that firms with FDI tend to be more productive than domestic firms (e.g., Djankov and Hoekman, 2000; Arnold and Javorcik, 2005).4 Finally, there is also some evidence that imports of intermediate inputs are positively correlated with firm and aggregate productivity (Kasahara and Rodrigue, 2005; Coe et al., 1997; Lumenga-Neso et al., 2005).5 The main findings of our paper can be summarized as follows. First, we find that there is a lot of heterogeneity at the firm-level in the acquisition of new technology, even after controlling for differences across countries and industries (accounting for approximately 13% of the total variation). Part of this heterogeneity is explained by the size of the firm and its human capital. For example, medium and large firms are 13 and 18 percentage points more likely to adopt new technology than micro firms, respectively. Second, we find a very 4Arnold and Javorcik (2005) find that a firm's total factor productivity increases significantly following its acquisition by foreign multinationals in Indonesia. 5Coe et al. (1997) and Lumenga-Neso et al. (2005) find that foreign knowledge embodied in imported inputs (from countries with large R&D stocks) has a positive effect on aggregate total factor productivity in developing countries. 5 strong positive correlation between trade and technological innovation. All else constant, importers and exporters are 4.3 and 7.3 percentage points more likely to report technological innovations than firms that do not engage in each of these activities, respectively. While it is possible that these estimates are biased upwards due to the selection of the more competitive firms (and hence more prone to adopt new technologies) into these activities, in our sample this problem does not seem to be too severe for exports. Firms that started exporting more than 10 years prior to the survey are still 4.2 percentage points more likely to report technological innovations than firms that do not export. Third, we find strong evidence that in low-tech industries majority foreign-owned firms are significantly less likely to engage in technological innovations than domestic-owned firms.6 This finding holds when we compare firms with similar managerial education and access to finance operating in the same region and industry. Moreover, majority foreign-owned firms are less likely to engage in technological innovation than minority foreign-owned firms. We interpret this finding as evidence that the technology transferred from multinational parents to majority-owned subsidiaries is more mature than the technology transferred to minority-owned subsidiaries. This provides some support to the idea that equity joint ventures maximize the quality of the technology transferred to local firms. Around the world, policies to attract FDI have been based on this presumption although the little empirical evidence available supports the contrary (Mansfield and Romeo, 1980; Ramachandran, 1993; Javorcik, 2006). The paper proceeds as follows. Section 2 describes the data and provides summary statistics. Section 3.1 describes our main findings on the profile of firms that engage in technological innovations and the importance of the different channels for the international transfer of knowledge. Section 3.2 tests the robustness of the results and Section 4 concludes. 6To the extent that foreigners select the best-performing firms and we are unable to fully account for this through our firm, industry and country controls, our negative point estimate could still be biased upwards. 6 2. Data We use a firm-level dataset collected by the World Bank across 43 developing countries (In- vestment Climate Surveys).7 The surveys were conducted between 2002 and 2005 and the samples were designed to be representative of the population of firms according to their industry and location within each country. Although only one wave of data is available for each country, the information available in the survey has several advantages for analyzing technological innovations. First, it is based on a common questionnaire across a large set of countries, which yields comparable information on several firm-level variables. Among oth- ers, the survey collects information on whether the firm recently adopted new technology, the main channels used to acquire technological innovations, its ownership structure, age, size, human capital composition, and whether it participates in international trade. The survey also collects information on the firm's R&D activities and technology licensing. Table A.1 defines in detail all the variables used in the analysis and Table A.2 reports the summary statistics. The final sample includes 17,723 firms distributed across a wide set of manufactur- ing industries - auto and auto components, beverages, chemicals, electronics, food, garments, leather, metals and machinery, non-metallic and plastic materials, paper, textiles, wood and furniture - in 43 countries in Africa, Asia, Eastern Europe, and Latin America.8 Second, the survey allows us to use a broad definition for technological innovation. Specifically, we define technological innovation as a dummy variable equal to one if a firm reports having introduced new technology that substantially changed the way in which its main product 7The Investment Climate Surveys were conducted in 68 developing countries. We use data for only 43 countries, due to the lack of information on the main variables of interest for the remaining countries. We focus the analysis on manufacturing firms only. The information collected in the surveys is based on a 1.5-2 hours interview with the firm manager. Detailed information on the surveys can be obtained at: http://iresearch.worldbank.org/ics/jsp/index.jsp. In what follows, we will refer to firms as being the unit of analysis but the unit of data collection was actually an establishment or plant. 8The countries included in the dataset are Albania (2005), Armenia (2005), Belarus(2005), Bosnia and Herzegovina (2005) Brazil (2003), Bulgaria (2005), Chile (2004), China (2003), Croatia (2005), Czech Re- public (2005), Ecuador (2003), Egypt (2004), El Salvador (2003), Estonia (2005), Georgia (2005), Guatemala (2003), Honduras (2003), Hungary (2005), Indonesia (2003), Kazakhstan (2005), Kyrgyzstan (2005), Latvia (2005), Lithuania (2005), Madagascar (2005), Malaysia (2003), Moldova (2005), Nicaragua (2003), Philip- pines (2003), Poland (2005), Romania (2005), Russia (2005), Serbia and Montenegro (2005), Slovakia (2005), Slovenia (2005), South Africa (2003), Tajikistan (2005), Thailand (2004), Turkey (2005), Ukraine (2005), Uzbekistan (2005), Vietnam (2005), and Zambia (2002). The year in parentheses indicates when the survey was conducted in each country. 7 is produced in the three years prior to the survey.9 Defining a technological innovation in this way is particularly important in the context of developing countries to understand how firms catch-up to the world technological frontier.10 This variable will capture not only the creation of new knowledge but also the adoption and adaptation of production processes, which although new to the firm, may not be new to the country nor to the world. In our sample, 48% of firms report having conducted R&D activities and, of these, 63.4% report having adopted new technology.11 The fact that 25% of the firms in our sample report having adopted new technology without having conducted any R&D activities reinforces the impor- tance of using alternative measures of technological innovation. Third, the survey collects detailed information on the major channels used by firms to acquire technological innova- tions. On average, 71% of firms in the sample report that their new technology was either embodied in new machinery, developed/adapted within the firm, transferred from the parent company, or developed by key personnel or by consultants. A lower share of firms (32%) reports that technology was acquired or developed in coordination with supplier/clients (in- cluding licensing or turnkey operations from domestic or international sources, developed in cooperation with client firms, or developed with equipment and machinery suppliers) and only 13% of firms report that technology was developed in coordination with other institu- tions (including universities and public institutions, business or industry associations, trade fairs or study groups). One shortcoming of our data is that it captures only the intensive margin of technological 9There is some arbitrariness in this definition which could introduce measurement error in the dependent variable in the regressions in Section 3. If this measurement error is classic, this is not a problem. However, biases could result if the measurement error is systematically related to firm characteristics. For example, if smaller firms report more often that they change their technology, the estimated coefficient for small firms would be biased downwards. 10The survey also collects information on whether the firm developed new major product line(s) or up- graded existing product line(s) in the three years prior to the survey. However, since the focus of this paper is on technological innovations and the channels for technology diffusion, we do not focus on product in- novations. Nevertheless, in Table A.4, we estimate our main specifications using as a dependent variable a dummy variable equal to one if the firm engages in technological innovation or product innovation. Those results provide evidence on the importance of different channels for international knowledge diffusion for the firm's propensity to innovate. 11The finding that almost half of the firms in our sample conduct some R&D activities suggests that managers very likely consider as R&D those activities that relate to the adaptation of technologies to local conditions (in contrast to considering only frontier research activities as R&D). 8 innovation, but does not have information on the extensive margin. This contrasts with the information available in the Community Innovation Surveys (CIS) recently conducted in most European countries by the OECD (e.g., OECD, 1992; Evangelista et al., 1997; Criscuolo et al., 2005; Mohnen et al., 2006). Nevertheless, relative to the CIS datasets, our data has the advantage of collecting information on firm characteristics for all the firms in the sample while in most countries CIS datasets collect that information only for firms that innovate.12 Table 1 reports the average share of firms engaged in technological innovations across continents and industries and Figure 1 shows this average for each country in our sample. The evidence shows that a large share of firms report being engaged in technological inno- vations (56.1%) but also that there is significant heterogeneity across countries.13 Egypt and Uzbekistan show up as the countries with the lowest frequency of technological innova- tions while Thailand and Brazil report the highest frequencies in our sample. Table 1 also shows substantial heterogeneity in the degree of technological innovation across industries. Traditional industries tend to have fewer innovative firms while high-tech industries tend to have more. The most innovative industries are electronics and auto-parts while the least innovative are paper and food.14 To assess whether our innovation measure is economically meaningful, we analyze its correlation with relevant firm outcomes. Figure 2 plots the kernel density of labor produc- 12One exception is the CIS dataset for the UK, used by Cricuolo et al. (2005), which collects information on firm characteristics for all firms, regardless of their innovation status. 13Firms in our sample are more likely to engage in product innovation (68.8%) than in technological innovation (56.1%). This finding holds across industries and countries, with the exceptions of electronics and auto-parts as well as China, Malaysia, and Thailand. But different firms engage in the two types of innovation since the average propensity to engage in technological or product innovation is 77.6%. The average share of firms that engage in both types of innovation is 47.5%, while the probability of engaging in product innovation for firms that have also introduced technological innovations is 84.7%. The latter shows that technological innovations are likely to result in changes in product design and quality and thus new or improved products. However, engaging in technological innovations does not result in product innovation with certainty. 14Using the CIS datasets for 10 countries, Evangelista et al. (1997) find that the average propensity of European firms to introduce process (technological) or product innovations is 53%. The comparable average in our sample is much higher, 78%. This difference holds even within industries. In European countries, the average propensity to innovate in the electronics (textiles) industry is 67% (33%), which compares with an average of 82% (77%) in our sample. The ranking of industries is very similar across the two samples. This difference in the propensity to innovate across the two samples can be explained by managers in developing countries being more likely to report small changes in technology or products as being an innovation than managers in developed countries. 9 tivity (measured as value added per employee) for firms that are engaged in technological innovations and for firms that are not. Figure 2a reports evidence for the country with the lowest frequency of technological innovations, Egypt, while Figure 2b reports evidence for the country with the highest frequency of technological innovations, Brazil. Both figures show that firms that adopt new technology have higher labor productivity than those that do not. This positive correlation suggests that our technological innovation measure captures an economically important activity. Table 2 reports the average frequency of technological innovations for different types of firms. While only 54% of domestic-owned firms report having adopted new technology, foreign-owned firms are substantially more innovative in this respect, particularly those with minority foreign ownership whose propensity to adopt new technology is 74.1%. Also, the share of firms reporting technological innovations is much higher for exporters (64.8%) and for importers (62.7%) than for the full sample.15 These statistics suggest that trade and FDI are associated with more dynamic firms in developing countries. Note that firms integrated into global markets also report higher innovation inputs, measured by the propensity to engage in R&D activities. Hence, it is possible that their higher probability of technological innovation is simply explained by their higher probability of conducting R&D activities. Alternatively, there may be other important factors influencing technological innovation such as the size of the firm.16 This issue will be investigated in the next section. We turn next to the differences across firms engaged in global activities in the chan- nels used to acquire technological innovations. Table A.3 reports the results from probit regressions of each of the sources of technological innovations (shown as column headings) on the foreign ownership variables, and exporter and importer status, controlling for in- dustry fixed effects and for countries' per capita GDP in 1995. For minority and majority foreign-owned firms, transfers from the parent company are significantly more important as 15The higher propensity to adopt new technology of importers and exporters is verified also within indus- tries and countries. 16Table 2 also reports the association between size categories (based on total employment) and technolog- ical innovations. Large firms are substantially more prone to adopt new technologies than smaller firms and the relationship is monotonic. 10 a source of technological innovations than for other firms. The development in cooperation with client firms is a significantly more important source of technological innovations for minority foreign-owned firms than for domestic-owned firms, while the reverse is true for majority foreign-owned firms. Moreover, for majority foreign-owned firms the acquisition of technological innovations from any of the external institutions is significantly less impor- tant than for other firms. For exporters, development in cooperation with client firms and equipment or machinery suppliers is a significantly more important source of technological innovations, but so is the development and adaptation within the firm. Finally, for importers there is a significantly more important role of technology embodied in new machinery and equipment relative to non-importers and a less important role of technology development with client firms. In sum, the findings in this section show interesting patterns in technological innovations and their sources for firms that are integrated into global markets. The rest of the paper tests whether the bivariate correlations found so far resist more rigorous econometric estimation. 3. Empirical Findings 3.1. Main Estimates Our empirical framework considers profit maximizing firms deciding whether or not to engage in technological innovation. A firm decides to innovate if this decision is expected to increase its profits or, in other words, if the benefits from this decision are larger than the costs. Let ijc be the profits of a firm i in industry j in country c. Then, we assume that: 1 if ijc > 0 Innovic = (3.1) 0 otherwise. where Innovijc is a dummy variable that equals one if firm i reports engaging in technological innovation. Since ijc is unobserved, Equation (3.1) cannot be estimated directly. Therefore, we assume that ijc is a function of firm, industry, and country characteristics. In particular, we assume a linear form so that ijc = Xijc + Ij + Ic + ijc, where Xijc is a vector of firm characteristics, Ij are industry fixed effects, Ic are country-level variables, and ijc 11 captures unobserved firm, industry, and country characteristics. For this functional form, the probability that firm i innovates is given by: Pr(Innovijc = 1) = Pr(ijc > -Xijc - Ij - Ic). (3.2) Assuming that ijc are normally distributed, we can estimate Equation (3.2) by maximum likelihood (probit). Standard errors are clustered to allow for the possibility that the ijc are correlated across firms within country-industry cells. Table 3 reports the results from estimating different versions of Equation (3.2) which add progressively firm characteristics to Xijc. For each regressor, we report the marginal effects on the propensity to engage in technological innovation, at mean values. We include in all specifications industry fixed effects to control for differences in the production technology, product demand, and market competition across industries that could facilitate or require a firm to adopt new technology more frequently (Cohen and Levin, 1989).17 Column (1) reports that within industries, foreign-owned firms are more likely to engage in technological innovations than domestic-owned firms. In column (2), when we control for export and importing activities, there is still evidence that minority foreign-owned firms are significantly more likely to innovate but majority foreign-owned firms are now as likely as domestic-owned firms to adopt new technology. Moreover, within industries, exporters and importers are significantly more likely to adopt new technology than firms that do not trade. Since globally-integrated firms may be larger or have better workers than other firms and these advantages may be driving their higher propensity to adopt new technology, columns (3) and (4) include in the regression firm age, size, a public ownership dummy, and proxies for the quality of the firm's human capital. The main findings of a strong positive association between minority foreign ownership, exports, and imports on the one hand, and the propensity to adopt new technology on the other hand are maintained. Controlling for these variables, majority foreign-owned firms are now significantly less likely to adopt new 17There could also be differences across industries in what is defined as a technological innovation. For example, in an industry where there is continuous change, a small technological change may not be considered an innovation. 12 technology than minority foreign-owned or domestic-owned firms. Countries with a more favorable environment for innovation may also offer better export and import opportunities, receive more FDI, and have a more educated workforce. Since several policy and institutional dimensions could be relevant for international activities and could also affect innovation, we control for countries' per capita GDP in 1995 and for country fixed effects in columns (5) and (6), respectively.18 The relationship between firm-level openness and technological innovations remains robust and the magnitude of the effects does not change much relative to the previous columns. The magnitude of the effects in our preferred specification (column (6)) is economically significant. Firms that export are 4.3 percentage points more likely to innovate than firms selling only to the domestic market while importers are 7.6 percentage points more likely to innovate than firms using only domestic suppliers.19 Firms with a minority foreign participation are 6.1 percentage points more likely, while majority foreign-owned firms are 4.4 percentage points less likely to adopt new technology than domestic-owned firms. The findings in this table also show other interesting patterns. First, there is a negative relation between the propensity to engage in technological innovations and the age of the firm. This finding could be the result of some "creative destruction" as young firms are more innovative and dynamic than older firms with weaker learning possibilities (Schumpeter, 1942).20 Second, larger firms are more likely to engage in technological innovations. In our preferred specification, in column (6), small, medium, and large firms are 6.4, 13.3, and 18.3 percentage points more likely to report technological innovations than micro firms, respectively. This size advantage can be the result of economies of scale in the adaptation or development of new technology (e.g., Cohen and Klepper, 1996) or it can reflect the greater capacity of large firms to finance innovation projects in the presence of imperfect financial 18We include past GDP per capita, since current GDP per capita could be correlated with omitted variables in our specification also affecting innovation. 19Our findings are robust to changing the definition of exporters (importers) to those firms that export (import) more than 10% of their output (intermediate inputs). These results are available upon request. 20Klepper (1996) develops a model that yields a negative relation between age and the propensity to innovate. To the extent that technological innovations improve the probability of survival, the negative point estimate reported in Table 3 would be an upper bound relative to the true effect of firm age on innovation (Audrestch, 1995). 13 markets.21 Finally, the results in Table 3 show that the firm's human capital - measured by the incidence of on-the-job training and the percentage of the workforce with more than secondary education - is positively related with the propensity to adopt new technology.22 These results are in line with the idea that a more qualified workforce improves the firm's absorptive capability and reduces the costs of adopting or creating new technologies (Cohen and Levinthal, 1989). Table 3 also shows that public-owned firms are less likely to adopt new technology, although the effect is not statistically significant. This finding could reflect the fact that public-owned firms tend to operate in more protected markets and thus have smaller incentives to innovate. In sum, the main results in this section suggest an important role of trade and FDI for technological innovations at the firm level. The technological advantage of exporters could be the result of their access to more and better knowledge through the interactions with foreign buyers or it could simply reflect a higher pressure to innovate driven by the strong competitive pressures felt in foreign markets.23 Similarly, the technological advantages of importers could reflect a process of reverse engineering of higher quality foreign inputs. This mechanism would allow firms to learn about the embodied technological knowledge which may not be available domestically (e.g., Grossman and Helpman, 1991; Keller, 2005). Minority foreign-owned firms are also more likely than domestic-owned firms to adopt new technology while majority foreign-owned firms are less likely.24 This evidence sheds some doubts on the dynamism imparted by the technology that multinational parents transfer to their fully-owned or majority-owned subsidiaries in developing countries. This finding could be suggesting that parent multinationals transfer more mature technologies less prone to 21Cohen (1995) also finds an innovation advantage for larger firms in developed countries. Cohen and Klepper (1996) develop a model where larger firms have more incentives to innovate since they face lower average fixed (innovation) costs per unit of output. 22A positive correlation between innovation or technology adoption and human capital is documented at the aggregate level by Comin and Hobjn (2004), Lederman and Maloney (2003), and Keller (2005). 23Exporters may also benefit from scale economies in innovation due to their access to larger foreign markets. Hobday (1997) provides evidence for the electronics industry that the large size of export orders received by firms in Asian countries played an important role in encouraging them to undertake R&D. 24Our findings seem to differ from those in Criscuolo et al. (2005) for the UK, where foreign-owned firms are more innovative (in the sense of being more likely to introduce new processes or products) and those in Vishwasrao and Bosshardt (2001) for India, where foreign-owned firms are more likely to adopt foreign technology than domestic-owned firms. 14 adaptations and improvements to their majority-owned subsidiaries than to their minority- owned subsidiaries.25 Thus, our evidence provides some support to the idea that equity joint ventures transfer a more dynamic technology to local firms. This idea has been at the center of policies to attract FDI across many developing countries that favor joint ventures with local partners over majority foreign ownership. However, the evidence so far has been of stronger technology transfers to majority-owned subsidiaries (Mansfield and Romeo, 1980; Ramachandran, 1993; Javorcik, 2006).26 3.2. Robustness Our findings in Table 3 identify important associations between international technology diffusion through trade and FDI and firm-level technological innovations in developing coun- tries. However, some of the estimates may be biased. First, if multinational parents tend to acquire the best - more innovative and productive - domestic firms, then the coefficients on foreign ownership could be biased upwards. This problem is particularly relevant for the coefficient on minority foreign ownership. Second, if technological innovations improve a firm's ability to enter and remain in foreign markets or to buy from foreign suppliers, then our point estimates of the coefficients on exporter and importer status could also be biased upwards. Finally, it is possible that some unobserved variables, such as, e.g., managerial ability, that are correlated with the propensity to innovate also affect the firm's global en- gagement. Since we do not observe firms for more than one period and valid instruments for our variables of interest are difficult to obtain, we address these selection and omitted variables' problems making use of the rich set of information available in each cross-section of the survey and show the results in Table 4. The effect of international activities on innovation may simply reflect better managerial 25The multinational parents' fear of leakage of proprietary knowledge and of the imminent threat of imitation by domestic-owned firms in countries with weaker property rights mentioned by Saggi (2002), seems to influence the nature of the technology transfers to majority-owned subsidiaries but not so much of those to minority-owned subsidiaries. 26Mansfield and Romeo (1980) find that U.S. multinational parents transfer more advanced technologies to fully-owned subsidiaries than to minority-owned subsidiaries. Ramachandran (1993) finds higher technology transfers from multinational parents to fully-owned subsidiaries than to minority-owned subsidiaries in India. Javorcik (2006) finds that multinationals with the most advanced technologies tend to enter into Eastern European countries through majority foreign ownership rather than minority foreign ownership. 15 quality in the firms that are engaged in those activities. More entrepreneurial managers are more likely to engage in technological innovations more often but are also more likely to export or import. In column (1), we add to our baseline specification (column (6) of Table 3) a dummy variable that equals one if the manager has a college or a post-graduate degree. This variable, which proxies for managerial ability, has a positive and significant effect on the firm's propensity to innovate. More importantly, the effects of technology diffusion through trade and FDI on technological innovation are robust to the control for managerial quality. Alternatively, foreign-owned firms or exporters may have easier access to finance through their multinational parents or due to export-promoting policies. If firms are credit con- strained, lower costs or increased access to finance can increase their ability to innovate.27 Hence, the estimated effects of openness on innovation could be largely due to open firms' better access to finance. The findings in column (2) show that access to finance is indeed positively correlated with the propensity to innovate.28 The effects of minority foreign own- ership, exports, and imports on innovation are maintained, suggesting that it is not just a better access to finance that underlies those effects, but likely it is the access by more open firms to a richer knowledge base. Some industries and regions with a large presence of firms integrated into global markets may provide a particularly dynamic environment for innovation to flourish. Thus, it is possible that both the engagement of the firm but also the engagement of other firms in the same industry and region in international activities matter. In column (3), we control for this external effect by including in the regression the share of firms in the industry-region that export and the share of firms in the industry-region that import. It is reassuring to see that the effects of minority foreign ownership, exports, and imports on innovation are robust to this control. The effects of these "spillover" variables are positive and significant for importers, but negative for exporters, suggesting some possible market-stealing effects. 27King and Levine (1993) argue that the development of financial intermediaries reduces the costs of identifying entrepreneurs more capable of generating innovations. Increased access to finance for firms can affect innovation through an improved screening of the quality and probability of success of the projects that are financed as well as through their effects on innovation inputs. 28Qualitatively similar findings are obtained by Ayyagari et al. (2006) using a smaller sample of countries from the Investment Climate Surveys. 16 More generally, in column (4) we address the problem of potential unobservable policy or geographical factors that could affect the degree of firm openness and the propensity to innovate, by controlling for industry-region fixed effects. Again our main findings are robust. To the extent that the degree of competition is constant across industries and regions, this specification controls for the degree of competition faced by firms. However, it is possible that different firms face different competitive pressures. In particular, it is likely that firms operating in more competitive markets face stronger pressures to innovate and may also be more engaged in international activities. To address this possibility, we include an explicit measure of competition in the regression shown in column (5). Based on the total number of competitors, we construct dummy variables indicating whether the firm faces no competition, weak, medium, or strong competition in the domestic market. Even though we find a positive effect of competition on technological innovation (stronger when the number of competitors is between 4 and 21 firms), the findings for the main variables of interest remain robust.29 A firm's geographical location in the country's capital city may enhance its innovation prospects due to the proximity to other firms, suppliers, clients, and other institutions, while also facilitating its access to global markets through exports and imports and increasing its attractiveness for FDI. The association between openness and innovation could therefore be spuriously due to location. However, the effects of minority foreign ownership, exports, and imports on innovation are qualitatively unchanged in column (6) when our main specification is estimated using the sample of firms located outside their country's capital city. Similar results are found when considering only those firms located outside each country's capital and second major cities. The findings in Table 3 show that industries with a higher degree of technological so- phistication (high-tech) face and take advantage of more innovation opportunities than tra- ditional industries with a lower degree of technological sophistication (low-tech). So far, we have controlled for such differences using industry fixed effects. More generally, there may be differences in the role of technology diffusion through trade and FDI for firm-level innova- 29A non-monotonic relation between innovation and competition is also found for the UK. Aghion et al. (2004) show that the effect of concentration on innovation has an inverted "U"-shape. 17 tion depending on the degree of technological sophistication of the industry. In columns (7) and (8), we report the results from estimating our main specification separately for firms in high-tech industries and low-tech industries, respectively. In both high-tech and low-tech in- dustries, importers have a significantly higher propensity to innovate and the coefficients are similar to those in column (6) of Table 3. Exporters also have a higher propensity to innovate in both high-tech and low-tech industries, but the effect is stronger in high-tech industries. In high-tech industries, foreign-owned firms are not more innovative than domestic-owned firms. Finally, in low-tech industries, there is strong evidence that majority foreign-owned firms are less likely to innovate than domestic-owned firms, while minority foreign-owned firms are more likely to innovate than domestic-owned firms, as in Table 3. For our de- veloping countries, these findings suggest that in low-tech industries multinational parents invest in majority-owned subsidiaries to take advantage of the availability of cheap labor, using them simply as export platforms. For this purpose, they set-up and use their more advanced technology to produce but do not devote more efforts to technological innovation than domestic-owned firms. In contrast, the technology transfers to the local partner by multinational parents that invest in minority-owned subsidiaries generate significantly more innovation. While the effect of exports on technological innovations is robust to the control for firm and industry-region characteristics, it may still be biased due to reverse causality. Exports may be facilitated by a firm's past technological choices which in turn may also influence its current technological innovation. We try to address this problem in column (9) by comparing non-exporters with those firms that have entered export markets more than 10 years prior to the survey. For such exporters, the causal effect is more likely to run from exports and the resulting knowledge diffusion to innovation, rather than the reverse. The point estimate in column (9) is very close to that shown in column (6) of Table 3, suggesting that the reverse causality problem may not be too severe. Similar results (available upon request) are obtained when considering only the firms that have entered export markets more than 20 years prior to the survey. As a robustness check to our main findings, we consider a broader measure of innovation 18 which is a dummy variable equal to one if the firm reports having substantially changed its production process (technological innovation) or if the firm reports having introduced a new product line or upgraded an existing product line (product innovation). The results from regressions using this broader measure of innovation are shown in Table A.4. Across all specifications, there is evidence of a strong positive correlation between trade and the propensity to innovate. We also find a strong positive effect of minority foreign ownership on the propensity to innovate, which contrasts with the negative effect of majority foreign ownership. Finally, we investigate to what extent differences in the sources of technology adoption across firms can explain why firms that are more integrated into global markets tend to innovate more. The baseline regression is our preferred specification in column (6) of Table 3. In column (1) of Table 5 we control for the firm's R&D efforts. The R&D activities within the firm may directly improve the likelihood of technological innovation, or may improve the firm's capacity to absorb external knowledge or technology (Cohen and Levinthal, 1989). Even after controlling for differences in R&D activities across firms, we find that the effects of trade and FDI on innovation are robust and similar to those reported in column (6) of Table 3. Moreover, the effect of R&D on technological innovations is positive and significant: firms that engage in R&D activities are 8.8 percentage points more likely to innovate. In column (2), we add to the previous specification a dummy variable for whether the firm has any technological license. The prior here is that firms that directly engage in the trade of knowledge or technology through licenses would be more likely to report technological innovations, all else constant. Although our data does not differentiate whether these licensing agreements are of domestic or foreign origin, for a smaller set of countries we have information on the use of technology licensed from a foreign-owned firm. The results for this smaller sample are reported in column (3). In both cases, the sign and the magnitudes of the trade and FDI variables remain robust. These results suggest that the use of foreign licenses matters for technological innovations but does not differ much for firms that engage in global activities relative to firms that do not.30 30This was also shown in column (4) of Table A3, where firms with foreign participation, exporters, or 19 In sum, the findings in this section show that there is a very robust positive association between trade (either exports or imports) and technological innovations. Minority foreign participations are also a better channel for technology diffusion than majority foreign partic- ipations. Furthermore, majority foreign-owned firms are significantly less likely to innovate technologically than domestic-owned firms, especially in low-tech industries. Although we do not analyze technological spillovers to domestic firms, this finding is consistent with minor- ity foreign participations bringing greater benefits to the domestic economy than majority foreign participations. These patterns do not seem to be driven by innovation inputs such as R&D activities or the use of licensed technology that directly affect technological innovation. 4. Conclusion This paper uses a firm-level dataset for a large number of developing countries to provide new evidence on the importance of technology diffusion through trade and FDI for technological innovations. Unlike most previous studies, we are able to identify whether firms have recently adopted new technology, which is arguably a better measure of a technological innovation for the firm than its R&D activities or its use of patents. Our findings show a very strong positive correlation between trade and technological innovations at the firm level. Importers and exporters are 4.3 and 7.3 percentage points more likely to adopt new technology than firms that do not engage in each of these activities, respectively. While these effects may be biased due to the selection of the more competitive firms (and hence more prone to adopt new technologies) into these activities, our evidence suggests that this problem is not too severe. We also find strong evidence that, all else constant, majority foreign-owned firms are significantly less likely to engage in technological innovations than domestic-owned firms with similar managerial education and access to finance operating in the same region and industry. This result is particularly strong in low-tech industries. We interpret this finding as evidence that the technology transferred from multinational parents to majority-owned subsidiaries is more mature than that transferred to minority-owned subsidiaries. This finding supports the idea that equity joint ventures transfer a more dynamic technology to local firms, which importers do not differ systematically in the use of licenses from abroad. 20 contrasts with previous empirical evidence and provides a rationale for the policies followed around the world to attract FDI. 21 References Aghion, P. and Howitt, P. (1998) Endogenous Growth Theory (MIT Press). Aghion, P., Bloom, N., Blundell, R., Griffith, R., and Howitt, P. (2005) Competition and Innovation: An Inverted-U Relationship. Quarterly Journal of Economics 120(2), pp. 701-728. Alvarez, R. and Lopez, R. (2005) Exporting and Performance: Evidence from Chilean Plants. Canadian Journal of Economics 38(4), pp. 1384-1400. Arnold, J. and Javorcik, B. (2005) Gifted Kids or Pushy Parents? Foreign Acquisitions and Plant Performance in Indonesia. World Bank Working Paper 3597. Audretsch, D., (1995) Innovation, Growth, and Survival. International Journal of Indus- trial Organization 13(4), pp. 441-457. Ayyagari, M., Demirguc-Kunt, A., and Maksimovic, V. (2006) Firm Innovation in Emerg- ing Markets: Role of Governance and Finance. World Bank mimeo. Bernard, A. and Jensen, J. (1999) Exceptional Export Performance: Cause, Effect, or Both? Journal of International Economics 47(1), pp. 1-25. Clerides, S., Lach, S., and Tybout, J. (1998) Is Learning-by-Exporting Important? Micro- Dynamic Evidence from Colombia, Mexico, and Morocco. Quarterly Journal of Economics 113(3), pp. 903-47. Coe, D., Helpman E., and Hoffmaister, A. (1997) North-South Spillovers. Economic Journal 107(440), pp. 134-149. Cohen, W., (1995) Empirical Studies of Innovative Activity, in: P. Stoneman (ed.) Hand- book of the Economics of Innovation and Technological Change. Blackwell (Oxford, UK), pp. 182-264. Cohen, W. and Klepper, S. (1996) Firm Size and the Nature of Innovation within Indus- tries: The Case of Process and Product R&D. Review of Economics and Statistics 78(2), pp. 232-243. Cohen, W. and Levin, R. (1989) Empirical Studies of Innovation and Market Structure, in: R. Schmalensee and R. Willig (eds.) Handbook of Industrial Organization Volume II (Elsevier Science Publishers), pp. 1059-1107. 22 Cohen, W., and Levinthal, D. (1989) Innovation and Learning: The Two Faces of R&D. Economic Journal 99(397), pp. 569-596. Comin, D. and Hobijn, B. (2004) Cross-Country Technology Adoption: Making the The- ories Face the Facts. Journal of Monetary Economics 51(1), pp. 39-83. Comin, D., Hobijn, B., and Rovito, E. (2006) Five Facts You Need to Know About Technology Diffusion. NBER Working Paper 11928. Criscuolo, C., Haskel, J., and Slaughter, M. (2005) Why Are Some Firms More Innov- ative? Knowledge Inputs, Knowledge Stocks and the Role of Global Engagement. NBER Working Paper 11479. Djankov, S. and Hoekman, B. (2000) Foreign Investment and Productivity Growth in Czech Enterprises. World Bank Economic Review 14(1), pp. 49-64. Evangelista, R., Sandven, T., Sirilli, G., and Smith, K. (1997) Innovation Expenditures in European Industry. Final Report to the European Commission DG-XIII, Project EIMS No. 93/54. Fernandes, A. and Isgut, A. (2006) Learning-by-Exporting Effects: Are They for Real? World Bank mimeo. Griliches, Z. (1998) R&D and Productivity: Econometric Results and Measurement Is- sues, in: P. Stoneman (ed.) Handbook of the Economics of Innovation and Technological Change (Blackwell. Oxford, UK), pp. 52-89. Grossman, G. and Helpman, E. (1991) Innovation and Growth in the World Economy (MIT Press. Cambridge, MA). Hall, R. and Jones, C. (1999) Why Do Some Countries Produce So Much More Output Per Worker Than Others? Quarterly Journal of Economics 114(1), pp. 83-116. Hobday, M. (1997) Innovation in East Asia (Edward Elgar Publishing, London). Javorcik, B. (2006) Technological Leadership and the Choice of Entry Mode by Foreign Investors, in: B. Hoekman and B. Javorcik (eds.) Global Integration and Technology Transfer (Palgrave Macmillan and the World Bank), pp. 179-206. Kasahara, H. and Rodrigue, J. (2005) Does the Use of Imported Intermediates Increase Productivity? Plant-Level Evidence. RBC Financial Group Economic Policy Research In- 23 stitute Working Paper 2005-7, University of Western Ontario. Keller, W. (2005) International Technology Diffusion. Journal of Economic Literature 42(3), pp. 752-782. King, R. and Levine, R. (1993) Finance and Growth: Schumpeter Might Be Right. Quar- terly Journal of Economics 108(3), pp. 717-737. Klenow, P. and A. Rodriguez-Clare (1997) The Neoclassical Revival in Growth Eco- nomics: Has It Gone Too Far? in B. Bernanke and J. Rotemberg (eds.) NBER Macroeco- nomics Annual (Cambridge, MA: MIT Press), pp. 73-102. Klepper, S. (1996) Entry, Exit, Growth, and Innovation over the Product Life Cycle. American Economic Review 86(3), pp. 562-583. Lederman D. and Maloney, W. (2003) R&D and Development. World Bank Working Paper No. 3024. Lumenga-Neso, O., Olarreaga, M., and Schiff, M. (2005) On 'Indirect' Trade-Related R&D Spillovers. European Economic Review 49(7), pp. 1785-1798. Mohnen, P., Mairesse, P., and Dagenais, M. (2006) Innovativity: A Comparison Across Seven European Countries. NBER Working Paper 12280. Mansfield, E. and Romeo, A. (1980) Technology Transfer to Overseas Subsidiaries by U.S.-Based Firms. Quarterly Journal of Economics 95(4), pp. 737-750. OECD (1992) OECD Proposed Guidelines for Collecting and Interpreting Technological Innovation Data - Oslo Manual (OECD, Paris). Parisi, M., Schiantarelli, F., and Sembenelli, A. (2006) Productivity, Innovation and R&D: Micro Evidence for Italy. European Economic Review, forthcoming. Ramachandran, V. (1993) Technology Transfer, Firm Ownership, and Investment in Hu- man Capital. Review of Economics and Statistics 75(4), pp. 664-670. Romer, P. (1990) Endogenous Technological Change. Journal of Political Economy 98(5), pp. 71-102. Saggi, K. (2002) Trade, Foreign Direct Investment, and International Technology Trans- fer: A Survey. World Bank Research Observer 17(2), pp. 191-235. Schumpeter, J. (1942) Capitalism, Socialism, and Democracy (Harper. New York). 24 Tybout, J. (2000) Manufacturing Firms in Developing Countries: How Well Do They Do, and Why? Journal of Economic Literature 38(1), pp. 11-44. Vishwasrao, S. and Bosshardt, W. (2001) Foreign Ownership and Technology Adoption: Evidence from Indian Firms. Journal of Development Economics 65(2), pp. 367-387. 25 Figure 1. Percentage of Firms Engaged in Technological Innovation 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% gypt hile ussia atvia onduras & ont. ania zech oldova tem ala bia erz. hina enia frica razil E zbekistanIndonesiaungary C elarus stonia urkey oland H R L M kraine C eorgia am C roatia lbania B B E T P ajikistan rgyzstan V ietnamulgaria om alaysia G Z cuador lovenia C A rm hailand H adagascar azakhstan B ithuania U M icaragua Slovakia E S acedonia A y L hilippines H R ua lSalvador M T U T P G E N M M K K Serbia osnia& SouthA B Source: Authors' calculations using data from the Investment Climate Surveys. Note: Technological innovation is defined in Table A.1. 26 pr inno oductiv vatio logarithm N ote Sour Fig s: ce :A ure 0 .02 .04 .06 ity n) 0 .02 .04 .06 .08 2. . ar of The utho e Labor excl For each labor graphs -4 rs'cal -6 ude pr us cul d co oductiv e Productiv E atio from untr pane ns the y, ity us chniko ing N the ke for on- N on- data Innov ity rneldens observatio inno v Innov -4 vativ kerne from -2 an ati Sal Figure ati Sal Figure d ity e ls the ve es ve ns Innov and w per es 27 estim in ith 2b. s Em atio the no n- optim Investm Firm s Em Firm per 2a. ation pl B pl -2 n. top inno ent oy oy E al ee A Cl raz ee gy llvariab and vativ bandw im (log) il (log) pt ate les botto e 0 m firm iths Innov Innov s ar e 5% de fine of (define to sm Surveys. ati ati 0 ve ve d the d ooth Firm bas Firm in the s s dis Tabl ed e tributio dis on 2 2 A .1. n techno tribut io of n labo logical of the r N Fu In Co T N ote Sour La Medi Sm Mi Im Ex D M Maj Full T ote Sour ll Wood T Paper O No Met L G ce the eat Food El C Bev A du L Eas A A ect ntin cro port port om ino Sam hem uto sia fri able :A ce rge al arm l ori extiles stry atin ca Sam able :A :A n-Met en ron ical erag tern en : A llvariabl :A um an an rM als uth er er estic rity ty d m t: E Forei ple 2.Fi utho ple an her d anufa allic ts d ics s es A eri es ors'cal Foreign-O rm llvariables rs'cal Mach an urope ca 1.Propen ar d Cha cul Furn an uto sity e C ar cul e iture ctur d Pharm om to de gn-O de atio ing atio inery w pon fine w ns ne d en us d ned in Mate ing ts Table racteristics, fine ns Plastic d us aceu in ing Tabl data tical Innovate data e T A s A .1. rials by .1. from from Co 28 the T Percen echno the ntin en Investm Inno 68.4% 63.0% 49.1% 41.4% 62.7% 64.8% 54.0% 74.1% 63.7% 56.1% echno E nga logi tag va tan of 1,484 2,848 2,779 2,494 1,526 4,654 2,577 8,460 2,032 Ob d ent tion logi ge e d calInnova Investm ent 17,723 381 455 975 834 977 Cl 1,1 ca in Cl l Firm im 43 402 1,425 s. im ate ate s tion, Surveys. Percen and Surveys. T Industry echno D Percen R&D 63.0 52.7 40.2 34.2 57.3 57.2 46.2 57.3% 55.4 48.0 oing E tag 56.1% 55.0% % % % % % % % % % R&D e of 51.5% 45.5% 54.7% 52.2% 55.2% 60.0% 53.9% 50.0% 73.4% 54.7% 52.2% 72.5% 60.9% 49.0% 59.9% 37.8% logi nga tag Firm calInnova ge e d of in Firm s tion s Table 3. Determinants of Technological Innovation (1) (2) (3) (4) (5) (6) Majority Foreign-Owned 0.068 -0.014 -0.048 -0.058 -0.053 -0.044 [0.021]*** [0.020] [0.020]** [0.019]*** [0.018]*** [0.017]*** Minority Foreign-Owned 0.179 0.126 0.100 0.091 0.091 0.061 [0.025]*** [0.024]*** [0.025]*** [0.024]*** [0.023]*** [0.022]*** Exporter 0.113 0.063 0.049 0.042 0.043 [0.018]*** [0.016]*** [0.016]*** [0.015]*** [0.012]*** Importer 0.092 0.079 0.067 0.065 0.076 [0.016]*** [0.015]*** [0.015]*** [0.015]*** [0.011]*** Age -0.001 -0.001 -0.002 -0.001 [0.000]*** [0.000]*** [0.000]*** [0.000]*** Small Firms 0.077 0.074 0.072 0.064 [0.027]*** [0.023]*** [0.023]*** [0.020]*** Medium Firms 0.184 0.151 0.152 0.133 [0.027]*** [0.024]*** [0.024]*** [0.021]*** Large Firms 0.223 0.173 0.181 0.183 [0.027]*** [0.023]*** [0.024]*** [0.021]*** Public-Owned -0.051 -0.062 -0.020 -0.015 [0.028]* [0.028]** [0.023] [0.018] Training 0.213 0.208 0.151 [0.016]*** [0.016]*** [0.011]*** Workforce More Second. Educat. 0.079 0.096 0.108 [0.030]*** [0.026]*** [0.019]*** Observations 17,622 17,407 17,165 16,697 16,697 16,697 Industry Dummies Included? Yes Yes Yes Yes Yes Yes Log Past per Capita GDP Included? No No No No Yes No Country Dummies Included? No No No No No Yes Source: Authors' calculations using data from the Investment Climate Surveys. Notes: The dependent variable is technological innovation. The table reports the marginal effects (at mean values) on the firm's propensity to innovate from probit regressions. Clustered standard errors by industry and country are in brackets. * significant at 10%, ** significant at 5%, *** significant at 1%. All variables are defined in Table A.1. Micro firms (with than 10 employees) is the omitted size group. 29 Table 4. Determinants of Technological Innovation - Robustness Sample of Sample Exporters for Excluding Sample of Sample of Full Full Full Full Full More than 10 Firms in High-Tech Low-Tech Sample Sample Sample Sample Sample Years and Country's Industries Industries Non- Capital City Exporters (1) (2) (3) (4) (5) (6) (7) (8) (9) Majority Foreign-Owned -0.039 -0.042 -0.042 -0.036 -0.055 -0.050 -0.037 -0.055 -0.040 [0.019]** [0.017]** [0.017]** [0.018]** [0.021]*** [0.018]*** [0.031] [0.019]*** [0.018]** Minority Foreign-Owned 0.073 0.063 0.063 0.068 0.049 0.057 0.018 0.091 0.050 [0.023]*** [0.023]*** [0.023]*** [0.025]*** [0.025]** [0.024]** [0.043] [0.027]*** [0.023]** Exporter 0.038 0.042 0.049 0.055 0.046 0.040 0.059 0.034 0.042 [0.013]*** [0.012]*** [0.011]*** [0.012]*** [0.013]*** [0.014]*** [0.018]*** [0.016]** [0.013]*** Importer 0.068 0.073 0.071 0.069 0.075 0.082 0.079 0.070 0.082 [0.012]*** [0.011]*** [0.011]*** [0.013]*** [0.012]*** [0.012]*** [0.018]*** [0.013]*** [0.012]*** Manager College Educat. or More 0.051 [0.014]*** Access to External Finance 0.031 [0.010]*** Share of Exporters in Region-Industry -0.063 [0.031]** Share of Importers in Region-Industry 0.100 [0.031]*** Weak Competition 0.044 [0.021]** Medium Competition 0.081 [0.019]*** Strong Competition 0.068 [0.022]*** Observations 13,208 14,609 16,697 15,808 13,743 12,039 5,966 10,731 14,582 Industry Dummies Included? Yes Yes Yes No Yes Yes Yes Yes Yes Country Dummies Included? Yes Yes Yes No Yes Yes Yes Yes Yes Log Past per Capita GDP Included? No No No Yes No No No No No Industry-Region Dummies Included? No No No Yes No No No No No Source: Authors' calculations using data from the Investment Climate Surveys. Notes: The dependent variable is technological innovation. The table reports the marginal effects (at mean values) on the firm's propensity to innovate from probit regressions. Clustered standard errors by industry and country are in brackets. * significant at 10%, ** significant at 5%, *** significant at 1%. All regressions include also firm age, size dummies, a public-owned dummy, a training dummy, and the percentage of the workforce with more than secondary education. No competition is the omitted competition category in column (5). In column (9), the sample includes all non-exporters but only the exporters that began exporting more than 10 years prior to the survey year. All variables are defined in Table A.1. 30 the A indus me N C In Obs Fo T R Im Ex Mi Maj T ll du & pe ote Sour port port rcentag e ofthe regressions an ce D tr echno y va s: stry erv reign nori ori able and lu The :A ountry at utho Du Du T logy er er ty ty 5.T co m echno Li incl untr es)on the depende rs'cal m ions Forei Forei m echno m ude y ies w ar fi nt cul ies logy censi gn-Ow orkforce rm's e In logi al atio In gn-Ow in ng so clu Li bracke prop variabl ns clu w firm e us ned ned calInno ith ag ts ensi is ing ded? ded? censi te ng data va m ore than e,size .* ty ch sig to no tion, nif dum innova logical from R&D icant the 31 seco m te inno [0.015]* [0.011]* [0.012]* [0.022]* [0.016]* ndar ies,a from at10%, 16,697 Yes Yes 0.088 0.070 0.041 0.059 -0.041 (1) ,and y publ prob vatio Investm educati Li * ic- ** sig itre n.The ent Cl ** ** ** ** on. ow nif gre im cens tabl A ned ssi ate es icant [0.023]* [0.012]* [0.011]* [0.013]* e [0.025]* [0.018]* llvariabl dum ons. 15,238 Yes Yes 0.066 0.066 0.039 0.051 -0.048 (2) m C es y,a at5%, lu reports Surveys. 0.212 st the ** ** ** ** * ** ar tr aining e *** ere d m de sig st ar fine gina [0.0 [0.0 [0.0 [0 [0 [0 .021]* d dum nificant andard l Yes Yes 10,204 0. .015]* .028]* 21]* in m errors 099 13]* 0.092 12]* 0.047 0.033 0.056 -0.041 (3) Tabl e y,and at1%. effects ** ** ** * * by (at No GDP (l Va Ac or M Wo Tr La M A Im Ex Dom M M Te Ta te:T og) lu chn chn chn chn w-T gh m R Mor Public ge po W ajor ble pe e Produc Te Sup Te Te Foreign Te R&D Lo Hi Co No, in A -T egi Share of cess Second rkfo rge icro, port inori he rC dde tInno Institution ology plier/C ology ology ology T ech aining ech eak, A Ex to e anager w rce Ow Sm er rter estic ty ity chnologi Va source api s For 1.Va d va Ac Ac Ac Li po ith E w all, re riable ta pe echno petitio E ary Fo Indus n Me on-Indu xte eig str is (log rEm tion qui lients qui qui logy cens dium Co ith nership Me n- tr M riable the ) pl red red red in Industries ie y rters lleg diu ign-Ow Ow calInnova Na L g s (im rnalFina ducation ore De e m m In vestm oyee from from Inte rnally icensi ,and por nc E than ned ned e ng Stron tion g ters) e en ducation ,and finiti ons tCl V Fi in Du Du Du Du D Du pe Du B A Du th ar Du Du Du D Du Y indir Du Du Du fo Du D m Du im ate alues Inte Indi rm from suppli sour cons umm m m m m mm m com e ran educ and ear in rnatio mm ain cator local 's the innova mm m m m um m m m m um m m m m ul uto m m ow m m m m 50, m m m m m m cur priorto Percent pr nom three y trade tion y ers. ces, technologi innova y ta tion y y y y y ation y y y y y y y y y y nts va th and petit y n gem be Su 1995 rv nalFi s) and rency de or e var en . age var ectly. in s s var everages, s Percentage tw ofthe reigners. year varia is varia varia rsonnel,m varia textiles,and w auto- firm in of ble . of varia varia ey ble velope variable ,de is ble ate ts. ood variable surv variable variable variable variable oductw variable in na alsa fairs s or eith riable iable ble iables ble the variable variable een iables ey s ble as unles cons nc the un eithe calinnov eq food, com the expo 51 ia its le stu equa d equa survey. do eq m equ equ equ pr equa s equa r velope er equa min odu dy d ual rials equ or and pone equa m firm and ta s to from l in ations varia em l garm ble to su al ual nt rters l equal 's w equalto equalto equal 15 alto alto equal alto other nt inus lto 200 1 nit s, to to to 1 1 equalto to 1 1 to 1 1 2000 mat 1 s, c (im 1 0 1 1 ced if 0 if on m 1 por 1 orkf the or 1 in w th tr if ur le che if if or ifth ifth year if ifth if10 ifm U . one in one if 32 ise m the e e e if or US lStatistics,IM exchange priorto fur us lto to to lto esti the bc to ent 1 groups coope and bodie 1 is 1 1 if if equ rate e ration eith d if al the the the the 0% the ifthe stat D SD F) of erials survey universit er to adapted e.T ne athe ic the ters) a ce ifthe w m e th fir . fir thr acting arketw w or co m of w w w oug of firm firm r, als fir firm firm in m firm hi share hen firm of tha m e ed (Sou . us ies 1 the us firm hese firm pr greaterth n ree the ing sts h the 's ch of m m s) fin m ov ha th an totalnum th the expor im e 50% introdu rce:W divi loc the introdu the ith and ith if in m cost non- and ithin one ide po than os the obtai has defini m os . its face tot tion eta ana s e 15 m s s 0, ber fi years rm ts firm of pr al de d ced public t of t ned ge fini lli in rts dir 0% im client licensin firm achinery es im techno al 's ior Defin a desig pharm firm ced s ain 0, ances r sh som th or ld De velopm currency coun c its te by new but a in sti firm g or lo new n foll pr bet num 's ofemp star capit e to tr or the aceut ed ow w be has e ectly tota porta gy and capita ted pr or firm th new y's tu ost e ition oduc tution porta s, m ,ortr and som nt tec een r respectiv of alis pla in less 's GDP w or ansf equipm nt w licens R & of seconda e low loyees its dir ow th entIndic relative lnu rnke oductlin opera s, m im y ed D st 1an fir investm un rnalorexte ely. capita technol ic d m ne in ne an be de r from ays ent, ays iv in te in develope ope por erred in ta from hnology expe icals,electr s d d survey. m oni 3, ents in of ersi ry,colle rnaltr by the tions. rm ectly. 50% by lis ogy to the flator tline(s) w from rela w ow wo or busi te a ndit hic nt th ty dom tha d d hic licen Parisietal.(2006 ge h w of th w e firm w h ator the e aterials, e,respective bet throu aini ee gh tr the ,or ng ediate gov is th ned tsubst U (Sou s) S rkers ugrade ness to ith rations from ays pa the sing ures ly ). cs,and n firm aini in by m . to estic 4 ng post- put e in the foreign- 's com bet its w dolla rce: in rent hi eta an ent for d or equipm w ow indus m ernm een s th exi indu firm dom firm ls gr w firm 's Wo e com ring agreemen ned (e.g., paper, d and tr ,colle adu 1 and iti antially r in rld firm st str acqu acqu labor oth 20, ercia ge es. ing l an suppli y ir ent estic hich of ke ir comp tin er m an ate orkers. entorst d eigners. 2000 D conv the d produc ass es te and or pany. y es te the cost m y-region ,or capital s eve anu achin m banks educ ate changed is es is (Source: lopm erte m any. w ore or po 10,bet dir th ow e d tlin chnologi int firm fro achin erna acqui personn chnologi three ith ery. ,exc post-gr ation. w lu le ne w R factur een d ay ent m e(s) ociations, tiona than din sitive ectly or cal ery yea l res elor cal & in rs D g, 20 asin th g g aduate . 11 or by e N Sour ote:A ota e port xpor no ce: Produc Te T Te T T Te T T T Tech T T T T Te T R& St M Weak N ech ech ech ech echn echn ech echno D rong ediu o Co Share Share M A W Trai Pu Lar Me Sm Mi T Ag Im E Mi Ma Tech T ork blic ge all cro ri able llvar A ccess Co -O lE A uthor tInnova m of dium T C to forc ning m er ter ty jority nolo iab gical gical gical gical gical gical echno logy m Ex w m om petitio ofIm anagerw E e po por w ned plo Fore gical 2.D les s' cal chnologicalIn nolo chnologicalIn nolo chnologicalIn nolo nolo nolo nolo In chnologicalIn Foreign m Co L pet pet ith itio n C xtern ith ym Foreign-O ign-O escr In ar itio M e n rters ters def n w w ined culations tion ologicalInn ologicalInn novation novation echnologicalInnova novation echnologicalInnova novati ova ova echnologicalInnova novation logy icensing petitio n in in Re R ollege ore ent iptiv ne ne e Ed in using s Innovations s Innovations tions s tions Innovations Innovations ons tions tions Innovations tions s Table A A A A A A A D D Tran A A D A Em Licensi gi egion alFinance than d d novation Statis ng da ta cqui cqui cqui cqui cqui on-I A bodi ndust -Indu ucation tics .1. from cquire d cquired cquire cquir evelo evelo evelo d ed red cquire d red sferred red red red or ed M the from Invest from In from fr from fr ped ped fr fr ped B ry stry Secondary om om w in Edu ith Co In Su U Consul T B from om om orA y in H N ore Li Li E ew m entC lim stitutions pplie ternally niver rade usi ne Par censi censi dap ted iring M en K cation rs/Cli sit ss quipm operatio ies or en n tC ng ng With ey at tants Fairs w e en Ind om in Perso achinery Surveys. ts orPu and ust torM ith orT orT ur ur or blic Study pan E nne ry ach Clien nk nk y Eq A in ey ey l 33 Institu G sso rou ery tFir stablis O O uipm tions ps ciation Su m s pplier perations perations hm en ent t fr fr om om D om Int est erna ic Sour tiona ces lSour ces 17 10 ,723 ,022 10,0 10 ,022 22 10,022 10,0 10 10,0 ,022 22 22 10,022 10,022 10,022 10,0 10 ,022 22 10,022 10,0 10,022 10,8 16 17 14 14 14 14 17,6 17 14 15 ,051 ,723 ,409 ,409 ,409 ,409 ,667 ,101 ,722 17,106 17 17 17 17 17 17 17 ,723 ,723 ,433 ,433 ,433 ,433 ,433 17,6 17 17 17 ,483 ,723 ,651 1765 17 Ob ,667 22 42 67 58 1 s. 226 0. 0. 69 13 0.32 0. 71 0.00 0.01 0. 03 0.01 0.05 0.07 0.04 0.01 0. 02 0.12 0.10 0. 53 0.13 0. 0. 0. 0. 0. 0. 10 48 38 37 13 13 0.37 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 18 0. 0. 0. 0. 0. M 37 75 67 22 54 08 27 21 36 16 .3 .2 48 40 04 13 56 ean 3 6 D 0. 652. Sta 17. evi - - - - - - - - - - - - - - - - - - - - - - - - - - - 25 - - - - - - 07 02 - - - - - ation ndard Table A3. Sources of Acquisition of Technological Innovations Licensing or Licensing or Developed Embodied Developed or Turnkey Turnkey Developed in with From From in New By Hiring Transferred Trade Fairs Adapted Operations Operations Cooperation Equipment Business or Universities Machinery Key from Parent and Study Consultants Within from from with Client or Industry or Public or Personnel Company Groups Establishment International Domestic Firms Machinery Association Institutions Equipment Sources Sources Supplier (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Majority Foreign-Owned -0.081 -0.009 0.006 -0.004 -0.058 0.191 -0.017 -0.017 -0.004 -0.011 -0.008 -0.004 [0.019]*** [0.012] [0.005] [0.003] [0.012]*** [0.021]*** [0.009]* [0.007]** [0.005] [0.003]*** [0.003]*** [0.001]*** Minority Foreign-Owned -0.059 -0.002 0.010 0.000 -0.039 0.096 0.026 0.006 -0.002 -0.008 -0.004 -0.002 [0.027]** [0.014] [0.008] [0.006] [0.016]** [0.029]*** [0.015]* [0.013] [0.005] [0.006] [0.004] [0.002] Exporter -0.039 -0.008 0.008 0.000 0.003 0.002 0.015 0.013 0.005 -0.004 0.001 0.001 [0.016]** [0.008] [0.004]** [0.003] [0.010] [0.003] [0.006]** [0.006]** [0.003]* [0.003] [0.002] [0.002] Importer 0.049 -0.012 0.005 0.000 -0.014 0.007 -0.025 -0.005 -0.009 0.006 -0.002 -0.001 [0.017]*** [0.009] [0.004] [0.003] [0.011] [0.003]** [0.010]** [0.005] [0.003]*** [0.003]* [0.002] [0.001] Industry Dummies Included? Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Log Past per Capita GDP Included? Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 7329 7329 7329 7329 7329 7329 7329 7329 7177 7177 6785 6366 Source: Authors' calculations using data from the Investment Climate Surveys. Notes: Each column shows the marginal effects (at mean values) from a probit regression of the source of technological innovation shown in the column heading on international activities. Robust standard errors are in brackets. * significant at 10%, ** significant at 5%, *** significant at 1%. All variables are defined in Table A.1. 34 Table A4. Determinants of Product and Technological Innovations Sample of Sample Exporters for Excluding Sample of Sample of Full Full Full Full More than 10 Full Firms in High-Tech Low-Tech Sample Sample Sample Sample Years and Sample Country's Industries Industries Non- Capital City Exporters (1) (2) (3) (4) (5) (6) (7) (8) (9) Majority Foreign-Owned -0.049 -0.039 -0.05 -0.038 -0.050 -0.055 -0.048 -0.051 -0.047 [0.016]*** [0.018]** [0.015]*** [0.018]** [0.017]*** [0.029]* [0.017]*** [0.017]*** [0.015]*** Minority Foreign-Owned 0.031 0.039 0.035 0.048 0.037 0.023 0.037 0.036 0.029 [0.017]* [0.016]** [0.016]** [0.020]** [0.018]** [0.031] [0.018]** [0.016]** [0.017]* Exporter 0.042 0.034 0.038 0.056 0.036 0.062 0.034 0.043 0.040 [0.008]*** [0.009]*** [0.008]*** [0.009]*** [0.009]*** [0.014]*** [0.010]*** [0.008]*** [0.008]*** Importer 0.089 0.081 0.085 0.098 0.091 0.075 0.094 0.097 0.085 [0.008]*** [0.009]*** [0.008]*** [0.009]*** [0.009]*** [0.014]*** [0.009]*** [0.009]*** [0.008]*** Manager College Educat. or More 0.042 [0.009]*** Access to External Finance 0.030 [0.007]*** R&D 0.072 [0.013]*** Observations 16,719 14,618 16,719 14,644 12,042 5,971 10,748 14,602 16,719 Industry Dummies Included? Yes Yes Yes Yes Yes Yes Yes Yes Yes Country Dummies Included? Yes Yes Yes Yes Yes Yes Yes Yes Yes Log Past per Capita GDP Included No No No No No Yes No No No Source: Authors' calculations using data from the Investment Climate Surveys. Notes: The dependent variable is a dummy variable equal to 1 if the firm introduced either a technological or a product innovation. The table reports the marginal effects on the firm's propensity to innovate from probit regressions. Clustered standard errors by industry and country are in brackets. * significant at 10%, ** significant at 5%, *** significant at 1%. All regressions include also firm age, size dummies, a public-owned dummy, a training dummy, and the percentage of the workforce with more than secondary education. In column (8), the sample includes all non-exporters but only the exporters that began exporting more than 10 years prior to the survey year. All variables are defined in Table A.1. 35