Human resource management for learning through knowledge exploitation and knowledge exploration: Pharmaceuticals in Mexico

Structural Change and Economic Dynamics 23 (2012) 530–546

Contents lists available at SciVerse ScienceDirect

Structural Change and Economic Dynamics journal homepage: www.elsevier.com/locate/sced

Human resource management for learning through knowledge exploitation and knowledge exploration: Pharmaceuticals in Mexico Fernando Santiago a,∗ , Ludovico Alcorta b a b

International Development Research Centre, Ottawa, Canada Research and Statistics Branch, UNIDO, Vienna International Centre, Wagramerstr. 5, P-O Box 300 A-1400, Vienna, Austria

a r t i c l e

i n f o

Article history: Received January 2010 Received in revised form June 2011 Accepted November 2011 Available online 8 December 2011 JEL classification: O32, O54, L65 Keywords: R&D Learning and innovation Human resource management Pharmaceuticals Mexico

a b s t r a c t This paper investigates the influence of human resource management practices on the likelihood that a firm performs in-house R&D. R&D is broadly interpreted as learning—a mechanism promoting absorptive capacity and supporting technology capability-building. Firms can choose between two learning strategies: they can exploit existing knowledge, or perform more complex explorations and acquire new knowledge. Different knowledge requirements associate with distinct R&D outcomes with varying degrees of novelty for the firm. Findings are supported with evidence from the pharmaceutical industry in Mexico. The analysis reveals positive linkages between human resource management practices and learning at firm level. The relationship is contingent on factors such as expected R&D outcomes, or the novelty of the knowledge required by the firm. The provision of training revealed the more consistent, positive influence on the likelihood that pharmaceuticals firms perform R&D in Mexico. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Literature on the linkages between human resource management and innovation performance at firm level is growing. Empirical work stems mostly from surveys of firms in developed countries. Scholars have addressed the extent to which sets of new and dynamic work practices influence innovation (Barton and Delbridge, 2001); the effects of distinct forms of labor flexibility on innovation performance (Michie and Sheehan, 1999, 2003), and the complementary relationships between human resource management practices underpinning innovation (Laursen and Foss, 2003). Research on the organization and learning

∗ Corresponding author. Present address: IDRC, 150 Kent Street, PO Box 8500, Ottawa, Canada K1G 3H9. Tel.: +1 613 696 2269. E-mail addresses: [email protected] (F. Santiago), [email protected] (L. Alcorta). 0954-349X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.strueco.2011.11.002

of agents involved in new product development is likewise significant (Lund, 2004a,b). Available literature documents positive relationships between human resource management and innovation performance at firm level. The influence of such practices varies according to the technological dynamics of different industries (Laursen, 2002; Laursen and Foss, 2003), establishment sizes and occupations (Lorenz and Valeyre, 2006), or the way national environments determine learning at individual and organizational levels (Arundel et al., 2007). Still missing, however, is a better understanding of mechanisms to explain such relationships (Laursen and Foss, 2003; Chung-Jen and Jing-Wen, 2009), and a consistent theory on what Delery (1998) termed the “transmission mechanism” from human resource management to innovation performance. Explaining how and why human resource management underpins innovation introduces innovation scholars into the more ample debate about how and why such

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

practices influence firms’ performance more generally. According to Boseli et al. (2005) and Combs et al. (2006) huge challenges stem from the diversity in the number and possible definitions of indicators on human resource management practices, together with the distinct multidisciplinary approaches to research. Arguably research in the field needs to be fine-tuned, specifically in the way the issues at stake are addressed. Lorenz and Wilkinson (2003) assert that researchers frequently assume linear relationships—from adoption of specific sets of management practices to innovation; leaving little room for more heterogeneous organizational strategies within single industries. It is also customary to look at innovation outcomes—products/processes; and their degrees of novelty—radical/incremental. Somewhat understimated is the study of the latent processes associated with the organization of people involved in innovation. Methodologicaly consideration of the intermediate latent processes linking human resource management to a firm’s performance is familiar for management scholars. Sternberg et al. (1997), Amabile (1997) and Mumford (2000), for instance, document how human resource management practices affect creativity and creative thinking. Relatedly Cohen and Levinthal (1989, 1990), Wright et al. (2001) and Chung-Jen and Jing-Wen (2009) assert that human resource management helps to capture and mobilize knowledge residing within and outside organizations. From the above, this paper enquiries about the intermediary factors that link human resource management to innovation. In particular it looks at learning processes supporting absorptive capacity, and the development of innovation capabilities by individuals and, ultimately, organizations. Learning arises from systematic performance of R&D by the firm. In such a way the paper grants research on human resource management practices and innovation greater relevance from a development perspective. White (2002) stressed the pertinence to understand how such practices contribute to research and other technological capabilities, particularly in developing countries. In his view, accumulated capacities can erode because of inadequate or poor management of people. To the best of our knowledge, this paper stems from one of the first systematic studies on the influence of human resource management over learning through R&D in developing countries. Based on literature on knowledge exploitation and knowledge exploration, the hypothesis is that the contribution of human resouces management to learning depends on factors such as the novelty of the knowledge required, and the expected outcomes from inhouse R&D. Empirical evidence refers to pharmaceutical firms in Mexico. In addition to being one of the most advanced developing economies, the country is the world’s ninth pharmaceutical market and the second in Latin America. As such, it has strong, although poorly realized potential to contribute to pharmaceutical innovation. Lack of sufficiently experienced and well trained workforce remains major bottleneck (Guzmán, 2005). Focus on the pharmaceutical industry in Mexico also helps to illustrate the importance of carefully considering the contexts in which human resource management practices work.

531

Macroeconomic conditions, the social environment around R&D, or even how countries contribute to innovation in specific industries dictate not only what is possible and feasible, but what can be expected from human resource management. Better understanding of the organizational practices around pharmaceutical R&D can inform strategies to support the development of human resources for the industry in Mexico and similar countries. The paper proceeds as follows: Section 2 brings together literature on human resource management and learning; the case of pharmaceuticals R&D in developing countries illustrates the discussion. Section 3 characterizes the specific management practices included for the analysis: training, remuneration, and worker’s participation in decision making; these practices are expected to enhance individuals’ and thereby, organizational learning. Section 4 presents the data, defines variables and the corresponding research strategy. Empirical results are provided in Section 5. Finally, Section 6 contains the discussion and conclusions. 2. Human resource management and learning through R&D This paper equates learning with absorptive capacity and capability-building processes by the firm. The literature documents the contribution that organizational practices, relating to R&D and innovation, can make toward the succes of firms. Such practices assist in continuous efforts to mobilize and organize resources that firms have at hand. In the case of Japan, for example, Odagiri (1998) highlighted the importance of building absorptive capabilities, making efforts in training and entrepreneurship and gaining sound scientific and technological understanding; including mastering the production and management of skilled personnel. Hemmert (1998) further underscored human resource management strategies to explain how Japanese firms have dealt with changing, often adverse, macroeconomic environments, and the challenges associated with business strategies posed by continuous technological innovation. Firms constantly reorganize and restructure R&D activities in general, and the management of R&D personnel in particular. Continuous improvement in personnel management underpins innovative organizational practices to promote incentives and motivation for, and productivity in R&D. Accordingly, Lundvall et al. (2002) argued that in addition to R&D efforts, analyses of firms’ innovation capabilities need to consider the influence emanating from the daily experiences of workers, engineers and salesmen, together with interactions among individuals within and outside the boundaries of a firm. Cohen and Levinthal (1989 and 1990)’s treatment of the dual role of R&D as learning mechanism links human resource management to R&D. R&D generates new information and knowledge underpinning searches for new market and technological opportunities through innovation. R&D is equally relevant for assimilating and exploiting existing information and knowledge. In other words, it helps to build absorptive capacity by tapping existing knowledge. The authors further distinguished between

532

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

expected goals from R&D. Firms can exploit existing knowledge bases, or engage in knowledge exploration and expansion of knowledge bases. Cohen and Levinthal stressed that the contribution of individuals’ cognitive processes to accumulate absorptive capacity is contingent on the nature of prior related knowledge and diversity of backgrounds. These elements depend on an individual’s capacity to absorb, assimilate, link, analyse and, eventually, create knowledge. Management scholars integrate the notions of knowledge exploitation and knowledge exploration, as central and distinguishable elements shaping organizational learning and capability-building, in the so-called knowledge-based theory of the firm (March, 1991). The primary role of firms, which is the basis of organizational capabilities, is the integration of specialized knowledge (Grant, 1996). The latter in turn, is often perceived in tacit form, and know-how, skills and practical knowledge embedded in individuals as core components of an organization (Barney, 1991). Succesful innovative firms couple local searches, through internal learning efforts, with distanct searches, knowledge diffusion and assimilation through, for instance, reverse engineering. Firms need to strategically combine stocks and flows of knowledge. Nelson and Winter (1982) argue that, over time, firms gain experience and, eventually, develop routines that increase efficiency and productivity in manufacturing and, in general, the management of current product portfolios. Improvements in products, processes or both are generally based on searches within a firm’s accumulated knowledge. Conversely, the more alien the intended innovation relative to what the firm knows, the larger the need to look beyond familiar cognitive boundaries. Management systems influence and play mediatory roles in these processes; they influence the organization and mobilization of individuals and their knowledge (Barney, 1991). They assist in the creation, transfer and integration of knowledge flows that enrich a firms’ human capital, as a stock (Wright et al., 2001), in ways that are valuable, rare and inimitable (Grant, 1996). That firms engage in either knowledge exploitation or knowledge exploration, or both, illustrates the heterogeneity, complexity and distinct uses of knowledge. Exploitation refers to the use and refinement of existing knowledge, technologies and products. It entails short-run perspectives, more certainty and proximity to potential benefits. Exploration, for its part, identifies searches for new knowledge, use of unfamiliar technologies, creation of products/services with unforeseen, or, at least, difficult to predict, demand (March, 1991; Greve, 2007). Exploration implies long-run mindset, greater uncertainty about future revenues and benefits. Although exploration and exploitation have potentially reinforcing effects on learning and capability-building, they lead to competing resource allocation, increased risks and tradeoffs in investment decisions. Finding the right balance is problematic, the choice of either strategy conditions survival and prosperity of firms: “. . .Systems that engage in exploration to the exclusion of exploitation are likely to find that they suffer the costs of experimentation without gaining many of its benefits. They exhibit too many

undeveloped new ideas and too little distinctive competence. Conversely, systems that engage in exploitation to the exclusion of exploration are likely to find themselves trapped in suboptimal stable equilibria” (March, 1991, p. 71). From the above, and based on Li et al. (2008), a practical interpretation of knowledge exploration and knowledge exploitation is in terms of the cognitive distance between knowledge requirements and a firm’s knowledge base. The latter is characterised by Kale and Little (2007, p. 594) “as simple and complex, based on the technological challenges involved in developing particular products and underlaying capabilities”. Knowledge exploitation refers to local searches for familiar, mature, current or proximate knowledge; it builds on existing technological capabilities. By contrast, knowledge exploration underpins searches for unfamiliar, distant knowledge. The proposed interpretation of innovation draws attention to the learning process occuring inside the firm. It induces some flexibility to the analysis while still capturing traditional views of innovation in terms of incremental and radical outcomes (Greve, 2007). Whereas local searches may lead to incremental innovations, distant searches could lead to radical ones. Nevertheless there is no reason for such match between knowledge searches and innovation outcomes to always occur. 2.1. The case of pharmaceuticals The pharmaceutical industry is illustrative of the issues discused above. Pharmaceuticals are highly R&D intensive; the capacity to perform R&D determines the firm’s viability and capacity to grow. As learning mechanism R&D intertwines capacities to exploit and explore technological and market opportunities. At basic level of technlogical capabilities, R&D supports accumulation of knowledge and experience needed to progressively introduce more sophisticated drugs to the market. Recent experiences in India support this argument. For instance, based on a capability building model, Kale and Little (2007) argued that “reverse engineering R&D capability—the ability to develop products by copying the process-is categorised as basic capability. Generics R&D involves incremental change representing intermediate capability while new chemical entity research involves creating new drugs and innovative therapies representing advanced capabilities” (p. 594). Building on the experience of Indian pharmaceutical firms, the authors illustrated how each stage of capability accumulation makes different demands from a firm’s knowledge base. Over time, firms use, acquire and accumulate different types of knowledge inputs for innovation with increasing degrees of novelty. Progression in the technology ladder accompanies expansion of learning activities outside familiar cognitive boundaries; knowledge searches become increasingly exploratory. Knowledge exploitation, however, remains relevant particularly for firms whose business strategies rely on the extension of life-cycles of existing pharmaceutical products. This experience, together with those presented by Cardinal and Hatfield (2000) and Kim et al. (1997) for example, show that although the technological dynamism of firms in catching

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

up modes generally lags behind that of large multinationals, R&D remains core ingredient for success. The major difference is that, in most cases, R&D in developing countries leads to incremental innovations. This is the case of most R&D by pharmaceutical firms in Mexico. Learning through R&D is relevant even for developing countries specialized in development and marketing of generics drugs. Development of generics starts a few years before patent expiry of the innovator product. Firms have to reproduce the knowledge needed to manufacture it while ensuring bioequivalence and biodisponibility, thus supporting its characteristic as generic interchangeable drug.2 Speed is necessary to the extent that first movers can gain and retain relevant market shares (Caves et al., 1991; Hollis, 2002). In most cases, the choice of products is linked to current product portfolios; what firms already know. Nevertheless, expected benefits increase if firms are able to enhance the characteristics of the innovator drug. Quality enhancement includes relatively simple improvements in product packaging, reformulation or recombinition of existing molecules. New products, in turn, include new applications of existing drugs, often in different therapeutic areas. The search for new knowledge may relate more to the methods and techniques used to synthesize the components—biotechnology techniques, for instance—than to the characteristics of the drug itself (Kale and Little, 2007). 3. Management practices and learning through R&D Section 1 commented on the complexities to define, based on widely accepted theoretical rationale, comprehensive checklists of management practices determining performance at firm level. The literature shows nevertheless that enhanced organizational practices frequently relate to Japanese management styles. Hemmert (1998) for example, indicated that relevant practices targeting R&D personnel include: hiring and firing, job rotation and continuity and compensation systems. Research on complementarities among human resource management practices identifies sets of interventions including: indicators on labor relations—incentives and compensation, recruitment and selection, teamwork, employment security, flexibility in job assignments, training, labormanagement communication and grievance rates (Michie and Sheehan, 1999, 2003; Laursen and Foss, 2003). In developing country contexts Tello and Greene (1996) and Vargas (2004) coincide on such indicators as the provision of training, worker’ empowerment, compensation and staff promotion as part of organizational techniques, such as total quality management (TQM) or just-in-time (JIT). Chung-Jen and Jing-Wen (2009) identified similar practices as mediators in knowledge management processes by Chinese firms. The following paragraphs present some human resource management practices likely to influence

2 Generic interchangeable (GI) denomination indicates that the reaction to a generic drug in the human body is exactly the same as that of an innovator drug.

533

learning within organizations. The discussion informs some hypotheses to be explored in Section 5. The provision of training: Training underpins development of technical and managerial skills among people, who are repositories of the tacit knowledge of an organization (Johnson et al., 1996). Tacit knowledge supports organizational structures, as well as the productive and innovation capabilities of a firm. Training takes two complementary forms: on-the-job and off-the-job. The former is most common. It supports learning of day-to-day operations and an understanding of basic concepts. The second, usually available for key personnel, contributes to enhancing the intellectual capital and skills by capturing existing knowledge, that is, latest developments in specific knowledge fields, research techniques and so on (Hara, 2003). Training contributes to strategies that can be devised to promote motivation and reward human resources. However Gray et al. (2004) stress that the influence of training depends very much on the creation of an environment where sufficient returns on investment in such activity can be expected. It needs to be accompanied by pertinent incentives and working conditions so that improved skills are adequately used (Laursen and Foss, 2003). At global level pharmaceuticals firms are strongly inclined to train personnel across operations (Bureau of Labor Statistics, 2008). Training requirements range from a few hours of on-the-job training to years of formal education, including job experience. Training not only includes development of general skills, but also those needed to carry out specific projects, develop particular processes, conduct specific analyses, handle specialized equipment and so on. Firms frequently train in safety, environmental and quality control and technological advances. Training in marketing and sales is expected to increase the market success of a product (Bureau of Labor Statistics, 2008). From the above, this paper explores the following hypotheses. H1. Training positively influences the likelihood that a firm performs in-house R&D. H2. The nature of training and its impact on learning through R&D will differ depending on the nature of the activities carried out by the firm. Remuneration for performance: Adequate compensation and reward for performance should positively and significantly impact on learning for innovation (Badawy, 1988). Appreciation of individual and professional aspirations promotes motivation and commitment towards an organization (Mumford, 2000; Quinn and Rubb, 2006). Effective reward systems encourage employees to take risks, pursue the development of new products and continuously generate ideas that can be realized (Mumford, 2000). Creativity can be encouraged if freedom, financial rewards, promotion and other forms of recognition exist (Amabile, 1997). Remunerations contribute to skill development cycles (Samstad and Pipkin, 2005); they may also strategically attract talent from outside thereby minimizing costs of internal development (Labarca, 1999). However, setting adequate remuneration systems is complex. Creative individuals may prefer an intellectually challenging environment over high salaries (Terziovski and Morgan,

534

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

2006). These considerations lead to formulate the following hypothesis.

4. Data sources, variable definition and research strategy

H3. In general, remuneration levels should positively influence learning through R&D.

This paper explores the likelihood that a pharmaceutical firm carries out in-house R&D in Mexico. A suitable approach for studying this type of decision variables is a probability model, such as binary probit regression (Greene, 2003). Analysis starts with a basic model that explores the extent to which human resource management practices, controlling for some firm characteristics, explain the likelihood that the firm performs in-house R&D. Then, the definition of R&D is iteratively changed to further identify expected R&D outcomes and corresponding types of knowledge requirements for the firm. Hence one can appreciate how adoption of similar set of human resource management practices influence decisions to conduct R&D for knowledge exploitation or knowledge exploitation, or for new/improved drugs or new/improved drug manufacturing processes.

H4. The influence from remunerations will vary depending on the complexity of knowledge requirements. Worker’s empowerment: Self-esteem—the feeling of power—is an important determinant of employee performance. Empowering employees is basic for highperformance work systems. According to Bartlett et al. (2002), people should be given the opportunity and means to tackle new problems, to gain varied experiences, and to be prepared to take on more challenging tasks. People may participate in the definition of personal objectives, the time they spend at work. Employees should be able to voluntarily involve in assignments that promote skills development, or establishment and management of effective mentoring relationships (Hemmert, 1998). In such a way firms can foster discovery activities (Mumford, 2000). However Bartlett et al. (2002) warn that mismatches between increased responsibility, and means and skills to perform the job can render empowerment meaningless, counterproductive even. Working conditions in the pharmaceutical industry tend to be among the best throughout manufacturing activities. Cleanliness, health and safety are paramount in the industry; worldwide pharmaceuticals firms customarily rank among the best places to work (GPWI). However, it must be acknowledged that strict regulations imposed on the pharmaceutical industry reduce opportunities to modify working conditions. Manufacturing processes and operations in general, must comply with strict good manufacturing practices3 and other quality and safety standards; firms work closely with regulatory authorities. Regarding R&D, the literature documents that drug development activities, such as those underpinning the formulation of generic drugs, are more structured and defined in terms of timing, nature of tasks, formality in the organization and conduction of activities. This is the general situation in a country such as Mexico, were the bulk of R&D leads to the obtaining of generic products. In light of this conflicting evidence, the expectation is as follows. H5. Workers’ empowerment can have positive albeit limited effects on in-house R&D. H6. The effects associated with worker’s empowerment will be contingent on the goals of R&D.

3 In most countries, sanitary authorities ensure effectiveness and safety of pharmaceutical products by implementing comprehensive safeguards and procedures of obligatory observance for drug manufacturers. These are summarized under good manufacturing practice (GMP) which, in simple terms, indicates the best rules/practices to manufacture drugs (FDA, 2004; Seiter, 2005). GMPs include layout and functionality of buildings, qualification and training of personnel, cleanliness and sanitation, monitoring, supervision and many other aspects. GMP’s are reviewed and adjusted according to scientific and technological advances, hence the term “current” or cGMPs.

4.1. Data and data sources Data used in this paper were extracted from the National Survey on Employment, Wages, Technology and Training in the Manufacturing Industry (Encuesta Nacional de Empleo, Salarios, Tecnología y Capacitación; henceforth ENESTYC). This survey was carried out by the National Institute for Geography, Statistics and Informatics in Mexico (Instituto Nacional de Estadística, Geografía e Informática; henceforth INEGI) on behalf of the Ministry of Labor and Social Provision in Mexico (Secretaría del Trabajo y Previsión Social, STPS). ENESTYC represents the entire Mexican manufacturing sector. The manufacturing establishment constitutes the unit of analysis. The survey builds on a stratified sample based on the establishment’s size, as measured by total employment: large 251+; medium: 101–250; small: 10–100 and micro: 0–5. Classification of activities is based on the North American Industrial Classification System (NASCI). Establishments with 100 or more employees are included together with a random sample of those with less than 100 employees. Confidence level is 95 percent, with an estimated non-response of 10 percent. Based on an agreement to comply with pertinent confidentiality requirements by INEGI, personnel from this Institute processed, on our behalf, the preliminary data for the event 2005. The information corresponds to the year 2004. ENESTYC provided information on technological and organizational profiles; employment and remuneration levels; management practices and the provision of training. This paper used data for the module for the pharmaceutical industry (NASCI code 3254). Such data included 141 data points; however, the effective working sample, excluding missing values, is 112 data points. Due to the inability to match data points with specific firms, the remaining part of this paper uses, indistinctly, the terms establishment and firm. However, firms can own more than one establishment. Additional data were collected through exploratory interviews carried out at some firms in Mexico. In total 20 firms, both multinational and of Mexican origin, participated in the study. Interviews were carried out in 2007.

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546 Table 1 Indicators on in-house R&D performance by pharmaceutical firms in Mexico. Variable

Definition

(1) (2)

rd inhouse rd improve process

(3)

rd design meq

(4)

rd drug improvement

(5)

rd drug design

(6)

rd exploit

(7)

rd explore

The firm carries out R&D in-house The goal of R&D is to improve existing manufacturing processes The goal of R&D is to improve or design new machinery and equipment for own use The goal of R&D is to improve existing pharmaceutical products The goal of R&D is to design new pharmaceutical products The firm performs R&D for knowledge exploitation The firm performs R&D for knowledge exploration

Authors, based on ENESTYC 2005, INEGI.

The aim was to learn about the nature of R&D activities and the associated human resource management practices in the local pharmaceutical industry. 4.2. Dependent variables Based on the information from ENESTYC, this paper studies the extent to which pharmaceuticals firms do R&D in Mexico; this is by means of the variable rd inhouse, item (1) in Table 1. It was also possible to identify the objectives of R&D. ENESTYC identifies R&D supporting cost-reducing innovations through: (2) improvements in existing drug manufacturing processes (rd improve process); and (3) improvement or design of new machinery and equipment for the firm’s own use (rd design meq). Variable rd design meq is broadly interpreted as R&D for new process innovation. Alternatively, R&D seeks demand-enhancing innovations including: (4) quality improvements on existing pharmaceutical products (rd drug improvement); and (5) design of new pharmaceutical products (rd drug design). The novelty of the R&D outcomes is defined taking the firm as reference; innovations can be new to the firm but not necessarily to the Mexican market or the world. Based on the discussion in Section 2, items (2) and (4) in Table 1 are interpreted as knowledge exploitation activities, improvements either in pharmaceutical products or drug manufacturing processes or both, lead to searches within familiar knowledge bases. By contrast, the introduction of some new drugs or new drug manufacturing processes, indicators (3) and (5), relate to knowledege searches outside familiar cognitive, including physical and geographical, boundaries of the firm.4 This distinction coincides with Kale and Little’s (2007) differenciation of pharmaceutical firms, based on their accumulated technological capabilities. By combining (2) and (4) a variable on R&D for knowledge exploitation, rd exploit was obtained. Likewise, by combining (3) and (5) the variable on R&D

4 Similar interpretations in the context of biotechnology and pharmaceuticals are found in Rothaermel and Deeds (2004), Gilsing (2006), and Kettler and Modi (2001).

535

for knowledge exploration, rd explore was obtained. In general, firms in Mexico pursue imitative and incremental innovations based on exploitation of already available knowledge. Correlation analysis in Appendix A reveals that knowledge exploitation in general, particularly for the improvement of drugs already in the firm’s portfolio, is what drives R&D of pharmaceutical firms in Mexico.5 There is high and statistically significant correlation between variables denoting R&D for knowledge exploitation. By contrast, the weakest correlations relate to R&D underpinning the design and improvement of machinery and equipment for the firm’s own use. Correlations among variables on R&D reflect the incremental nature of innovation carried out by pharmaceutical firms in Mexico. For example, quality enhancements of pharmaceutical products refer to changes in formualtions so that pharmaceutical products meet requirements of bioequivalence and biodisponibility of the active ingredient. Alternatively, firms seek to improve the packaging of products. New pharmaceutical products, in turn, include new vaccines, new applications of existing drugs by combining excipients, reformulating or recombining existing molecules—often in different therapeutic area—designing novel medical devices and so on. Some local firms develop new generics and excipients based on the application of biotechnology techniques. 4.3. Explanatory variables Table 2 presents the explanatory and control variables used in this paper. As for the former, Boseli et al. (2005, p. 74) acknowledge three forms to measure human resource management variables: “by its presence (that is, a dichotomous scale for whether it is actually in effect ‘yes’ or ‘no’), by its coverage (that is, a continuous scale for the proportion of the workforce covered by it) or by its intensity (that is, a continuous scale for the degree to which an individual employee is exposed to the practice or policy). The overwhelming majority [of studies] rely only on measures of presence.” In general, this is the case with ENESTYC. Only a few variables in the dataset reflect intensity in the use of the corresponding human resource management practice. For example, the indicator on worker’s participation in decision-making. Wright et al. (2001) and Boseli et al. (2005) also advise caution on differences in measuring management variables in terms of either policies or practices. Whereas the former reflect an organization’s stated intentions regarding management activities, the latter are the actual, functioning, observable activities, as experienced by employees. Written policies will influence performance only if

5 To further identify factors driving pharmaceutical R&D in Mexico exploratory factor analysis was conducted including rd process improvement, rd design meq, rd drug improvement and rd drug design; from the analysis one single factor was retained. Within such factor, rd drug improvement had the largest factor loading, albeit somewhat close to those for rd drug improvement and rd drug design. For the sake of simplicity of the analysis the paper reports only results from the correlation analysis.

536

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

Table 2 Human resource management and control variables included in the analysis. Description train04 training internal external training internal external tr ln avg rem rem size imp empowerment Control variables modern practice firm size export dummy fdi fdi expt

1 if the firm provided training to its employees in 2004; 0 otherwise 1 if training is provided by colleagues in-house; 0 otherwise 1 if the firm provides training through external providers (specialized public job training centres, public universities, private universities, other firms, consultants or the industry’s trade organization); 0 otherwise Interaction term between training internal and external training. 1 if the firm provides training both in-house and through external providers; 0 otherwise Natural logarithm of the average remuneration per worker: total remuneration (salaries and benefits) paid in 2004 divided by total number of employees in that same year Interaction term between ln avg rem and firm size as defined below 1 if workers participate in decision making and the firm declares that such practice is important; 2 not important; 0 workers do not participate 1 if the firm reports the use of total quality management and/or just-in-time organizational practices irrespective of actual importance; 0 otherwise Size of the firm 1 = Medium, small and micro, 2 = Large 1 if firms report participation in export markets; 0 otherwise 1 if firms report foreign ownership; 0 otherwise Interaction term between fdi and export dummy. 1 if firms report both foreign ownership and participation in export markets; 0 otherwise

Authors, based on ENESTYC 2005, INEGI. Notes: Information for the 112 data points in working sample. Except for variables train04 and training internal, the rest were created by the authors with information from the source.

individuals perceive them as important for organizational well-being. ENESTYC contains several variables representing regulations on management practices. Unfortunately, since the dataset lacks detailed information on how such rules translate into actual practices, the analysis in this paper focuses on the actual human resource management practices reported by the firm. From the above, and based on Delery’s (1998) the analysis in this paper used alternative constructs of human resource management variables; hence it is possible to capture how different ways to implement a specific practice result in distinct influences on performance at the firm level. The paper includes four alternative definitions of the provision of training, namely, whether training was provided in 2004, train04, and whether it took place in-house, training internal, or through external suppliers, external training. We also used an interaction term between internal and external training, internal external tr. The variable on staff remuneration is expressed as the average of remmunerations, in natural logs, ln avg rem, and by controling for the size of the firm, rem size. Finally, worker’s participation in decision making takes into account the importance of such practice as perceived by the employers, imp empowerment.

4.4. Control variables Arundel et al. (2007) in the case of Europe, OECD (1998) for the OECD countries and Kaplinsky (1995) for developing countries document the interrelation between modern human resource management practices and organizational strategies adopted by firms. Such strategies correspond with the type of management practices available for firms, and shape the environment in which learning takes place (Arundel et al., 2007). In the case of pharmaceutical firms, and in the context of cGMPs, TQM practices assist in meeting the strict quality controls required by

regulatory authorities. In this study, the variable modern practice controls for the use of JIT and/or TQM practices. Capital origin and export behavior condition strongly the technological performance of pharmaceutical firms in ˜ developing countries (Kim et al., 1989; Zúniga et al., 2007). Foreign ownership will determine the perceived importance of R&D for the firm’s business strategy in the host country. In the case of developing countries such as Mexico, multinational affiliates will generaly show rather passive technological behavior, as measured by R&D performance; R&D remains concentrated at the parent company. In the case of exports, systematic R&D efforts assist firms in meeting some of the challenges derived from the increased exposure to competition in foreign markets. This paper incorporates these observations via two dummy variables, fdi and export dumy, respectively.

4.5. Research strategy In conducting the analysis, several checks were performed to ensure accuracy and robustness of results. Models were included, where each dependent variable was regressed on the explanatory variables only; then compared to full specification models. Equations were also run including only those explanatory and control variables that revealed some statistical significance, at five percent or less, in the full specification model. Estimates from all those alternative models are consistent with the results reported in this paper. Note a minor difference in the definition of training used in the model with R&D for new drug manufacturing processes, rd design meq. The majority of pharmaceutical firms reported to have provided training to employees during 2004. Models with train04 had problems converging; the variable predicted perfectly the probability that a firm performs such type of R&D (Long and Freese, 2006). The

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

choice was for the alternative, internal external tr, which is an interaction term between in-house and externaly provided training. Normalizing the log of remunerations with respect to firm’s size (rem size) corrected problems of high and positive correlations between ln rem avg and the variables on FDI and exports, respectively (Appendix A and Table 2). This procedure also captures some scale effects associated with firm size (Cockburn and Henderson, 2001). Hight correlations also involved variables on the use of JIT, fdi and exports. In the first case, high correlation with the indicator on worker’s empowerment was eliminated by using the interaction term, modern practice, between JIT and TQM. Then for fdi and export dummy, the interaction term fdi expt solved some problems. The latter allowed some no-linearities into the models but with a downsize. Stimates including all three variables, fdi, export dummy and fdi expt, were problematic as they tended to predict perfectly the learnimg behavior of firms. Correcting for this—see Long and Freese (2006), was partial solution as colinearity problems increased. In STATA the solution to colinearity is to drop the redundant variable, fdi expt. The paper reports models with and withouth fdi expt; results remain consistent. 5. Empirical results 5.1. Learning behavior of pharmaceutical firms in Mexico Appendix B summarizes the learning behavior of pharmaceutical firms in Mexico. The 74.1 percent of firms performed R&D in 2004, with 63.4 percent and 70.5 percent focusing on process and product innovations, respectively. Of those firms conducting R&D for process innovation, 25.3 percent did so to improve or design machinery for own use, while 63.4 percent to improve drug manufacturing processes. As for demand-enhancing innovations, 61.1 percent of firms pursued new pharmaceutical products, and 66.1 percent focused on improvements in existing drugs. Indicators on sales and employment show that, on average, R&D performers slightly outperform those reporting no R&D. For instance, average employment, total sales and sales per employee are, respectively, 1.4, 1.6 and 1.1 times larger in firms with active learning strategies. By contrast, indicators on capital origin and export orientation tended to favor non-R&D performers. Some 70 percent of firms carried out either knowledge exploitation or exploration. The corresponding figures on employment, sales and so on, are very close among each group, yet with a slight advantage for active learners. Some 60 percent of firms in the sample participated in export markets. However, since the average share of exports in total sales of the industry is rather modest, one can argue that pharmaceutical firms predominantly serve the local market. In line with the cGMP’s requirement, ENESTYC reports extensive adoption of modern manufacturing practices in the pharmaceutical industry. As for the nature of human resource management in the pharmaceutical industry in Mexico, firms show great propensity to provide training to employees. The practice is more frequent in connection with R&D for knowledge

537

exploitation, and by firms pursuing drug innovation. Firms combine both internal and external sources of training in search for synergistic effects between the two types of training. ENESTYC documents that in Mexico, remunerations in the pharmaceutical industry are higher than in other manufacturing industries. They are even higher in firms performing in-house R&D. Nevertheless, our interviews suggest that, as a mechanism to motivate and retain workers, adjustment in remunerations is frequently limited by the need to maintain balance of the firm’s overall compensations structure. ENESTYC also documents that firms in the local pharmaceutical industry seldom allow worker participation in decision-making about working conditions. Even in those areas where workers have a voice, the practice is reported as having little importance for the company. In this regard the exploratory interviews revealed that in firms with some incipient R&D efforts, R&D staff frequently subordinate to manufacturing and quality control. 5.2. Econometric analysis 5.2.1. Knowledge exploitation or exploration The literature suggests that knowledge exploration, in the sense of research, experimentation and technological capability-building, should associate with stronger exigencies on human resource management. This is as compared to knowledge exploitation. This section explores this hypothesis. Table 3 presents estimates from the econometric analysis. Model (1) corresponds to in-house R&D, irrespective of the goal pursued by the firm. Then, model (2) captures R&D for knowledge exploitation (rd exploit) and finally, model (3) identifies R&D for knowledge exploration (rd explore). Each model in the table splits in four sections: models with explanatory variables but without controls, and then those with the full set of variables. In order to ensure that correlation between fdi and export dummy does not cause major problems, a third column includes models with the interaction term fdi expt. Finally, the computation of marginal effects for the model with the full set of explanatory variables is presented. The Wald tests for the value of X2 , which is different from zero, confirm that the models are statistically significant at standard confidence levels. The count R2 show that, in general, the predictive power of each model is acceptable (Long and Freese, 2006). The values of the Cragg–Uhler R2 suggest that the models adequatly explain the probability that a firm performs R&D. Individual estimates reveal that training has the most significant effect on learning through different kinds of R&D; yet the effect looks fairly similar in the case of both knowledge exploitation and knowledge exploration. Remunerations in turn, show positive impact on rd explore but the effect seems not to be robust. Scale effects are also captured by the variable on remunerations, as it is normalized by the firm’s size. Export participation and foreign ownership report relevant influences on R&D performance, albeit with effects that run in opposite directions. Whereas export participation induces learning, foreign ownership inhibits it. In fact, the influence of capital ownership is

538

Table 3 Influence of management practices on knowledge exploitation and exploration by pharmaceutical firms in Mexico. (1) rd inhouse

(2) rd exploit

(3) rd explore

Mg effect train04 rem size imp empowerment

1.25*** (0.41) 0.04 (0.05) −0.11 (0.20)

modern practice export dummy fdi

1.32*** (0.44) 0.06 (0.06) −0.14 (0.23) 0.26 (0.35) 0.71** (0.35) −0.99*** (0.38)

−0.63 (0.46)

−1.04** (0.51)

−0.68** (0.34) −0.98* (0.52)

−58.6 10.5 [3]** 0.137 0.777

−55.1 19.7 [6]*** 0.218 0.768

−56.5 14.0 [5]** 0.186 0.768

fdi expt Constant Observations Log Likelihood full X2 Cragg–Uhler R2 Count R2

1.30*** (0.43) 0.09 (0.06) −0.16 (0.23) 0.28 (0.34)

0.48*** (0.15) 0.18 (0.02) −0.04 (0.07) 0.08 (0.11) 0.22** (0.11) −0.33** (0.13)

Mg effect 1.18*** (0.41) 0.04 (0.05) −0.19 (0.19)

1.27*** (0.43) 0.08 (0.06) −0.18 (0.22) 0.17 (0.35) 0.46 (0.35) −0.94** (0.37)

– –

−0.62 (0.45)

−1.01** (0.50)

– – – –

−62.0 9.84 [3]** 0.123 0.750

−58.9 16.1 [6]** 0.194 0.750

1.28*** (0.43) 0.10 (0.06) −0.20 (0.22) 0.20 (0.34)

0.47*** (0.15) 0.03 (0.02) −0.06 (0.07) 0.06 (0.12) 0.15 (0.11) −0.33** (0.13)

Mg effect 1.05** (0.42) 0.06 (0.04) 0.01 (0.19)

−0.81** (0.34) −1.01* (0.51)

– –

−1.04** (0.47)

112 −59.1 14.3 [5]** 0.190 0.741

– – – –

−69.0 9.06 [3]*** 0.117 0.679

1.22** (0.49) 0.10* (0.06) 0.023 (0.22) 0.29 (0.32) 0.98*** (0.36) −1.39*** (0.41)

1.15** (0.47) 0.14** (0.06) −0.010 (0.21) 0.33 (0.31)

0.46*** (0.16) 0.04* (0.02) 0.01 (0.08) 0.11 (0.12) 0.36*** (0.12) −0.51*** (0.13)

−1.69*** (0.60)

−0.96*** (0.34) −1.56*** (0.57)

– –

−61.5 20.4 [6]*** 0.275 0.741

−64.6 14.9 [5]** 0.213 0.696

– – – –

Authors, based on ENESTYC 2005, INEGI. Notes: Robust standard errors in parentheses. Degrees of freedom within squared brackets. MgEffects: computation of marginal effects correspond to the variables in the full specification model. For variables definitions see Tables 1 and 2. * p < 0.1. ** p < 0.05. *** p < 0.01.

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

Variables

Table 4 Results from probit analysis: management practices and learning in the Mexican pharmaceutical industry. (4) rd improve process

train04

1.13*** (0.42)

1.10** (0.44)

(5) rd design meq

internal external tr rem size imp empowerment

0.02 (0.04) −0.04 (0.19)

modern practice export dummy fdi

0.06 (0.05) −0.13 (0.21) 0.51 (0.32) 0.28 (0.31) −0.70** (0.34)

Constant Observations Log likelihood full X2 Cragg–Uhler R2 Count R2 a

0.08 (0.05) −0.15 (0.21) 0.53 (0.32)

−0.81* (0.46)

−1.26** (0.51)

−0.67** (0.32) −1.27** (0.51)

−69.3 8.11 [3]** 0.101 0.688

−66.3 13.6 [6]** 0.166 0.661

−66.0 13.6 [5]** 0.172 0.670

fdi expt

(6) rd drug improvement

1.12** (0.44) 0.05 (0.05) 0.53*** (0.20) 0.42** (0.20)

0.45** (0.19) 0.11** (0.05) 0.74*** (0.21) −0.29 (0.35) 0.21 (0.36) −0.95** (0.44)

0.47** (0.20) 0.13** (0.05) 0.74*** (0.22) −0.28 (0.36)

−2.65*** (0.58)

−2.94*** (0.64)

−1.03** (0.45) −3.01*** (0.67)

−45.7 16.5 [3]*** 0.225 0.795

−42.8 24.1 [6]*** 0.295 0.795

−42.2 23.8 [5]*** 0.307 0.821

(7) rd drug design

0.95** (0.41)

1.01** (0.43)

1.03** (0.43)

1.05** (0.42)

1.16** (0.50)

1.08** (0.47)

0.06 (0.04) −0.13 (0.19)

0.10* (0.05) −0.13 (0.21) 0.16 (0.32) 0.31 (0.32) −0.76** (0.35)

0.12** (0.05) −0.14 (0.21) 0.18 (0.32)

0.07 (0.04) −0.07 (0.19)

0.11** (0.06) −0.12 (0.21) 0.48 (0.32) 1.01*** (0.36) −1.39*** (0.42)

0.15*** (0.06) −0.15 (0.21) 0.51 (0.32)

−0.77* (0.45) 112 −67.6 8.25 [3]** 0.100 0.696

Authors, based on ENESTYC 2005, INEGI. Notes: Robust standard errors in parentheses. Degrees of freedom within squared brackets. For variables definitions see Tables 1 and 2. a Percentages. * p < 0.1. ** p < 0.05. *** p < 0.01.

−1.11** (0.50)

−0.73** (0.33) −1.13** (0.51)

−1.09** (0.47)

−1.80*** (0.61)

−0.95*** (0.35) −1.66*** (0.59)

−65.3 13.2 [6]** 0.150 0.679

−65.1 12.7 [5]** 0.156 0.679

−69.4 9.54 [3]*** 0.121 0.670

−61.1 20.4 [6]*** 0.290 0.723

−64.5 14.4 [5]** 0.224 0.732

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

Variables

539

540

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

Table 5 Changes in probabilities and marginal effects for models in Table 3.

rd impr proc train04 rem size imp empowerment modern practice export dummy fdi rd design meq internal external tr rem size imp empowerment modern practice export dummy fdi rd drug imp train04 rem size imp empowerment modern practice export dummy fdi rd design train04 rem size imp empowerment modern practice export dummy fdi

(1) min → max

(2) 0 → 1

(3) −+1/2

(4) −+sd/2

(5) MargEfcta

0.418 0.199 −0.101 0.192 0.103 −0.265

0.418 0.025 −0.050 0.192 0.103 −0.265

0.394 0.024 −0.050 0.187 0.103 −0.255

0.127 0.070 −0.035 0.089 0.052 −0.120

0.412 0.024 −0.050 0.189 0.103 −0.260

0.188 0.218 0.420 −0.062 0.042 −0.160

0.027 0.008 0.154 −0.062 0.042 −0.160

0.093 0.023 0.152 −0.059 0.042 −0.197

0.099 0.069 0.106 −0.028 0.021 −0.091

0.092 0.023 0.151 −0.059 0.042 −0.195

0.387 0.295 −0.093 0.057 0.112 −0.283

0.387 0.039 −0.045 0.057 0.112 −0.283

0.354 0.036 −0.045 0.057 0.111 −0.269

0.113 0.107 −0.032 0.027 0.056 −0.127

0.366 0.036 −0.045 0.057 0.111 −0.274

0.437 0.341 −0.092 0.181 0.368 −0.510

0.437 0.042 −0.045 0.181 0.368 −0.510

0.413 0.042 −0.045 0.176 0.364 −0.484

0.134 0.124 −0.032 0.084 0.187 −0.238

0.433 0.042 −0.045 0.177 0.377 −0.518

(6) MargEfctb , c 0.642 0.418 (0.146)*** 0.024 (0.019) −0.050 (0.080) 0.192 (0.123) 0.103 (0.117) −0.265 (0.128)** 0.124 0.092 (0.035)*** 0.023 (0.011)** 0.151 (0.050)*** −0.062 (0.081) 0.042 (0.071) −0.160 (0.062)** 0.672 0.387 (0.152)** 0.036 (0.019)* −0.045 (0.076) 0.057 (0.119) 0.111 (0.117) −0.283 (0.132)** 0.642 0.437 (0.161)*** 0.042 (0.021)** −0.045 (0.079) 0.180 (0.121) 0.368 (0.123)*** −0.510 (0.136)***

Authors, based on ENESTYC 2005, INEGI. Min → Max: change in predicted probability as x changes from minimum to maximum; 0 → 1: change in predicted probability as x changes from 0 to 1; −+1/2: change in predicted probability as x changes from 1/2 unit below base value to 1/2 unit above; −+sd/2: change in predicted probability as x changes from 1/2 standard deviation below base to 1/2 standard deviation above; MargEfct: partial derivative of the predicted probability/rate with respect to a given independent variable. For variables definitions see Tables 1 and 2. a Computed based on the method of discrete changes. b Computed based on the method of marginal changes; robust standard errors in parentheses. c Changes for binary variables from 0 to 1. * Significance at the 10% level. ** Significant at the 5% level. *** Significant at the 1% level.

stronger than that of exports. This is evident by looking at models with fdi expt as explanatory variable. Modern practice and imp empowerment did not reveal any specific effect on learning. Overall, the estimates suggest a passive learning behavior of the pharmaceutical industry in Mexico. The constant term is consistently negative and statistically significant. If all right-hand side coefficients were set at zero, the probability that a firm carries out R&D is rather low. A complementary way to look at results from probit models is by computing the marginal effects associated with modifications in the value of a given explanatory variable (Christofides et al., 1997, 2000). A fourth column for each model in Table 3 presents the marginal and discrete probability changes for the variables in the full specification models. Estimates confirm that training has positive and statistically significant impacts on learning. If all remaining variables in the equation are left constant, in this case at their mean value, the shift from non- to provision of training increases, by some 48 percent, the probability that a firm carries out in-house R&D. The effects of training associated with knowledge exploitation were not statistically different from those of training in the context of R&D for knowledge exploration.

Contrary to our research hypothesis, findings in Table 3 suggest that the influence of human resource management on the likelihood that a pharmaceutical firm does R&D is rather limited. Moreover, it is difficult to perceive how distinct types of learning activities associate with different human resource management practices. Based on the discussion in Section 2, in what follows a further distinction is made on the innovation outcomes expected from R&D.

5.2.2. Learning through different kinds of R&D Table 4 presents results from models that incorporate distinct goals pursued through in-house R&D. Models (4) and (5) include cost-reducing R&D, while models (6) and (7) relate to demand-enhancing R&D. Customary indicators on goodness of fit corroborate the adequacy of the models. As expected, the more detailed definitions of R&D provide better information on how the contribution of management practices to learning differs. Relevant practices vary both in number and strenght of the corresponding effects. This supports the idea that firms with divergent learning and innovation strategies should gain differently from adoption of even comparable human resource management practices (Laursen and Mahnke, 2001; Laursen and Foss, 2003).

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

541

Table 6 Testing the influence of internal and external training on performance of in-house R&D. Variable

rd inhouse

rd exploit

rd explore

rd improve process

rd drug design

rd drug improvement

training internal

0.58* (0.33) 0.63** (0.31) 0.02 (0.06) −0.11 (0.24) 0.31 (0.35) 0.71** (0.33) −0.94** (0.38) −0.56 (0.44)

0.70** (0.32) 0.46 (0.31) 0.04 (0.06) −0.15 (0.23) 0.22 (0.34) 0.42 (0.34) −0.89** (0.38) −0.57 (0.43)

0.31 (0.33) 0.79** (0.32) 0.06 (0.06) 0.035 (0.22) 0.34 (0.31) 0.99*** (0.33) −1.31*** (0.40) −1.21*** (0.44)

0.59* (0.32) 0.50* (0.31) 0.03 (0.05) −0.12 (0.22) 0.54* (0.32) 0.23 (0.32) −0.63* (0.35) −0.90** (0.42)

0.24 (0.33) 0.75** (0.32) 0.08 (0.06) −0.11 (0.22) 0.52* (0.32) 1.01*** (0.33) −1.29*** (0.41) −1.32*** (0.45)

0.40 (0.32) 0.35 (0.31) 0.08 (0.06) −0.11 (0.21) 0.21 (0.32) 0.28 (0.31) −0.69* (0.36) −0.68* (0.41)

−55.5 23.2*** 0.207 0.777

−59.3 15.6** 0.185 0.741

−60.9 30.1*** 0.287 0.696

−65.9 14.1*** 0.135 0.705

−60.9 29.3*** 0.295 0.723

−66.4 11.5 0.127 0.688

external training rem size imp empowerment modern practice export dummy fdi Constant Observations Log likelihood full X2 [7] Cragg–Uhler R2 Count R2

112

Authors, based on ENESTYC 2005, INEGI. Notes: Robust standard errors in parentheses. Degrees of freedom within squared brackets. For variables definitions see Tables 1 and 2. * p < 0.1. ** p < 0.05. *** p < 0.01.

Estimates indicate that the provision of training remains the most significant practice across learning activities. And yet, the more evident contribution corresponds to R&D for new drug designs, followed by improvements in drug manufacturing processes. Variables on remunerations and worker’s empowerment also gain statistical significance. Remunerations are important for exploratory R&D leading to new drug manufacturing processes, and new drug designs. Worker’s empowerment has positive effects on rd design meq. Export dummy and fdi continue to play relevant roles, with effects running in opposite directions, for rd drug design. Table 5 presents the computation of marginal effects for variables in the basic models of Table 4. Unlike the analysis in Table 3 this new excercise is much more detailed. Estimates confirm training as the intervention with the largest positive and statistically significant impact on the likelihood that a firm performs R&D. In fact, the largest effect of training is on rd drug design, 43.7 percent. By contrast, the lowest influence, some nine percent, is in the case of rd design meq. The latter variable is also the one where worker’s empowerment has perceptible and positive contributions to learning. Marginal increases in remunerations have positive and statistically significant influence on knowledge exploration. This result suggests that as firm grow, so does their capacity to remunerate R&D personnel. Interpretation of discrete probability changes should be handled with care, as they are meaningful only for variables spanning over a sufficiently large range of values (Long and Freese, 2006). A pertinent case is that of remunerations. Column (1) in Table 5 reveals that a change in the log of remunerations, equivalent to an increase from minimum to maximum, raises the likelihood that a firm

conducts rd drug design by some 0.341. Changes in remunerations are stronger for demand-enhancing R&D than for cost-reducing activities. Changes induced by increases of half a standard deviation in the log of remunerations, column (4), are larger for rd design than for any other type of process R&D. Except for rd design meq, worker’s participation in decision-making failed to provide meaningful information about its likely influence on the likelihood of R&D performance. 5.3. Effects from different types of training Similar to previous studies on innovation and human resource management, this paper has documented that the provision of training has positive and robust influences on the likelihood that firms do R&D. In order to extract some more meaningful conclusions, more disaggregated measures on the actual nature of training were introduced. Section 4.3 identified two complementary forms: internal (on-the-job) and external (off-the-job). The former was expected to support knowledge diffusion and sharing within the organization, it would more closely relate to knowledge exploitation strategies. External training, in turn, was expected to support expansion of knowledge bases through interactions with other knowledge producers (Casas, 2005). In order to explore this dual nature of training, two additional variables, namely, training internal and external training, were brought into the analysis. Table 6 contains estimates for models where train04 is replaced by the two new variables.6 The Wald

6 The analysis excluded the variable on rd design meq because training internal tended to predict perfectly the probability that a firm performs this specific activity.

542

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

tests show that, with the exception of rd drug improvement, the remaining models are statistically significant at conventional confidence levels.7 Estimates in Table 6 confirm that internal training is more closely related to knowledge exploitation, while that provided by external agents impacts more directly on learning through knowledge exploration, particularly rd drug design. Note that remunerations and worker’s empowerment lose the explanatory power found in Tables 3 and 4. Exposure to competition through participation in export markets stimulates learning, particularly for (new) pharmaceutical products. 6. Discussion and conclusions This paper investigated the influence of human resource management practices on the likelihood that a firm performs in-house R&D. Based on the relevant literature, the focus was on the provision of training, remunerations and worker’s participation in decision-making. The analysis was carried out in the context of a developing country such as Mexico, and was based on the notions of knowledge exploitation and knowledge exploration, respectively. Results show that human resource management influences innovation by stimulating, first, learning and capacity-building through in-house R&D. The number of relevant human resource management practices and their corresponding influence is contingent on the novelty of the knowledge required by the firm. The latter in turn, is linked to expected R&D outcomes. The more novel the expected R&D outcome, the more notorious influence of human resource managament practices tends to be. Differences in the nature of R&D induces distinctive demands on the human resources shaping an organization. In line with the literature on human resource management and new product development, R&D for new drug designs was found to be positively associated with practices such as training and remunerations (Lund, 2004a). Nevertheless evidence was also obtained on the positive impact of management practices on R&D for new process innovation, technical change more broadly defined. Training, remuneration for performance and incorporation of workers into decision-making supported R&D for the design or improvement of machinery and equipment. To the best of our knowledge, this is one of the first studies in this field documenting this issue in the context of developing countries. Data limitations prevented further investigation into this finding; nevertheless this is a relevant issue considering that process innovations enjoy a significant share of innovations in developing countries. As for the effects associated to specific personnel management interventions, and in light of the hypotheses presented in Section 3, relevant findings are as follows: The provision of training systematically and positively affects the likelihood that firms pursue R&D; hence

7 Although not included in Table 6, models were also ran where fdi expt replaced the individual variables fdi and export dummy. Results from such models are consistent with the conclusions presented here.

hypothesis H1 is confirmed. This lends support to Samstad and Pipkin (2005)’s claim that training and general qualifications of the labor force dictate the type of human resource management practices needed and feasible in countries such as Mexico. Raising skill levels facilitates adoption of advanced management systems in Mexican firms; moreover, it assists in building required capabilities to more sistematically conduct R&D. This paper supported the pertinence to promote interactions between firms and other external agents, at least for the provision of R&D-relevant training. Hence firms can access new knoweldge and expand existing knowledge bases. Further research should shed light on the nature of the actual knowledge flows being involved. However, exploratory interviews among Mexican drug manufactures suggest that interactions are broad. They include learning about new excipients and formulations, to methodologies for the synthesis of chemical ingredients. In yet some other cases, external training provides understanding of advanced research methodologies and applications, particularly in areas such as biotechnology. Overall, hypothesis H2 was also confirmed. The literature review in Section 3 suggested that, in principle, remunerations should influence learning positively. Estimates revealed that raising remunerations increases the probability that a firm performs in-house R&D, particularly for knowledge exploration. However, the effect was not robust. Consequently hypothesis H3 is partially supported; hypothesis H4 seems more plausible. Remunerations underpin learning but only under certain conditions and for specific types of R&D. Albeit difficult to corroborate based on data used here, a possible explanation results from the frequent mark-up on pecuniary remunerations, more specifically wages. In countries such as Mexico opportunities to training and/or perspectives for promotion become equally or even more relevant as reward mechanisms. Alternatively, remunerations serve more as determinants of labor mobility within the pharmaceutical industry; thus promoting a continuous transfer of research capabilities, however limited, within the industry (Santiago, 2010). Somewhat inconclusive results were drawn in the case of worker’s empowerment. The practice was positive and significant only in the case of exploration-related R&D for new drug manufacturing processes. This is at odds with previous literature indicating that delegation of decisionmaking capacity is key for new product development, as it fosters creativity and discovery (Mumford, 2000). A possible explanation points at the traditional perception that paternalistic work environments, rigid and hierarchical oganizational structures, such as those generaly found in Mexico and other similar countries, are unsuitable for enhanced performance. Nevertheless, as Section 3 in this paper acknowledged, the nature of drug manufacturing can limit the influence of worker’s empowerment on the scope of R&D. Concerns over drug quality and safety lead to close scrutiny and approval by sanitary authorities thereby limiting the capacity to change both existing drugs and the corresponding manufacturing processes. Any alteration in either of them may require additional reviews and approval by the regulatory authorities. FDA (2004) asserts that this

Table A1 Correlation analysis of variables on management practices and firm characteristics considered for the analyses. Max

Min

rd inhouse rd exploit

0.741 0.714

0 0

1 1

(3)

rd explore

0.625

0

1

(4)

rd improve process

0.634

0

1

(5)

rd design meq

0.188

0

1

(5)

rd drug design

0.616

0

1

(6)

rd drug improvement

0.661

0

1

(7)

ln avg rema

4.735

3.137

5.914

(8)

rem sizeb

6.846

3.137

11.828

(9)

train04

0.893

0

1

(10)

training internal

0.786

0

1

(11)

external training

0.750

0

1

(12)

internal external trc

2.286

0

3

(13)

imp empowerment d

0.491

0

2

(14)

modern practice

0.670

0

1

(15)

export dummy

0.536

0

1

(16)

fdi

0.313

0

1

(17)

fdi expt

0.295

0

1

(18)

firm sizee

1.411

1

2

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

1.00 0.93 (0.00) 0.76 (0.00) 0.78 (0.00) 0.28 (0.00) 0.75 (0.00) 0.82 (0.00)

0.69 (0.00) 0.83 (0.00) 0.30 (0.00) 0.68 (0.00) 0.88 (0.00)

0.64 (0.00) 0.37 (0.00) 0.98 (0.00) 0.73 (0.00)

0.37 (0.00) 0.62 (0.00) 0.71 (0.00)

0.33 (0.00) 0.34 (0.00)

1.00 0.71 (0.00)

1.00

0.22 (0.02) 0.11 (0.27) 0.32 (0.00) 0.24 (0.01) 0.27 (0.00) 0.31 (0.00)

0.17 (0.08) 0.10 (0.29) 0.29 (0.00) 0.25 (0.01) 0.23 (0.02) 0.28 (0.00)

0.22 (0.02) 0.17 (0.07) 0.27 (0.00) 0.18 (0.06) 0.32 (0.00) 0.33 (0.00)

0.13 (0.16) 0.08 (0.39) 0.28 (0.00) 0.24 (0.01) 0.25 (0.01) 0.29 (0.00)

0.12 (0.20) 0.20 (0.04) 0.17 (0.08) 0.25 (0.01) 0.17 (0.07) 0.24 (0.01)

0.23 (0.01) 0.18 (0.05) 0.26 (0.01) 0.17 (0.07) 0.31 (0.00) 0.32 (0.00)

0.22 (0.02) 0.15 (0.11) 0.24 (0.01) 0.18 (0.06) 0.20 (0.04) 0.23 (0.02)

1.00 0.71 (0.00) 0.14 (0.14) 0.22 (0.02) 0.21 (0.03) 0.26 (0.01)

1.00 0.13 (0.19) 0.27 (0.00) 0.27 (0.00) 0.33 (0.00)

1.00 0.66 (0.00) 0.60 (0.00) 0.74 (0.00)

1.00 0.30 (0.00) 0.63 (0.00)

1.00 0.93 (0.00)

1.00

0.01 (0.94) 0.10 (0.27) 0.14 (0.13) −0.04 (0.67) −0.07 (0.50) 0.08 (0.41)

−0.04 (0.70) 0.06 (0.53) 0.08 (0.37) −0.09 (0.37) −0.11 (0.24) 0.09 (0.37)

0.07 (0.47) 0.12 (0.20) 0.20 (0.03) −0.07 (0.43) −0.11 (0.27) 0.16 (0.09)

0.03 (0.75) 0.18 (0.06) 0.07 (0.44) −0.05 (0.62) −0.08 (0.41) 0.07 (0.47)

0.29 (0.00) 0.05 (0.63) 0.08 (0.40) −0.03 (0.77) −0.06 (0.53) 0.20 (0.03)

0.03 (0.76) 0.15 (0.12) 0.22 (0.02) −0.06 (0.52) −0.09 (0.33) 0.17 (0.07)

−0.01 (0.92) 0.06 (0.54) 0.09 (0.35) −0.05 (0.63) −0.07 (0.43) 0.14 (0.15)

0.10 (0.27) 0.14 (0.14) 0.63 (0.00) 0.51 (0.00) 0.52 (0.00) 0.50 (0.00)

0.12 (0.20) −0.01 (0.88) 0.41 (0.00) 0.48 (0.00) 0.48 (0.00) 0.96 (0.00)

0.16 (0.09) 0.25 (0.01) 0.14 (0.14) 0.17 (0.07) 0.16 (0.09) 0.11 (0.23)

0.09 (0.36) 0.14 (0.14) 0.26 (0.01) 0.26 (0.01) 0.24 (0.01) 0.26 (0.01)

0.11 (0.24) 0.12 (0.21) 0.12 (0.19) 0.08 (0.41) 0.06 (0.55) 0.27 (0.00)

0.12 (0.19) 0.15 (0.11) 0.20 (0.03) 0.16 (0.09) 0.14 (0.14) 0.32 (0.00)

(13)

(14)

(15)

(16)

(17)

(18)

1.00 1.00

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

Mean (1) (2)

1.00 1.00

1.00 0.44 (0.00) 0.09 (0.34) 0.19 (0.05) 0.16 (0.09) 0.12 (0.23)

1.00 0.15 (0.13) 0.19 (0.05) 0.16 (0.09) −0.07 (0.47)

1.00 0.55 (0.00) 0.60 (0.00) 0.27 (0.00)

1.00 0.96 (0.00) 0.38 (0.00)

1.00 0.38 (0.00)

1.00

Authors based on ENESTYC, 2005; INEGI. Notes: standard deviations: a 0.675, b 2.997, c 1.069, d 0.697, e 0.494. For variables definitions see Tables 1 and 2. p values in parenthesis.

543

544

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

can be cumbersome for the firm; a major barrier for process innovation in the pharmaceutical industry. In the context of countries specialized in the manufacturing of generic drugs, development of such products is restricted by the need to comply with specific parameters and qualities set by the drug innovator. If firms only reproduce the knowledge behind such products, it makes little sense to allow workers to play around with the technology. Hypothesis H5 is supported but H6 requires further scrutiny.

Some comments in relation to estimates associated with the control variables are pertinent. In relation to foreign ownership, the findings here contradict the usual perception that foreign firms are more technologically dynamic than domestic firms. The choice of performance indicators is important. In terms of R&D, a careful reflection points to the position that countries such as Mexico occupy within business and innovation strategies of multinationals. Local affiliates maintain low profiles when assisting in

Table B1 Summary statistics for the pharmaceutical industry in Mexico, 2004. Mean c

d

Standar deviation

Min

Max

R&D in-house Employment Total salesa Domestic sales Export share Share of FDI Ageb

Internal (I) 475.7 694094.8 609320.3 0.07 0.30 33.2

No R&D (II) 331.2 433261.5 394477.4 0.08 0.34 27.5

(I)/(II) 1.4 1.6 1.5 0.9 0.9 1.2

Internal 555.2 1270892 1055332 0.13 0.46 19.4

No R&D 259.1 694938.1 634741.2 0.20 0.48 16.6

Internal 1.1 2394 2394 0 0 1

No R&D 63 12127.5 0 0 0 0

Internal 3391.5 6958020 6334508 0.69 1 74

No R&D 1158.4 2297038 2069799 1 1 70

Improved process Employment Total salesa Domestic sales Export share Share of FDI Ageb

Imp proce 492.5 741488.3 656732.5 0.1 0.3 33.2

No R&Df 344.3 427531.3 375254.1 0.1 0.3 29.2

1.4 1.7 1.8 0.6 0.9 1.1

Imp proc 589.2 1354405.0 1120739.0 0.1 0.5 20.6

No R&D 261.3 641430.5 583423.9 0.2 0.5 15.1

Imp proc 1.1 2394.0 2394.0 0.0 0.0 1.0

No R&D 63.0 12127.5 0.0 0.0 0.0 0.0

Imp proc 3391.5 6958020.0 6334508.0 0.6 1.0 74.0

No R&D 1158.4 2297038.0 2069799.0 1.0 1.0 70.0

New process Employment Total salesa Domestic sales Export share Share of FDI Ageb

Mach & equipg 655.0 1140099.0 919528.0 0.1 0.3 39.2

No R&Dh 388.3 508048.1 469267.5 0.1 0.3 30.0

1.7 2.2 2.0 1.4 0.9 1.3

Mach & equip 804.2 1808071.0 1307236.0 0.2 0.5 17.7

No R&D 386.7 914265.9 856102.8 0.1 0.5 18.7

Mach & equip 2.2 31859.5 31859.5 0.0 0.0 16.0

No R&D 1.1 2394.0 0.0 0.0 0.0 0.0

Mach & equip 3391.5 6958020.0 4359928.0 0.6 1.0 74.0

No R&D 2852.9 6772189.0 6334508.0 1.0 1.0 72.0

Improved drug Employment Total salesa Domestic sales Export share Share of FDI Ageb

Imp drugi 496.6 738053.8 654131.2 0.1 0.3 34.2

No R&Dj 324.7 409433.5 358097.9 0.1 0.3 26.8

1.5 1.8 1.8 0.6 0.9 1.3

Imp drug 577.4 1328800.0 1101783.0 0.1 0.5 19.8

No R&D 261.5 653886.7 587750.3 0.2 0.5 15.9

Imp drug 1.1 2394.0 2394.0 0.0 0.0 1.0

No R&D 63.0 7717.9 0.0 0.0 0.0 0.0

IImp drug 3391.5 6958020.0 6334508.0 1.0 1.0 74.0

No R&D 1158.4 2297038.0 2069799.0 1.0 1.0 70.0

New drug Employment Total salesa Domestic sales Export share Share of FDI Ageb

drug designk 526.1 765674.0 676530.3 0.1 0.3 34.6

No R&Dl 297.4 403324.4 356577.6 0.1 0.3 27.1

1.8 1.9 1.9 0.7 0.8 1.3

drug design 592.1 1367771.0 1134408.0 0.1 0.5 19.9

No R&D 238.0 631241.8 564394.8 0.2 0.5 16.2

drug design 2.2 2394.0 2394.0 0.0 0.0 1.0

No R&D 1.1 7717.9 0.0 0.0 0.0 0.0

drug design 3391.5 6958020.0 6334508.0 0.7 1.0 74.0

No R&D 1158.4 2297038.0 2069799.0 1.0 1.0 70.0

Exploitation Employment Total salesa Domestic sales Export share Share of FDI Ageb

rd exploit (I)m 475.9 708400 626251.9 0.1 0.3 33.3

No R&D (II)n 344.3 421951.7 372290 0.1 0.4 27.7

1.4 1.7 1.7 1 0.7 1.2

rd exploit 560.7 1291759 1071360 0.1 0.4 19.7

No R&D 276.6 665506.5 607670.3 0.2 0.5 16.1

rd exploit 1.12 2394 2394 0 0 1

No R&D 63 12127.5 0 0 0 0

rd exploit 3391.5 6958020 6334508 0.6 1 74

No R&D 1158.4 2297038 2069799 1 1 70

Exploration Employment Total salesa Domestic sales Export share Share of FDI Ageb

rd exploreo 488.1 705485.3 620633.8 0.1 0.3 33.6

No R&Dp 319.0 437609.6 393435.2 0.1 0.3 27.2

1.5 1.6 1.6 1 1 1.2

rd explore 565.6 1292822 1073893 0.1 0.5 19.4

No R&D 249.1 693483.2 623592.3 0.2 0.5 16.7

rd explore 1.12 2394 2394 0 0 1

No R&D 63 7717.9 0 0 0 0

rd explore 3391.5 6958020 6334508 0.7 1 74

No R&D 1158.4 2297038 2069799 1 1 70

Authors, based on ENESTYC 2005, INEGI. Firms in sample: 112. Number of firms: c (83); d (29); e (71); f (41); g (21); h (91); i (74); j (38); k (69); l (43); m (80); n (32); o (79); p (33). a Thousand Mexican pesos. b Difference between the year in which a firm started operations in current business and the year of the survey, 2004.

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

the exploitation of knowledge generated at the parent location or elsewhere in the developed world (von Zedtwitz and Ove, 2002). Acquisition of new knowledge, demanding more advanced R&D activities, is seldom carried out in developing countries. By contrast, exposure to external competition and larger market opportunities, via exports, was found to increase the likelihood that a firm pursues R&D. The strongest effect is associated with new drug designs. In line with Kale and Little’s (2007), the managing director of a foreign affiliate argued that “Success requires strong commitment of financial and human resources, particularly in research. The goal is to develop a portfolio of products to be launched in export markets over a significant time horizon”. In the case of the Mexican industry, strong reliance on the local pharmaceutical market inhibit incentives to innovate (Santiago, 2010); management strategies aim to increase productivity and efficiency. In other words, adoption of modern organizational practices may simply contribute to the making of what Cimoli (2000) identifies as “global modern manufacturing centre”. Acknowledgements Earlier versions of this paper benefited greatly from comments by two anonymous referees and the editors to this special issue of SCED; as well as Gabriela Dutrénit; Nobuya Haraguchi; Robin Cowan; Wilfred Dolfsma, Branka Urem, Jojo Jacob and other members of the research group on Innovation, global business strategies and host country development at UNU-MERIT; Leonel Corona, Javier Jasso and staff of the División de Investigación, Facultad de Contaduría, Administración e Informática of the National Autonomous University, Mexico. Suggestions by Martin Shrolec and other participants at the 7th Annual conference of the Globelics network in Dakar, Senegal are appreciated. Fernando Santiago is greatly indebted to the following at INEGI for granting access to the data used in this paper: Gerardo Leyva and Abigail Durán. Special thanks to Adriana Ramírez, Gabriel Romero and Cándido Aguilar also at INEGI for help in the processing of the data. Maria Fermie helped in editing and correcting previous versions. Substantial part of the work was carried out while Fernando Santiago was a visiting researcher at the Research and Statistics Branch, UNIDO, and a PhD researcher at UNU-MERIT. Omissions and errors remain the responsibility of the authors. The ideas expressed in this paper do not reflect those of the organisations hosting the authors. Appendix A. See Table A1. Appendix B. See Table B1. References Amabile, T., 1997. Motivating creativity in organizations: on doing what you love and loving what you do. California Management Review 40, 39–58.

545

Arundel, A., Lorenz, E., Lundvall, B.-A., Valeyre, A., 2007. How Europe’s economies learn: a comparison of work organization and innovation mode for the EU-15. Industrial and Corporate Change 16, 1175–1210. Badawy, M., 1988. What we’ve learned: managing human resources. Research Technology Management 31, 139–145. Barney, J., 1991. Firm resources and sustained competitive advantage. Journal of Management 17, 99–120. Bartlett, K., Lawler, J., Bae, J., Chen, S., Wan, D., 2002. Differences in international human resource development among indigenous firms and multinational affiliates in East and Southeast Asia. Human Resource Development Quarterly 13 (4), 383–405. Barton, H., Delbridge, R., 2001. Development in the learning factory: training human capital. Journal of European Industrial Training 25, 465–472. Boseli, P., Dietz, G., Boon, C., 2005. Commonalities and contradictions in HRM and performance research. Human Resource Management Journal 15, 67–94. Cardinal, L., Hatfield, D., 2000. Internal knowledge generation: the research laboratory and innovative productivity in the pharmaceutical industry. Journal of Engineering and Technology Management 17, 247–271. Casas, R., 2005. Exchange and knowledge flows between large firms and research institutions. Innovation: Management, Policy and Practice 7, 188–199. Caves, R., Hurwitz, M., Whinston, M., 1991. Patent expiration, entry and competition in the US pharmaceutical industry. In: Brookings Papers in Economic Activity, pp. 1–48. Christofides, L., Hardin, Stengos, T., 2000. Corrigendum to “On the calculation of marginal effects in the bivariate probit model”. Economics Letters 68, 339. Christofides, L., Stengos, T., Swidinsky, R., 1997. On the calculation of marginal effects in the bivariate probit model. Economics Letters 54, 203–208. Chung-Jen, C., Jing-Wen, H., 2009. Strategic human resource practices and innovation performance – the mediating role of knowledge management capacity. Journal of Business Research 62, 104–114. Cimoli, M. (Ed.), 2000. Developing Innovation Systems: Mexico in a global context. Thomson Learning, London and New York. Cockburn, I., Henderson, R., 2001. Scale and scope in drug development: unpacking the advantages of size in pharmaceutical research. Journal of Health Economics 20 (6), 1033–1057. Cohen, W., Levinthal, D., 1989. Innovation and learning: the two faces of R and D. The Economic Journal 99, 569–596. Cohen, W., Levinthal, D., 1990. Absorptive capacity: a new perspective on learning and innovation. Administrative Science Quarterly 35, 128–152. Combs, J., Liu, Y., Hall, A., Ketchen, D., 2006. How much do highperformance work practices matter? A meta-analysis of their effects on organizational performance. Personnel Psychology 59, 501–528. Delery, J., 1998. Issues of fit in strategic human resource management: implications for research. Human Resource Management Review 8 (3), 289–309. Gilsing, V., 2006. Exploration and exploitation in innovation systems: the case of pharmaceutical biotechnology. Research Policy 35, 1–23. Grant, R., 1996. Prospering in dynamically-competitive environments: organizational capability as knowledge integration. Organization Science 7, 375–387. Gray, D., Cundell, S., Hay, D., O’Neill, J., 2004. Learning Through the Workplace: A Guide to Work-Based Learning. Nelson Thrones, Ltd., Spain. Greene, W., 2003. Econometric Analysis. Pearson Education. Greve, H., 2007. Exploration and exploitation in product innovation. Industrial and Corporate Change 16, 945–975. Guzmán, A., 2005. Naturaleza de la IyD y las patentes de la industria farmacéutica en México, in Industria Farmacéutica y Propiedad Intelectual: Los países en desarrollo. In: Guzmán, A., Viniegra, G. (Eds.), H. Cámara de Diputados LIX Legislatura, UAM-Iztapalapa, M.A. Porrúa. México, pp. 267–322. Hara, T., 2003. Innovation in the Pharmaceutical Industry: The Process of Drug Discovery and Development. Edward Elgar. Hemmert, M., 1998. Reorganization of R and D in Japanese manufacturing firms: preserving competitiveness for the twenty-first century. In: Hemmert, M., Oberländer, C. (Eds.), Technology and Innovation in Japan: Policy and Management for the Twenty-First Century. Routledge, pp. 129–150 (Chapter 8). Hollis, A., 2002. The importance of being first: evidence from Canadian generic pharmaceuticals. Health Economics 11, 723–734. Johnson, J., Baldwin, J., Diverty, B., 1996. The implications of innovation for human resources strategy. Futures 28, 103–119.

546

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530–546

Kale, D., Little, L., 2007. From imitation to innovation: the evolution of R&D capabilities and learning processes in the Indian pharmaceutical industry. Technology Analysis & Strategic Management 19, 589–609. Kaplinsky, R., 1995. Technique and system: the spread of Japanese management techniques to developing countries. World Development 23, 57–71. Kettler, H., Modi, R., 2001. Building local research and development capacity for the prevention and cure of neglected diseases: the case of India. Bulletin of the World Health Organization 79, 742–747. Kim, L., 1997. Imitation to Innovation: The Dynamics of Korea’s Technological Learning. Harvard University Press, Boston. Kim, Y., Kim, L.S., Lee, J., 1989. Innovation strategy of local pharmaceutical firms in Korea: a multivariate analysis. Technology Analysis and Strategic Management 1 (1), 29–44. Labarca, G., 1999. Formación para el Trabajo: Observaciones en América Latina y El Caribe. In: Guillermo Labarca (Ed.), Formación y empresa, CINTERFOR/OIT/CEPAL/GTZ, pp. 19–58 (Chapter 1). Laursen, K., Mahnke, V., 2001. Knowledge strategies, firm types, and complementarity in human-resource practices. Journal of Management and Governance 5, 1–27. Laursen, K., 2002. The importance of sectoral differences in the application of complementary HRM practices for innovation performance. International Journal of the Economics of Business 9, 139–156. Laursen, K., Foss, N., 2003. New human resource management practices, complementarities and the impact on innovation performance. Cambridge Journal of Economics 27, 243–263. Li, Y., Vanhaverbeke, W., Schoemakers, W., 2008. Exploration and exploitation in innovation: reframing the interpretation. Creativity and Innovation Management 17, 107–126. Long, S., Freese, J., 2006. Regression Models for Categorical Dependent Variables Using STATA. STATA Press. Lorenz, E., Valeyre, A., 2006. Organizational forms and innovative performance: a comparison of the EU-15. In: Lorenz, E., Lundvall, B.-A. (Eds.), How Europe’s Economies Learn: Coordinating Competing Models. Oxford University Press, pp. 140–160 (Chapter 6). Lorenz, E., Wilkinson, F., 2003. Forum: Organizational change, human resource management and innovative performance: comparative perspectives. Cambridge Journal of Economics 27, 239–241. Lund, R., 2004a. Some Danish experiences related to the organization of new product development. In: Christensen, J., Lundvall, B.-A. (Eds.), Research on Technological Innovation, Management and Policy, 8. Elsevier, Amsterdam, pp. 229–254. Lund, R., 2004b. The organization of actors’ learning in connection with new product development. In: Christensen, J., Lundvall, B.-A. (Eds.), Research on Technological Innovation, Management and Policy, 8. Elsevier, Amsterdam, pp. 129–153. Lundvall, B.-A., Johnson, B., Andersen, E., Dalum, B., 2002. National systems of production, innovation and competence building. Research Policy 31, 213–231. March, J., 1991. Exploration and exploitation in organizational learning. Organization Science 2, 71–87. Michie, J., Sheehan, M., 1999. HRM practices, R&D expenditure and innovation investment: evidence from the UK’s 1990 Workplace Industrial Relations Survey (WIRS). Industrial and Corporate Change 8, 211–234. Michie, J., Sheehan, M., 2003. Labour market deregulation, ‘flexibility’ and innovation. Cambridge Journal of Economics 27, 123–143. Mumford, M., 2000. Managing creative people: strategies and tactics for innovation. Human Resource Management Review 10, 313–351.

Nelson, R., Winter, S., 1982. An Evolutionary Theory of Technical Change. Beknap Harvard, Cambridge, MA. Odagiri, H., 1998. The Japanese business system for creation and diffusion of technological knowledge: time for a change? In: Hemmert, M., Oberländer, C. (Eds.), Technology and Innovation in Japan: Policy and Management for the Twenty-First Century. Routledge, pp. 115–128 (Chapter 7). OECD, 1998. The OECD Jobs Strategy: Technology, Productivity and Job Creation: Best Policy Practices. OECD, Paris. Quinn, M., Rubb, S., 2006. Mexico’s labor market: the importance of education–occupation matching on wages and productivity in developing countries. Economics of Education 25, 147–156. Rothaermel, F., Deeds, D., 2004. Exploration and exploitation alliances in biotechnology: a system of new product development. Strategic Management Journal 25, 201–222. Samstad, J., Pipkin, S., 2005. Bringing the firm back in: Local decision making and human capital development in Mexico’s maquiladora sector. World Development 33, 805–822. Santiago, F., 2010. Human resource management practices and learning for innovation in developing countries: pharmaceutical firms in Mexico. Doctoral Dissertation, UNU-MERIT/University of Maastricht, The Netherlands. Seiter, A., 2005. Pharmaceuticals: local manufacturing. Health Nutrition and Population, vol. 5, The World Bank. Sternberg, R., O’Hara, L., Lubbart, T., 1997. Creativity as investment. California Management Review 40, 8–21. Tello, M.P., Greene, W., 1996. US managerial strategies and applications for retaining personnel in Mexico. International Journal of Man Power 17, 54–94. Terziovski, M., Morgan, J., 2006. Management practices and strategies to accelerate the innovation cycle in the biotechnology industry. Technovation 26, 1–8. Vargas, F., 2004. 40 preguntas sobre competencia laboral. Montevideo, Uruguay: CINTERFOR/OIT/CEPAL/GTZ. von Zedtwitz, M., Ove, G., 2002. Market versus technology drive in R and D internationalization: four different patterns of managing research and development. Research Policy 31, 569–588. White, F., 2002. Capacity-building for health research in developing countries: a manager’s approach. Pan American Journal of Public Health 12, 165–172. Wright, P., Dunford, B., Scott, S., 2001. Human resources and the resource based view of the firm. Journal of Management 27, 701–721. ˜ M., Guzmán, A., Brown, F., 2007. Technology acquisition strategies Zúniga, in the pharmaceutical industry in Mexico. Comparative Technology Transfer and Society 5, 274–296.

Websites Bureau of Labor Statistics, US Department of Labor; Career Guide to Industries, 2008–09 Edition, Pharmaceutical and Medicine Manufacturing; http://www.bls.gov/oco/cg/cgs009.htm; 26 January, 2008. FDA, 2004. Innovation and Continuous Improvement in Pharmaceutical Manufacturing: Pharmaceutical CGMPs for the 21st Century. US Food and Drug Administration; http://www.fda.gov/Drugs/ DevelopmentApprovalProcess/Manufacturing/Questionsand AnswersonCurrentGoodManufacturingPracticescGMPforDrugs/ default.htm; 8 January, 2009. GPWI; Las mejores empresas para trabajar en México; http://www. greatplacetowork.com.mx/best/list-mx.htm; 8 January, 2009.