human resource management for learning through knowledge exploitation and knowledge exploration:...

He

Fa

b

a

ARRAA

JO

KRLHPM

1

mgfieiepcma

8

(

0d

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

Contents lists available at SciVerse ScienceDirect

Structural Change and Economic Dynamics

journa l h o me pag e: www.elsev ier .com/ locate /sced

uman resource management for learning through knowledgexploitation and knowledge exploration: Pharmaceuticals in Mexico

ernando Santiagoa,∗, Ludovico Alcortab

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

r t i c l e i n f o

rticle history:eceived January 2010eceived in revised form June 2011ccepted November 2011vailable online 8 December 2011

EL classification:32, O54, L65

a b s t r a c t

This paper investigates the influence of human resource management practices on thelikelihood that a firm performs in-house R&D. R&D is broadly interpreted as learning—amechanism 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 knowledgerequirements associate with distinct R&D outcomes with varying degrees of novelty for thefirm. Findings are supported with evidence from the pharmaceutical industry in Mexico.

eywords:&Dearning and innovationuman resource managementharmaceuticalsexico

The analysis reveals positive linkages between human resource management practices andlearning at firm level. The relationship is contingent on factors such as expected R&D out-comes, or the novelty of the knowledge required by the firm. The provision of trainingrevealed the more consistent, positive influence on the likelihood that pharmaceuticalsfirms perform R&D in Mexico.

© 2011 Elsevier B.V. All rights reserved.

. Introduction

Literature on the linkages between human resourceanagement and innovation performance at firm level is

rowing. Empirical work stems mostly from surveys ofrms in developed countries. Scholars have addressed thextent to which sets of new and dynamic work practicesnfluence innovation (Barton and Delbridge, 2001); theffects of distinct forms of labor flexibility on innovationerformance (Michie and Sheehan, 1999, 2003), and the
omplementary relationships between human resourceanagement practices underpinning innovation (Laursen
nd Foss, 2003). Research on the organization and learning

∗ Corresponding author. Present address: IDRC, 150 Kent Street, PO Box500, Ottawa, Canada K1G 3H9. Tel.: +1 613 696 2269.

E-mail addresses: [email protected] (F. Santiago), [email protected]. Alcorta).

954-349X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.strueco.2011.11.002

of agents involved in new product development is likewisesignificant (Lund, 2004a,b).

Available literature documents positive relationshipsbetween human resource management and innovationperformance at firm level. The influence of such practicesvaries according to the technological dynamics of differentindustries (Laursen, 2002; Laursen and Foss, 2003), estab-lishment sizes and occupations (Lorenz and Valeyre, 2006),or the way national environments determine learning atindividual and organizational levels (Arundel et al., 2007).Still missing, however, is a better understanding of mech-anisms to explain such relationships (Laursen and Foss,2003; Chung-Jen and Jing-Wen, 2009), and a consistenttheory on what Delery (1998) termed the “transmissionmechanism” from human resource management to inno-
vation performance.
Explaining how and why human resource manage-ment underpins innovation introduces innovation scholarsinto the more ample debate about how and why such

dx.doi.org/10.1016/j.strueco.2011.11.002

http://www.sciencedirect.com/science/journal/0954349X

http://www.elsevier.com/locate/sced

mailto:[email protected]

mailto:[email protected]

dx.doi.org/10.1016/j.strueco.2011.11.002

ge and E
F. Santiago, L. Alcorta / Structural Chan
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 num-ber and possible definitions of indicators on humanresource management practices, together with the dis-tinct multidisciplinary approaches to research. Arguablyresearch in the field needs to be fine-tuned, specificallyin the way the issues at stake are addressed. Lorenzand Wilkinson (2003) assert that researchers frequentlyassume linear relationships—from adoption of specific setsof management practices to innovation; leaving little roomfor more heterogeneous organizational strategies withinsingle industries. It is also customary to look at innova-tion outcomes—products/processes; and their degrees ofnovelty—radical/incremental. Somewhat understimated isthe study of the latent processes associated with the orga-nization of people involved in innovation.

Methodologicaly consideration of the intermediatelatent processes linking human resource management toa firm’s performance is familiar for management schol-ars. Sternberg et al. (1997), Amabile (1997) and Mumford(2000), for instance, document how human resource man-agement practices affect creativity and creative thinking.Relatedly Cohen and Levinthal (1989, 1990), Wright et al.(2001) and Chung-Jen and Jing-Wen (2009) assert thathuman resource management helps to capture and mobi-lize knowledge residing within and outside organizations.

From the above, this paper enquiries about the inter-mediary factors that link human resource managementto innovation. In particular it looks at learning processessupporting absorptive capacity, and the development ofinnovation capabilities by individuals and, ultimately,organizations. Learning arises from systematic perfor-mance of R&D by the firm. In such a way the paper grantsresearch on human resource management practices andinnovation greater relevance from a development perspec-tive. White (2002) stressed the pertinence to understandhow such practices contribute to research and other tech-nological capabilities, particularly in developing countries.In his view, accumulated capacities can erode because ofinadequate or poor management of people.

To the best of our knowledge, this paper stems fromone of the first systematic studies on the influence ofhuman resource management over learning through R&Din developing countries. Based on literature on knowledgeexploitation and knowledge exploration, the hypothesisis that the contribution of human resouces managementto learning depends on factors such as the novelty of theknowledge required, and the expected outcomes from in-house R&D.

Empirical evidence refers to pharmaceutical firms inMexico. In addition to being one of the most advanceddeveloping economies, the country is the world’s ninthpharmaceutical market and the second in Latin America.As such, it has strong, although poorly realized poten-tial to contribute to pharmaceutical innovation. Lackof sufficiently experienced and well trained workforce
remains major bottleneck (Guzmán, 2005). Focus on thepharmaceutical industry in Mexico also helps to illus-trate the importance of carefully considering the contextsin which human resource management practices work.
conomic Dynamics 23 (2012) 530– 546 531

Macroeconomic conditions, the social environment aroundR&D, or even how countries contribute to innovation inspecific industries dictate not only what is possible andfeasible, but what can be expected from human resourcemanagement. Better understanding of the organizationalpractices around pharmaceutical R&D can inform strate-gies to support the development of human resources forthe industry in Mexico and similar countries.

The paper proceeds as follows: Section 2 brings togetherliterature on human resource management and learning;the case of pharmaceuticals R&D in developing coun-tries illustrates the discussion. Section 3 characterizes thespecific management practices included for the analy-sis: training, remuneration, and worker’s participation indecision making; these practices are expected to enhanceindividuals’ and thereby, organizational learning. Section 4presents the data, defines variables and the correspondingresearch strategy. Empirical results are provided in Section5. Finally, Section 6 contains the discussion and conclu-sions.

2. Human resource management and learningthrough R&D

This paper equates learning with absorptive capacityand capability-building processes by the firm. The lit-erature documents the contribution that organizationalpractices, relating to R&D and innovation, can make towardthe succes of firms. Such practices assist in continuousefforts to mobilize and organize resources that firms haveat hand. In the case of Japan, for example, Odagiri (1998)highlighted the importance of building absorptive capabil-ities, making efforts in training and entrepreneurship andgaining sound scientific and technological understanding;including mastering the production and management ofskilled personnel.

Hemmert (1998) further underscored human resourcemanagement strategies to explain how Japanese firms havedealt with changing, often adverse, macroeconomic envi-ronments, and the challenges associated with businessstrategies posed by continuous technological innovation.Firms constantly reorganize and restructure R&D activitiesin general, and the management of R&D personnel in partic-ular. Continuous improvement in personnel managementunderpins innovative organizational practices to promoteincentives and motivation for, and productivity in R&D.Accordingly, Lundvall et al. (2002) argued that in additionto R&D efforts, analyses of firms’ innovation capabilitiesneed to consider the influence emanating from the dailyexperiences of workers, engineers and salesmen, togetherwith interactions among individuals within and outside theboundaries of a firm.

Cohen and Levinthal (1989 and 1990)’s treatment ofthe dual role of R&D as learning mechanism links humanresource management to R&D. R&D generates new infor-mation and knowledge underpinning searches for newmarket and technological opportunities through innova-
tion. R&D is equally relevant for assimilating and exploitingexisting information and knowledge. In other words, ithelps to build absorptive capacity by tapping existingknowledge. The authors further distinguished between

5 ge and E

ekasctbcc

ettkpc(tez

tkrb(epapwmfiiaettfiw

oeEkpbnotir

rtabdelw

32 F. Santiago, L. Alcorta / Structural Chan

xpected goals from R&D. Firms can exploit existingnowledge bases, or engage in knowledge explorationnd expansion of knowledge bases. Cohen and Levinthaltressed that the contribution of individuals’ cognitive pro-esses to accumulate absorptive capacity is contingent onhe nature of prior related knowledge and diversity ofackgrounds. These elements depend on an individual’sapacity to absorb, assimilate, link, analyse and, eventually,reate knowledge.

Management scholars integrate the notions of knowl-dge exploitation and knowledge exploration, as cen-ral and distinguishable elements shaping organiza-ional learning and capability-building, in the so-callednowledge-based theory of the firm (March, 1991). Therimary role of firms, which is the basis of organizationalapabilities, is the integration of specialized knowledgeGrant, 1996). The latter in turn, is often perceived inacit form, and know-how, skills and practical knowledgembedded in individuals as core components of an organi-ation (Barney, 1991).

Succesful innovative firms couple local searches,hrough internal learning efforts, with distanct searches,nowledge diffusion and assimilation through, for instance,everse engineering. Firms need to strategically com-ine stocks and flows of knowledge. Nelson and Winter1982) argue that, over time, firms gain experience and,ventually, develop routines that increase efficiency androductivity in manufacturing and, in general, the man-gement of current product portfolios. Improvements inroducts, processes or both are generally based on searchesithin a firm’s accumulated knowledge. Conversely, theore alien the intended innovation relative to what the

rm knows, the larger the need to look beyond famil-ar cognitive boundaries. Management systems influencend play mediatory roles in these processes; they influ-nce the organization and mobilization of individuals andheir knowledge (Barney, 1991). They assist in the creation,ransfer and integration of knowledge flows that enrich arms’ human capital, as a stock (Wright et al., 2001), inays that are valuable, rare and inimitable (Grant, 1996).

That firms engage in either knowledge exploitationr knowledge exploration, or both, illustrates the het-rogeneity, complexity and distinct uses of knowledge.xploitation refers to the use and refinement of existingnowledge, technologies and products. It entails short-runerspectives, more certainty and proximity to potentialenefits. Exploration, for its part, identifies searches forew knowledge, use of unfamiliar technologies, creationf products/services with unforeseen, or, at least, difficulto predict, demand (March, 1991; Greve, 2007). Explorationmplies long-run mindset, greater uncertainty about futureevenues and benefits.

Although exploration and exploitation have potentiallyeinforcing effects on learning and capability-building,hey lead to competing resource allocation, increased risksnd tradeoffs in investment decisions. Finding the rightalance is problematic, the choice of either strategy con-
itions survival and prosperity of firms: “. . .Systems thatngage in exploration to the exclusion of exploitation areikely to find that they suffer the costs of experimentation
ithout gaining many of its benefits. They exhibit too many

conomic Dynamics 23 (2012) 530– 546

undeveloped new ideas and too little distinctive compe-tence. Conversely, systems that engage in exploitation tothe exclusion of exploration are likely to find themselvestrapped in suboptimal stable equilibria” (March, 1991, p.71).

From the above, and based on Li et al. (2008), a practicalinterpretation of knowledge exploration and knowledgeexploitation is in terms of the cognitive distance betweenknowledge requirements and a firm’s knowledge base. Thelatter is characterised by Kale and Little (2007, p. 594) “assimple and complex, based on the technological challengesinvolved in developing particular products and underlay-ing capabilities”. Knowledge exploitation refers to localsearches for familiar, mature, current or proximate knowl-edge; it builds on existing technological capabilities. Bycontrast, knowledge exploration underpins searches forunfamiliar, distant knowledge. The proposed interpreta-tion of innovation draws attention to the learning processoccuring inside the firm. It induces some flexibility to theanalysis while still capturing traditional views of innova-tion in terms of incremental and radical outcomes (Greve,2007). Whereas local searches may lead to incrementalinnovations, distant searches could lead to radical ones.Nevertheless there is no reason for such match betweenknowledge searches and innovation outcomes to alwaysoccur.

2.1. The case of pharmaceuticals

The pharmaceutical industry is illustrative of the issuesdiscused above. Pharmaceuticals are highly R&D intensive;the capacity to perform R&D determines the firm’s via-bility and capacity to grow. As learning mechanism R&Dintertwines capacities to exploit and explore technologicaland market opportunities. At basic level of technlogi-cal capabilities, R&D supports accumulation of knowledgeand experience needed to progressively introduce moresophisticated drugs to the market. Recent experiences inIndia support this argument. For instance, based on acapability building model, Kale and Little (2007) arguedthat “reverse engineering R&D capability—the ability todevelop products by copying the process-is categorised asbasic capability. Generics R&D involves incremental changerepresenting intermediate capability while new chemicalentity research involves creating new drugs and inno-vative therapies representing advanced capabilities” (p.594). Building on the experience of Indian pharmaceuticalfirms, the authors illustrated how each stage of capabil-ity accumulation makes different demands from a firm’sknowledge base. Over time, firms use, acquire and accu-mulate different types of knowledge inputs for innovationwith increasing degrees of novelty. Progression in the tech-nology ladder accompanies expansion of learning activitiesoutside familiar cognitive boundaries; knowledge searchesbecome increasingly exploratory. Knowledge exploitation,however, remains relevant particularly for firms whosebusiness 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 thatalthough the technological dynamism of firms in catching

ge and E
up modes generally lags behind that of large multination-als, R&D remains core ingredient for success. The majordifference is that, in most cases, R&D in developing coun-tries leads to incremental innovations. This is the case ofmost R&D by pharmaceutical firms in Mexico.

Learning through R&D is relevant even for develop-ing countries specialized in development and marketingof generics drugs. Development of generics starts a fewyears before patent expiry of the innovator product. Firmshave to reproduce the knowledge needed to manufactureit while ensuring bioequivalence and biodisponibility, thussupporting its characteristic as generic interchangeabledrug.2 Speed is necessary to the extent that first moverscan gain and retain relevant market shares (Caves et al.,1991; Hollis, 2002). In most cases, the choice of prod-ucts is linked to current product portfolios; what firmsalready know. Nevertheless, expected benefits increaseif firms are able to enhance the characteristics of theinnovator drug. Quality enhancement includes relativelysimple improvements in product packaging, reformulationor recombinition of existing molecules. New products, inturn, include new applications of existing drugs, often indifferent therapeutic areas. The search for new knowledgemay relate more to the methods and techniques used tosynthesize the components—biotechnology techniques, forinstance—than to the characteristics of the drug itself (Kaleand Little, 2007).

3. Management practices and learning through R&D

Section 1 commented on the complexities to define,based on widely accepted theoretical rationale, compre-hensive checklists of management practices determiningperformance at firm level. The literature shows never-theless that enhanced organizational practices frequentlyrelate to Japanese management styles. Hemmert (1998)for example, indicated that relevant practices targetingR&D personnel include: hiring and firing, job rota-tion and continuity and compensation systems. Researchon complementarities among human resource manage-ment practices identifies sets of interventions including:indicators on labor relations—incentives and compensa-tion, recruitment and selection, teamwork, employmentsecurity, flexibility in job assignments, training, labor-management communication and grievance rates (Michieand Sheehan, 1999, 2003; Laursen and Foss, 2003). Indeveloping country contexts Tello and Greene (1996) andVargas (2004) coincide on such indicators as the provi-sion of training, worker’ empowerment, compensation andstaff promotion as part of organizational techniques, suchas total quality management (TQM) or just-in-time (JIT).Chung-Jen and Jing-Wen (2009) identified similar prac-tices as mediators in knowledge management processes by
Chinese firms.
The following paragraphs present some humanresource management practices likely to influence

2 Generic interchangeable (GI) denomination indicates that the reactionto a generic drug in the human body is exactly the same as that of aninnovator drug.


learning within organizations. The discussion informssome hypotheses to be explored in Section 5.

The provision of training: Training underpins develop-ment of technical and managerial skills among people,who are repositories of the tacit knowledge of an orga-nization (Johnson et al., 1996). Tacit knowledge supportsorganizational structures, as well as the productive andinnovation capabilities of a firm. Training takes two com-plementary forms: on-the-job and off-the-job. The formeris most common. It supports learning of day-to-day opera-tions and an understanding of basic concepts. The second,usually available for key personnel, contributes to enhanc-ing the intellectual capital and skills by capturing existingknowledge, that is, latest developments in specific knowl-edge fields, research techniques and so on (Hara, 2003).Training contributes to strategies that can be devised topromote motivation and reward human resources. How-ever Gray et al. (2004) stress that the influence of trainingdepends very much on the creation of an environmentwhere sufficient returns on investment in such activitycan be expected. It needs to be accompanied by pertinentincentives and working conditions so that improved skillsare adequately used (Laursen and Foss, 2003).

At global level pharmaceuticals firms are stronglyinclined to train personnel across operations (Bureau ofLabor Statistics, 2008). Training requirements range from afew hours of on-the-job training to years of formal educa-tion, including job experience. Training not only includesdevelopment of general skills, but also those needed tocarry out specific projects, develop particular processes,conduct specific analyses, handle specialized equipmentand so on. Firms frequently train in safety, environmentaland quality control and technological advances. Trainingin marketing and sales is expected to increase the marketsuccess of a product (Bureau of Labor Statistics, 2008). Fromthe above, this paper explores the following hypotheses.

H1. Training positively influences the likelihood that afirm performs in-house R&D.

H2. The nature of training and its impact on learningthrough R&D will differ depending on the nature of theactivities carried out by the firm.

Remuneration for performance: Adequate compensationand reward for performance should positively and signif-icantly impact on learning for innovation (Badawy, 1988).Appreciation of individual and professional aspirationspromotes motivation and commitment towards an organi-zation (Mumford, 2000; Quinn and Rubb, 2006). Effectivereward systems encourage employees to take risks, pur-sue the development of new products and continuouslygenerate ideas that can be realized (Mumford, 2000). Cre-ativity can be encouraged if freedom, financial rewards,promotion and other forms of recognition exist (Amabile,1997). Remunerations contribute to skill developmentcycles (Samstad and Pipkin, 2005); they may also strate-gically attract talent from outside thereby minimizing
costs of internal development (Labarca, 1999). However,setting adequate remuneration systems is complex. Cre-ative individuals may prefer an intellectually challengingenvironment over high salaries (Terziovski and Morgan,

5 ge and E

2h

Hi

Hi

pfp(ttmttdta2bpc

tiiriomoffi

dminslc

Hi

Hw

oaap2qiat


006). These considerations lead to formulate the followingypothesis.

3. In general, remuneration levels should positivelynfluence learning through R&D.

4. The influence from remunerations will vary depend-ng on the complexity of knowledge requirements.

Worker’s empowerment: Self-esteem—the feeling ofower—is an important determinant of employee per-ormance. Empowering employees is basic for high-erformance work systems. According to Bartlett et al.2002), people should be given the opportunity and meanso tackle new problems, to gain varied experiences, ando be prepared to take on more challenging tasks. People

ay participate in the definition of personal objectives,he time they spend at work. Employees should be ableo voluntarily involve in assignments that promote skillsevelopment, or establishment and management of effec-ive mentoring relationships (Hemmert, 1998). In such

way firms can foster discovery activities (Mumford,000). However Bartlett et al. (2002) warn that mismatchesetween increased responsibility, and means and skills toerform the job can render empowerment meaningless,ounterproductive even.

Working conditions in the pharmaceutical industryend to be among the best throughout manufacturing activ-ties. Cleanliness, health and safety are paramount in thendustry; worldwide pharmaceuticals firms customarilyank among the best places to work (GPWI). However,t must be acknowledged that strict regulations imposedn the pharmaceutical industry reduce opportunities toodify working conditions. Manufacturing processes and

perations in general, must comply with strict good manu-acturing practices3 and other quality and safety standards;rms work closely with regulatory authorities.

Regarding R&D, the literature documents that drugevelopment activities, such as those underpinning the for-ulation of generic drugs, are more structured and defined

n terms of timing, nature of tasks, formality in the orga-ization and conduction of activities. This is the generalituation in a country such as Mexico, were the bulk of R&Deads to the obtaining of generic products. In light of thisonflicting evidence, the expectation is as follows.

5. Workers’ empowerment can have positive albeit lim-ted effects on in-house R&D.

6. The effects associated with worker’s empowermentill be contingent on the goals of R&D.

3 In most countries, sanitary authorities ensure effectiveness and safetyf pharmaceutical products by implementing comprehensive safeguardsnd procedures of obligatory observance for drug manufacturers. Thesere summarized under good manufacturing practice (GMP) which, in sim-le terms, indicates the best rules/practices to manufacture drugs (FDA,004; Seiter, 2005). GMPs include layout and functionality of buildings,ualification and training of personnel, cleanliness and sanitation, mon-

toring, supervision and many other aspects. GMP’s are reviewed anddjusted according to scientific and technological advances, hence theerm “current” or cGMPs.


4. Data sources, variable definition and researchstrategy

This paper explores the likelihood that a pharmaceu-tical firm carries out in-house R&D in Mexico. A suitableapproach for studying this type of decision variables isa probability model, such as binary probit regression(Greene, 2003). Analysis starts with a basic model thatexplores the extent to which human resource manage-ment 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 furtheridentify expected R&D outcomes and corresponding typesof knowledge requirements for the firm. Hence one canappreciate how adoption of similar set of human resourcemanagement practices influence decisions to conduct R&Dfor knowledge exploitation or knowledge exploitation, orfor new/improved drugs or new/improved drug manufac-turing processes.

4.1. Data and data sources

Data used in this paper were extracted from the NationalSurvey on Employment, Wages, Technology and Training inthe Manufacturing Industry (Encuesta Nacional de Empleo,Salarios, Tecnología y Capacitación; henceforth ENESTYC).This survey was carried out by the National Institute forGeography, Statistics and Informatics in Mexico (InstitutoNacional de Estadística, Geografía e Informática; henceforthINEGI) on behalf of the Ministry of Labor and Social Pro-vision in Mexico (Secretaría del Trabajo y Previsión Social,STPS). ENESTYC represents the entire Mexican manufactur-ing sector. The manufacturing establishment constitutesthe unit of analysis. The survey builds on a stratified sam-ple based on the establishment’s size, as measured bytotal employment: large 251+; medium: 101–250; small:10–100 and micro: 0–5. Classification of activities is basedon the North American Industrial Classification System(NASCI). Establishments with 100 or more employees areincluded together with a random sample of those with lessthan 100 employees. Confidence level is 95 percent, withan estimated non-response of 10 percent.

Based on an agreement to comply with pertinent con-fidentiality requirements by INEGI, personnel from thisInstitute processed, on our behalf, the preliminary datafor the event 2005. The information corresponds to theyear 2004. ENESTYC provided information on technologicaland organizational profiles; employment and remunera-tion levels; management practices and the provision oftraining. This paper used data for the module for the phar-maceutical industry (NASCI code 3254). Such data included141 data points; however, the effective working sample,excluding missing values, is 112 data points. Due to theinability to match data points with specific firms, theremaining part of this paper uses, indistinctly, the termsestablishment and firm. However, firms can own more thanone establishment.

Additional data were collected through exploratoryinterviews carried out at some firms in Mexico. In total20 firms, both multinational and of Mexican origin, par-ticipated in the study. Interviews were carried out in 2007.

F. Santiago, L. Alcorta / Structural Change and E

Table 1Indicators on in-house R&D performance by pharmaceutical firms inMexico.

Variable Definition

(1) rd inhouse The firm carries out R&D in-house(2) rd improve process The goal of R&D is to improve

existing manufacturing processes(3) rd design meq The goal of R&D is to improve or

design new machinery andequipment for own use

(4) rd drug improvement The goal of R&D is to improveexisting pharmaceutical products

(5) rd drug design The goal of R&D is to design newpharmaceutical products

(6) rd exploit The firm performs R&D forknowledge exploitation

management activities, the latter are the actual, function-ing, observable activities, as experienced by employees.Written policies will influence performance only if

(7) rd explore The firm performs R&D forknowledge exploration

Authors, based on ENESTYC 2005, INEGI.

The aim was to learn about the nature of R&D activitiesand the associated human resource management practicesin the local pharmaceutical industry.

4.2. Dependent variables

Based on the information from ENESTYC, this paperstudies the extent to which pharmaceuticals firms doR&D in Mexico; this is by means of the variablerd inhouse, item (1) in Table 1. It was also possibleto identify the objectives of R&D. ENESTYC identifiesR&D supporting cost-reducing innovations through: (2)improvements in existing drug manufacturing processes(rd improve process); and (3) improvement or design ofnew machinery and equipment for the firm’s own use(rd design meq). Variable rd design meq is broadly inter-preted as R&D for new process innovation. Alternatively,R&D seeks demand-enhancing innovations including:(4) quality improvements on existing pharmaceuticalproducts (rd drug improvement); and (5) design of newpharmaceutical products (rd drug design).

The novelty of the R&D outcomes is defined taking thefirm as reference; innovations can be new to the firmbut not necessarily to the Mexican market or the world.Based on the discussion in Section 2, items (2) and (4) inTable 1 are interpreted as knowledge exploitation activi-ties, improvements either in pharmaceutical products ordrug manufacturing processes or both, lead to searcheswithin familiar knowledge bases. By contrast, the intro-duction of some new drugs or new drug manufacturingprocesses, indicators (3) and (5), relate to knowledegesearches outside familiar cognitive, including physical andgeographical, boundaries of the firm.4 This distinctioncoincides with Kale and Little’s (2007) differenciation ofpharmaceutical firms, based on their accumulated techno-
logical capabilities. By combining (2) and (4) a variable onR&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 pharma-ceuticals are found in Rothaermel and Deeds (2004), Gilsing (2006), andKettler and Modi (2001).


for knowledge exploration, rd explore was obtained. Ingeneral, firms in Mexico pursue imitative and incremen-tal innovations based on exploitation of already availableknowledge.

Correlation analysis in Appendix A reveals thatknowledge exploitation in general, particularly for theimprovement of drugs already in the firm’s portfolio, iswhat drives R&D of pharmaceutical firms in Mexico.5 Thereis high and statistically significant correlation betweenvariables denoting R&D for knowledge exploitation. Bycontrast, the weakest correlations relate to R&D under-pinning the design and improvement of machinery andequipment for the firm’s own use. Correlations amongvariables on R&D reflect the incremental nature of inno-vation carried out by pharmaceutical firms in Mexico.For example, quality enhancements of pharmaceuticalproducts refer to changes in formualtions so that phar-maceutical products meet requirements of bioequivalenceand biodisponibility of the active ingredient. Alternatively,firms seek to improve the packaging of products. New phar-maceutical products, in turn, include new vaccines, newapplications of existing drugs by combining excipients,reformulating or recombining existing molecules—oftenin different therapeutic area—designing novel medicaldevices and so on. Some local firms develop new genericsand excipients based on the application of biotechnologytechniques.

4.3. Explanatory variables

Table 2 presents the explanatory and control variablesused in this paper. As for the former, Boseli et al. (2005, p.74) acknowledge three forms to measure human resourcemanagement variables: “by its presence (that is, a dichoto-mous scale for whether it is actually in effect ‘yes’ or ‘no’),by its coverage (that is, a continuous scale for the propor-tion of the workforce covered by it) or by its intensity (thatis, a continuous scale for the degree to which an individualemployee is exposed to the practice or policy). The over-whelming majority [of studies] rely only on measures ofpresence.” In general, this is the case with ENESTYC. Onlya few variables in the dataset reflect intensity in the useof the corresponding human resource management prac-tice. For example, the indicator on worker’s participationin decision-making.

Wright et al. (2001) and Boseli et al. (2005) also advisecaution on differences in measuring management variablesin terms of either policies or practices. Whereas the for-mer reflect an organization’s stated intentions regarding

5 To further identify factors driving pharmaceutical R&D inMexico exploratory factor analysis was conducted includingrd process improvement, rd design meq, rd drug improvement andrd drug design; from the analysis one single factor was retained. Withinsuch factor, rd drug improvement had the largest factor loading, albeitsomewhat close to those for rd drug improvement and rd drug design.For the sake of simplicity of the analysis the paper reports only resultsfrom the correlation analysis.

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

Table 2Human resource management and control variables included in the analysis.

Description

train04 1 if the firm provided training to its employees in 2004; 0 otherwisetraining internal 1 if training is provided by colleagues in-house; 0 otherwiseexternal training 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 otherwiseinternal external tr Interaction term between training internal and external training. 1 if the firm provides training both in-house and

through external providers; 0 otherwiseln avg rem 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 yearrem size Interaction term between ln avg rem and firm size as defined belowimp empowerment 1 if workers participate in decision making and the firm declares that such practice is important; 2 not important; 0

workers do not participateControl variables

modern practice 1 if the firm reports the use of total quality management and/or just-in-time organizational practices irrespective of actualimportance; 0 otherwise

firm size Size of the firm 1 = Medium, small and micro, 2 = Largeexport dummy 1 if firms report participation in export markets; 0 otherwisefdi 1 if firms report foreign ownership; 0 otherwisefdi expt Interaction term between fdi and export dummy. 1 if firms report both foreign ownership and participation in export

markets; 0 otherwise

AN ariablei

iwisrpp

yrctfiowieiartpii

4

(dmncatca

uthors, based on ENESTYC 2005, INEGI.otes: Information for the 112 data points in working sample. Except for v

nformation from the source.

ndividuals perceive them as important for organizationalell-being. ENESTYC contains several variables represent-

ng regulations on management practices. Unfortunately,ince the dataset lacks detailed information on how suchules translate into actual practices, the analysis in thisaper focuses on the actual human resource managementractices reported by the firm.

From the above, and based on Delery’s (1998) the anal-sis in this paper used alternative constructs of humanesource management variables; hence it is possible toapture how different ways to implement a specific prac-ice result in distinct influences on performance at therm level. The paper includes four alternative definitionsf the provision of training, namely, whether trainingas provided in 2004, train04, and whether it took place

n-house, training internal, or through external suppliers,xternal training. We also used an interaction term betweennternal and external training, internal external tr. The vari-ble on staff remuneration is expressed as the average ofemmunerations, in natural logs, ln avg rem, and by con-roling for the size of the firm, rem size. Finally, worker’sarticipation in decision making takes into account the

mportance of such practice as perceived by the employers,mp empowerment.

.4. Control variables

Arundel et al. (2007) in the case of Europe, OECD1998) for the OECD countries and Kaplinsky (1995) foreveloping countries document the interrelation betweenodern human resource management practices and orga-

izational strategies adopted by firms. Such strategiesorrespond with the type of management practices avail-
ble for firms, and shape the environment in which learningakes place (Arundel et al., 2007). In the case of pharma-eutical firms, and in the context of cGMPs, TQM practicesssist in meeting the strict quality controls required by
s train04 and training internal, the rest were created by the authors with

regulatory authorities. In this study, the variable mod-ern practice controls for the use of JIT and/or TQM practices.Capital origin and export behavior condition stronglythe technological performance of pharmaceutical firms indeveloping countries (Kim et al., 1989; Zúniga et al., 2007).Foreign ownership will determine the perceived impor-tance of R&D for the firm’s business strategy in the hostcountry.

In the case of developing countries such as Mexico,multinational affiliates will generaly show rather passivetechnological behavior, as measured by R&D performance;R&D remains concentrated at the parent company. In thecase of exports, systematic R&D efforts assist firms in meet-ing some of the challenges derived from the increasedexposure to competition in foreign markets. This paperincorporates these observations via two dummy variables,fdi and export dumy, respectively.

4.5. Research strategy

In conducting the analysis, several checks were per-formed to ensure accuracy and robustness of results.Models were included, where each dependent variable wasregressed on the explanatory variables only; then com-pared to full specification models. Equations were also runincluding only those explanatory and control variables thatrevealed some statistical significance, at five percent or less,in the full specification model. Estimates from all thosealternative models are consistent with the results reportedin this paper.

Note a minor difference in the definition of trainingused in the model with R&D for new drug manufacturingprocesses, rd design meq. The majority of pharmaceutical
firms reported to have provided training to employees dur-ing 2004. Models with train04 had problems converging;the variable predicted perfectly the probability that a firmperforms such type of R&D (Long and Freese, 2006). The

ge and E
choice was for the alternative, internal external tr, which isan interaction term between in-house and externaly pro-vided training.

Normalizing the log of remunerations with respect tofirm’s size (rem size) corrected problems of high and pos-itive correlations between ln rem avg and the variables onFDI and exports, respectively (Appendix A and Table 2).This procedure also captures some scale effects associ-ated with firm size (Cockburn and Henderson, 2001). Hightcorrelations also involved variables on the use of JIT, fdiand exports. In the first case, high correlation with theindicator on worker’s empowerment was eliminated byusing the interaction term, modern practice, between JITand TQM. Then for fdi and export dummy, the interac-tion term fdi expt solved some problems. The latter allowedsome no-linearities into the models but with a downsize.Stimates including all three variables, fdi, export dummyand fdi expt, were problematic as they tended to predictperfectly the learnimg behavior of firms. Correcting forthis—see Long and Freese (2006), was partial solution ascolinearity problems increased. In STATA the solution tocolinearity is to drop the redundant variable, fdi expt. Thepaper reports models with and withouth fdi expt; resultsremain consistent.

5. Empirical results

5.1. Learning behavior of pharmaceutical firms in Mexico

Appendix B summarizes the learning behavior of phar-maceutical firms in Mexico. The 74.1 percent of firmsperformed R&D in 2004, with 63.4 percent and 70.5 percentfocusing on process and product innovations, respectively.Of those firms conducting R&D for process innovation, 25.3percent did so to improve or design machinery for own use,while 63.4 percent to improve drug manufacturing pro-cesses. As for demand-enhancing innovations, 61.1 percentof firms pursued new pharmaceutical products, and 66.1percent focused on improvements in existing drugs.

Indicators on sales and employment show that, on aver-age, R&D performers slightly outperform those reportingno R&D. For instance, average employment, total sales andsales per employee are, respectively, 1.4, 1.6 and 1.1 timeslarger in firms with active learning strategies. By contrast,indicators on capital origin and export orientation tendedto favor non-R&D performers. Some 70 percent of firmscarried out either knowledge exploitation or exploration.The corresponding figures on employment, sales and soon, are very close among each group, yet with a slightadvantage for active learners. Some 60 percent of firms inthe sample participated in export markets. However, sincethe average share of exports in total sales of the indus-try is rather modest, one can argue that pharmaceuticalfirms predominantly serve the local market. In line with thecGMP’s requirement, ENESTYC reports extensive adoptionof modern manufacturing practices in the pharmaceuticalindustry.

As for the nature of human resource management inthe pharmaceutical industry in Mexico, firms show greatpropensity to provide training to employees. The practiceis more frequent in connection with R&D for knowledge


exploitation, and by firms pursuing drug innovation. Firmscombine both internal and external sources of training insearch for synergistic effects between the two types oftraining.

ENESTYC documents that in Mexico, remunerationsin the pharmaceutical industry are higher than in othermanufacturing industries. They are even higher in firmsperforming in-house R&D. Nevertheless, our interviewssuggest that, as a mechanism to motivate and retain work-ers, adjustment in remunerations is frequently limited bythe need to maintain balance of the firm’s overall compen-sations structure.

ENESTYC also documents that firms in the local phar-maceutical industry seldom allow worker participation indecision-making about working conditions. Even in thoseareas where workers have a voice, the practice is reportedas having little importance for the company. In this regardthe exploratory interviews revealed that in firms with someincipient R&D efforts, R&D staff frequently subordinate tomanufacturing and quality control.

5.2. Econometric analysis

5.2.1. Knowledge exploitation or explorationThe literature suggests that knowledge exploration,

in the sense of research, experimentation and techno-logical capability-building, should associate with strongerexigencies on human resource management. This is as com-pared to knowledge exploitation. This section explores thishypothesis.

Table 3 presents estimates from the econometric analy-sis. Model (1) corresponds to in-house R&D, irrespective ofthe goal pursued by the firm. Then, model (2) captures R&Dfor 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 withexplanatory variables but without controls, and then thosewith the full set of variables. In order to ensure that corre-lation between fdi and export dummy does not cause majorproblems, a third column includes models with the inter-action term fdi expt. Finally, the computation of marginaleffects for the model with the full set of explanatory vari-ables is presented. The Wald tests for the value of X2, whichis different from zero, confirm that the models are statisti-cally significant at standard confidence levels. The count R2

show that, in general, the predictive power of each modelis acceptable (Long and Freese, 2006). The values of theCragg–Uhler R2 suggest that the models adequatly explainthe probability that a firm performs R&D.

Individual estimates reveal that training has the mostsignificant effect on learning through different kinds ofR&D; yet the effect looks fairly similar in the case ofboth knowledge exploitation and knowledge exploration.Remunerations in turn, show positive impact on rd explorebut the effect seems not to be robust. Scale effects are alsocaptured by the variable on remunerations, as it is nor-malized 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. Whereasexport participation induces learning, foreign ownershipinhibits it. In fact, the influence of capital ownership is

538F.

Santiago, L.

Alcorta

/ Structural

Change and

Economic

Dynam

ics 23 (2012) 530– 546

Table 3Influence of management practices on knowledge exploitation and exploration by pharmaceutical firms in Mexico.

Variables (1) rd inhouse (2) rd exploit (3) rd exploreMg effect Mg effect Mg effect

train04 1.25*** 1.32*** 1.30*** 0.48*** 1.18*** 1.27*** 1.28*** 0.47*** 1.05** 1.22** 1.15** 0.46***

(0.41) (0.44) (0.43) (0.15) (0.41) (0.43) (0.43) (0.15) (0.42) (0.49) (0.47) (0.16)rem size 0.04 0.06 0.09 0.18 0.04 0.08 0.10 0.03 0.06 0.10* 0.14** 0.04*

(0.05) (0.06) (0.06) (0.02) (0.05) (0.06) (0.06) (0.02) (0.04) (0.06) (0.06) (0.02)imp empowerment −0.11 −0.14 −0.16 −0.04 −0.19 −0.18 −0.20 −0.06 0.01 0.023 −0.010 0.01

(0.20) (0.23) (0.23) (0.07) (0.19) (0.22) (0.22) (0.07) (0.19) (0.22) (0.21) (0.08)modern practice 0.26 0.28 0.08 0.17 0.20 0.06 0.29 0.33 0.11

(0.35) (0.34) (0.11) (0.35) (0.34) (0.12) (0.32) (0.31) (0.12)export dummy 0.71** 0.22** 0.46 0.15 0.98*** 0.36***

(0.35) (0.11) (0.35) (0.11) (0.36) (0.12)fdi −0.99*** −0.33** −0.94** −0.33** −1.39*** −0.51***

(0.38) (0.13) (0.37) (0.13) (0.41) (0.13)fdi expt −0.68** −0.81** −0.96***

(0.34) (0.34) (0.34)Constant −0.63 −1.04** −0.98* – −0.62 −1.01** −1.01* – −1.04** −1.69*** −1.56*** –

(0.46) (0.51) (0.52) – (0.45) (0.50) (0.51) – (0.47) (0.60) (0.57) –

Observations 112Log Likelihood full −58.6 −55.1 −56.5 – −62.0 −58.9 −59.1 – −69.0 −61.5 −64.6 –X2 10.5 [3]** 19.7 [6]*** 14.0 [5]** – 9.84 [3]** 16.1 [6]** 14.3 [5]** – 9.06 [3]*** 20.4 [6]*** 14.9 [5]** –Cragg–Uhler R2 0.137 0.218 0.186 – 0.123 0.194 0.190 – 0.117 0.275 0.213 –Count R2 0.777 0.768 0.768 – 0.750 0.750 0.741 – 0.679 0.741 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 variablesdefinitions see Tables 1 and 2.

* p < 0.1.** p < 0.05.

*** p < 0.01.

F. Santiago,

L. A

lcorta /

Structural Change

and Econom

ic D

ynamics

23 (2012) 530– 546539

Table 4Results from probit analysis: management practices and learning in the Mexican pharmaceutical industry.

Variables (4) rd improve process (5) rd design meq (6) rd drug improvement (7) rd drug design

train04 1.13*** 1.10** 1.12** 0.95** 1.01** 1.03** 1.05** 1.16** 1.08**

(0.42) (0.44) (0.44) (0.41) (0.43) (0.43) (0.42) (0.50) (0.47)internal external tr 0.05 0.45** 0.47**

(0.05) (0.19) (0.20)rem size 0.02 0.06 0.08 0.53*** 0.11** 0.13** 0.06 0.10* 0.12** 0.07 0.11** 0.15***

(0.04) (0.05) (0.05) (0.20) (0.05) (0.05) (0.04) (0.05) (0.05) (0.04) (0.06) (0.06)imp empowerment −0.04 −0.13 −0.15 0.42** 0.74*** 0.74*** −0.13 −0.13 −0.14 −0.07 −0.12 −0.15

(0.19) (0.21) (0.21) (0.20) (0.21) (0.22) (0.19) (0.21) (0.21) (0.19) (0.21) (0.21)modern practice 0.51 0.53 −0.29 −0.28 0.16 0.18 0.48 0.51

(0.32) (0.32) (0.35) (0.36) (0.32) (0.32) (0.32) (0.32)export dummy 0.28 0.21 0.31 1.01***

(0.31) (0.36) (0.32) (0.36)fdi −0.70** −0.95** −0.76** −1.39***

(0.34) (0.44) (0.35) (0.42)fdi expt −0.67** −1.03** −0.73** −0.95***

(0.32) (0.45) (0.33) (0.35)Constant −0.81* −1.26** −1.27** −2.65*** −2.94*** −3.01*** −0.77* −1.11** −1.13** −1.09** −1.80*** −1.66***

(0.46) (0.51) (0.51) (0.58) (0.64) (0.67) (0.45) (0.50) (0.51) (0.47) (0.61) (0.59)

Observations 112Log likelihood full −69.3 −66.3 −66.0 −45.7 −42.8 −42.2 −67.6 −65.3 −65.1 −69.4 −61.1 −64.5X2 8.11 [3]** 13.6 [6]** 13.6 [5]** 16.5 [3]*** 24.1 [6]*** 23.8 [5]*** 8.25 [3]** 13.2 [6]** 12.7 [5]** 9.54 [3]*** 20.4 [6]*** 14.4 [5]**

Cragg–Uhler R2 0.101 0.166 0.172 0.225 0.295 0.307 0.100 0.150 0.156 0.121 0.290 0.224Count R2a 0.688 0.661 0.670 0.795 0.795 0.821 0.696 0.679 0.679 0.670 0.723 0.732

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.

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

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

(1) min → max (2) 0 → 1 (3) −+1/2 (4) −+sd/2 (5) MargEfcta (6) MargEfctb,c

rd impr proc 0.642train04 0.418 0.418 0.394 0.127 0.412 0.418 (0.146)***

rem size 0.199 0.025 0.024 0.070 0.024 0.024 (0.019)imp empowerment −0.101 −0.050 −0.050 −0.035 −0.050 −0.050 (0.080)modern practice 0.192 0.192 0.187 0.089 0.189 0.192 (0.123)export dummy 0.103 0.103 0.103 0.052 0.103 0.103 (0.117)fdi −0.265 −0.265 −0.255 −0.120 −0.260 −0.265 (0.128)**

rd design meq 0.124internal external tr 0.188 0.027 0.093 0.099 0.092 0.092 (0.035)***

rem size 0.218 0.008 0.023 0.069 0.023 0.023 (0.011)**

imp empowerment 0.420 0.154 0.152 0.106 0.151 0.151 (0.050)***

modern practice −0.062 −0.062 −0.059 −0.028 −0.059 −0.062 (0.081)export dummy 0.042 0.042 0.042 0.021 0.042 0.042 (0.071)fdi −0.160 −0.160 −0.197 −0.091 −0.195 −0.160 (0.062)**

rd drug imp 0.672train04 0.387 0.387 0.354 0.113 0.366 0.387 (0.152)**

rem size 0.295 0.039 0.036 0.107 0.036 0.036 (0.019)*

imp empowerment −0.093 −0.045 −0.045 −0.032 −0.045 −0.045 (0.076)modern practice 0.057 0.057 0.057 0.027 0.057 0.057 (0.119)export dummy 0.112 0.112 0.111 0.056 0.111 0.111 (0.117)fdi −0.283 −0.283 −0.269 −0.127 −0.274 −0.283 (0.132)**

rd design 0.642train04 0.437 0.437 0.413 0.134 0.433 0.437 (0.161)***

rem size 0.341 0.042 0.042 0.124 0.042 0.042 (0.021)**

imp empowerment −0.092 −0.045 −0.045 −0.032 −0.045 −0.045 (0.079)modern practice 0.181 0.181 0.176 0.084 0.177 0.180 (0.121)export dummy 0.368 0.368 0.364 0.187 0.377 0.368 (0.123)***

fdi −0.510 −0.510 −0.484 −0.238 −0.518 −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 changesfrom 1/2 standard deviation below base to 1/2 standard deviation above; MargEfct: partial derivative of the predicted probability/rate with respect to agiven 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.

saespMtsr

mwaepmsitificdk

* Significance at the 10% level.** Significant at the 5% level.

*** Significant at the 1% level.

tronger than that of exports. This is evident by lookingt models with fdi expt as explanatory variable. Mod-rn practice and imp empowerment did not reveal anypecific effect on learning. Overall, the estimates suggest aassive learning behavior of the pharmaceutical industry inexico. The constant term is consistently negative and sta-

istically significant. If all right-hand side coefficients wereet at zero, the probability that a firm carries out R&D isather low.

A complementary way to look at results from probitodels is by computing the marginal effects associatedith modifications in the value of a given explanatory vari-

ble (Christofides et al., 1997, 2000). A fourth column forach model in Table 3 presents the marginal and discreterobability changes for the variables in the full specificationodels. Estimates confirm that training has positive and

tatistically significant impacts on learning. If all remain-ng variables in the equation are left constant, in this case atheir mean value, the shift from non- to provision of train-ng increases, by some 48 percent, the probability that a
rm carries out in-house R&D. The effects of training asso-iated with knowledge exploitation were not statisticallyifferent from those of training in the context of R&D fornowledge exploration.
Contrary to our research hypothesis, findings in Table 3suggest that the influence of human resource managementon the likelihood that a pharmaceutical firm does R&D israther limited. Moreover, it is difficult to perceive how dis-tinct types of learning activities associate with differenthuman resource management practices. Based on the dis-cussion in Section 2, in what follows a further distinctionis made on the innovation outcomes expected from R&D.

5.2.2. Learning through different kinds of R&DTable 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 ongoodness of fit corroborate the adequacy of the models. Asexpected, the more detailed definitions of R&D provide bet-ter information on how the contribution of managementpractices to learning differs. Relevant practices vary bothin number and strenght of the corresponding effects. This
supports the idea that firms with divergent learning andinnovation strategies should gain differently from adoptionof even comparable human resource management prac-tices (Laursen and Mahnke, 2001; Laursen and Foss, 2003).

F. Santiago, L. Alcorta / Structural Change and Economic Dynamics 23 (2012) 530– 546 541

Table 6Testing 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* 0.70** 0.31 0.59* 0.24 0.40(0.33) (0.32) (0.33) (0.32) (0.33) (0.32)

external training 0.63** 0.46 0.79** 0.50* 0.75** 0.35(0.31) (0.31) (0.32) (0.31) (0.32) (0.31)

rem size 0.02 0.04 0.06 0.03 0.08 0.08(0.06) (0.06) (0.06) (0.05) (0.06) (0.06)

imp empowerment −0.11 −0.15 0.035 −0.12 −0.11 −0.11(0.24) (0.23) (0.22) (0.22) (0.22) (0.21)

modern practice 0.31 0.22 0.34 0.54* 0.52* 0.21(0.35) (0.34) (0.31) (0.32) (0.32) (0.32)

export dummy 0.71** 0.42 0.99*** 0.23 1.01*** 0.28(0.33) (0.34) (0.33) (0.32) (0.33) (0.31)

fdi −0.94** −0.89** −1.31*** −0.63* −1.29*** −0.69*

(0.38) (0.38) (0.40) (0.35) (0.41) (0.36)Constant −0.56 −0.57 −1.21*** −0.90** −1.32*** −0.68*

(0.44) (0.43) (0.44) (0.42) (0.45) (0.41)

Observations 112Log likelihood full −55.5 −59.3 −60.9 −65.9 −60.9 −66.4X2 [7] 23.2*** 15.6** 30.1*** 14.1*** 29.3*** 11.5Cragg–Uhler R2 0.207 0.185 0.287 0.135 0.295 0.127Count R2 0.777 0.741 0.696 0.705 0.723 0.688

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.

ing internal and external training, were brought into theanalysis. Table 6 contains estimates for models wheretrain04 is replaced by the two new variables.6 The Wald

* p < 0.1.** p < 0.05.

*** p < 0.01.

Estimates indicate that the provision of training remainsthe most significant practice across learning activities. Andyet, the more evident contribution corresponds to R&Dfor new drug designs, followed by improvements in drugmanufacturing processes. Variables on remunerations andworker’s empowerment also gain statistical significance.Remunerations are important for exploratory R&D lead-ing to new drug manufacturing processes, and new drugdesigns. Worker’s empowerment has positive effects onrd design meq. Export dummy and fdi continue to play rel-evant roles, with effects running in opposite directions, forrd drug design.

Table 5 presents the computation of marginal effectsfor variables in the basic models of Table 4. Unlike theanalysis in Table 3 this new excercise is much moredetailed. Estimates confirm training as the interventionwith the largest positive and statistically significant impacton the likelihood that a firm performs R&D. In fact, thelargest effect of training is on rd drug design, 43.7 percent.By contrast, the lowest influence, some nine percent, isin the case of rd design meq. The latter variable is also theone where worker’s empowerment has perceptible andpositive contributions to learning. Marginal increases inremunerations have positive and statistically significantinfluence on knowledge exploration. This result suggeststhat as firm grow, so does their capacity to remunerateR&D personnel.

Interpretation of discrete probability changes should behandled with care, as they are meaningful only for vari-ables spanning over a sufficiently large range of values
(Long and Freese, 2006). A pertinent case is that of remu-nerations. Column (1) in Table 5 reveals that a change inthe log of remunerations, equivalent to an increase fromminimum to maximum, raises the likelihood that a firm
conducts rd drug design by some 0.341. Changes in remu-nerations are stronger for demand-enhancing R&D than forcost-reducing activities. Changes induced by increases ofhalf a standard deviation in the log of remunerations, col-umn (4), are larger for rd design than for any other typeof process R&D. Except for rd design meq, worker’s partic-ipation in decision-making failed to provide meaningfulinformation about its likely influence on the likelihood ofR&D performance.

5.3. Effects from different types of training

Similar to previous studies on innovation and humanresource management, this paper has documented that theprovision of training has positive and robust influenceson the likelihood that firms do R&D. In order to extractsome more meaningful conclusions, more disaggregatedmeasures on the actual nature of training were intro-duced. Section 4.3 identified two complementary forms:internal (on-the-job) and external (off-the-job). The formerwas expected to support knowledge diffusion and shar-ing within the organization, it would more closely relateto knowledge exploitation strategies. External training,in turn, was expected to support expansion of knowl-edge bases through interactions with other knowledgeproducers (Casas, 2005). In order to explore this dualnature of training, two additional variables, namely, train-

6 The analysis excluded the variable on rd design meq because train-ing internal tended to predict perfectly the probability that a firm performsthis specific activity.

5 ge and E

ttv

mtoreTpf

6

mfttacormabteineb

dIawalovreiodttei

apTa

rm


ests show that, with the exception of rd drug improvement,he remaining models are statistically significant at con-entional confidence levels.7

Estimates in Table 6 confirm that internal training isore closely related to knowledge exploitation, while

hat provided by external agents impacts more directlyn learning through knowledge exploration, particularlyd drug design. Note that remunerations and worker’smpowerment lose the explanatory power found inables 3 and 4. Exposure to competition through partici-ation in export markets stimulates learning, particularlyor (new) pharmaceutical products.

. Discussion and conclusions

This paper investigated the influence of human resourceanagement practices on the likelihood that a firm per-

orms in-house R&D. Based on the relevant literature,he focus was on the provision of training, remunera-ions and worker’s participation in decision-making. Thenalysis was carried out in the context of a developingountry such as Mexico, and was based on the notionsf knowledge exploitation and knowledge exploration,espectively. Results show that human resource manage-ent influences innovation by stimulating, first, learning

nd capacity-building through in-house R&D. The num-er of relevant human resource management practices andheir corresponding influence is contingent on the nov-lty of the knowledge required by the firm. The lattern turn, is linked to expected R&D outcomes. The moreovel the expected R&D outcome, the more notorious influ-nce of human resource managament practices tends toe.

Differences in the nature of R&D induces distinctiveemands on the human resources shaping an organization.

n line with the literature on human resource managementnd new product development, R&D for new drug designsas found to be positively associated with practices such

s training and remunerations (Lund, 2004a). Neverthe-ess evidence was also obtained on the positive impactf management practices on R&D for new process inno-ation, technical change more broadly defined. Training,emuneration for performance and incorporation of work-rs into decision-making supported R&D for the design ormprovement of machinery and equipment. To the best ofur knowledge, this is one of the first studies in this fieldocumenting this issue in the context of developing coun-ries. Data limitations prevented further investigation intohis finding; nevertheless this is a relevant issue consid-ring that process innovations enjoy a significant share ofnnovations in developing countries.

As for the effects associated to specific personnel man-gement interventions, and in light of the hypotheses
resented in Section 3, relevant findings are as follows:he provision of training systematically and positivelyffects the likelihood that firms pursue R&D; hence
7 Although not included in Table 6, models were also ran where fdi expteplaced the individual variables fdi and export dummy. Results from suchodels are consistent with the conclusions presented here.


hypothesis H1 is confirmed. This lends support to Samstadand Pipkin (2005)’s claim that training and general qualifi-cations of the labor force dictate the type of human resourcemanagement practices needed and feasible in countriessuch as Mexico. Raising skill levels facilitates adoption ofadvanced management systems in Mexican firms; more-over, it assists in building required capabilities to moresistematically conduct R&D.

This paper supported the pertinence to promote inter-actions between firms and other external agents, at leastfor the provision of R&D-relevant training. Hence firmscan access new knoweldge and expand existing knowledgebases. Further research should shed light on the natureof the actual knowledge flows being involved. However,exploratory interviews among Mexican drug manufacturessuggest that interactions are broad. They include learningabout new excipients and formulations, to methodolo-gies for the synthesis of chemical ingredients. In yet someother cases, external training provides understanding ofadvanced research methodologies and applications, partic-ularly in areas such as biotechnology. Overall, hypothesisH2 was also confirmed.

The literature review in Section 3 suggested that, inprinciple, remunerations should influence learning pos-itively. Estimates revealed that raising remunerationsincreases the probability that a firm performs in-houseR&D, particularly for knowledge exploration. However, theeffect was not robust. Consequently hypothesis H3 is par-tially supported; hypothesis H4 seems more plausible.Remunerations underpin learning but only under certainconditions and for specific types of R&D. Albeit difficultto corroborate based on data used here, a possible expla-nation results from the frequent mark-up on pecuniaryremunerations, more specifically wages. In countries suchas Mexico opportunities to training and/or perspectivesfor promotion become equally or even more relevant asreward mechanisms. Alternatively, remunerations servemore as determinants of labor mobility within the pharma-ceutical industry; thus promoting a continuous transfer ofresearch capabilities, however limited, within the industry(Santiago, 2010).

Somewhat inconclusive results were drawn in the caseof worker’s empowerment. The practice was positive andsignificant only in the case of exploration-related R&D fornew drug manufacturing processes. This is at odds withprevious literature indicating that delegation of decision-making capacity is key for new product development,as it fosters creativity and discovery (Mumford, 2000). Apossible explanation points at the traditional perceptionthat paternalistic work environments, rigid and hierarchi-cal oganizational structures, such as those generaly foundin Mexico and other similar countries, are unsuitable forenhanced performance. Nevertheless, as Section 3 in thispaper acknowledged, the nature of drug manufacturing canlimit the influence of worker’s empowerment on the scopeof R&D. Concerns over drug quality and safety lead to closescrutiny and approval by sanitary authorities thereby lim-
iting the capacity to change both existing drugs and thecorresponding manufacturing processes. Any alteration ineither of them may require additional reviews and approvalby the regulatory authorities. FDA (2004) asserts that this

F. Santiago,

L. A

lcorta /

Structural Change

and Econom

ic D

ynamics

23 (2012) 530– 546543

Table A1Correlation analysis of variables on management practices and firm characteristics considered for the analyses.

Mean Min Max (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18)

(1) rd inhouse 0.741 0 1 1.00(2) rd exploit 0.714 0 1 0.93 1.00

(0.00)(3) rd explore 0.625 0 1 0.76 0.69 1.00

(0.00) (0.00)(4) rd improve process 0.634 0 1 0.78 0.83 0.64 1.00

(0.00) (0.00) (0.00)(5) rd design meq 0.188 0 1 0.28 0.30 0.37 0.37 1.00

(0.00) (0.00) (0.00) (0.00)(5) rd drug design 0.616 0 1 0.75 0.68 0.98 0.62 0.33 1.00

(0.00) (0.00) (0.00) (0.00) (0.00)(6) rd drug improvement 0.661 0 1 0.82 0.88 0.73 0.71 0.34 0.71 1.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

(7) ln avg rema 4.735 3.137 5.914 0.22 0.17 0.22 0.13 0.12 0.23 0.22 1.00(0.02) (0.08) (0.02) (0.16) (0.20) (0.01) (0.02)

(8) rem sizeb 6.846 3.137 11.828 0.11 0.10 0.17 0.08 0.20 0.18 0.15 0.71 1.00(0.27) (0.29) (0.07) (0.39) (0.04) (0.05) (0.11) (0.00)

(9) train04 0.893 0 1 0.32 0.29 0.27 0.28 0.17 0.26 0.24 0.14 0.13 1.00(0.00) (0.00) (0.00) (0.00) (0.08) (0.01) (0.01) (0.14) (0.19)

(10) training internal 0.786 0 1 0.24 0.25 0.18 0.24 0.25 0.17 0.18 0.22 0.27 0.66 1.00(0.01) (0.01) (0.06) (0.01) (0.01) (0.07) (0.06) (0.02) (0.00) (0.00)

(11) external training 0.750 0 1 0.27 0.23 0.32 0.25 0.17 0.31 0.20 0.21 0.27 0.60 0.30 1.00(0.00) (0.02) (0.00) (0.01) (0.07) (0.00) (0.04) (0.03) (0.00) (0.00) (0.00)

(12) internal external trc 2.286 0 3 0.31 0.28 0.33 0.29 0.24 0.32 0.23 0.26 0.33 0.74 0.63 0.93 1.00(0.00) (0.00) (0.00) (0.00) (0.01) (0.00) (0.02) (0.01) (0.00) (0.00) (0.00) (0.00)

(13) imp empowerment d 0.491 0 2 0.01 −0.04 0.07 0.03 0.29 0.03 −0.01 0.10 0.12 0.16 0.09 0.11 0.12 1.00(0.94) (0.70) (0.47) (0.75) (0.00) (0.76) (0.92) (0.27) (0.20) (0.09) (0.36) (0.24) (0.19)

(14) modern practice 0.670 0 1 0.10 0.06 0.12 0.18 0.05 0.15 0.06 0.14 −0.01 0.25 0.14 0.12 0.15 0.44 1.00(0.27) (0.53) (0.20) (0.06) (0.63) (0.12) (0.54) (0.14) (0.88) (0.01) (0.14) (0.21) (0.11) (0.00)

(15) export dummy 0.536 0 1 0.14 0.08 0.20 0.07 0.08 0.22 0.09 0.63 0.41 0.14 0.26 0.12 0.20 0.09 0.15 1.00(0.13) (0.37) (0.03) (0.44) (0.40) (0.02) (0.35) (0.00) (0.00) (0.14) (0.01) (0.19) (0.03) (0.34) (0.13)

(16) fdi 0.313 0 1 −0.04 −0.09 −0.07 −0.05 −0.03 −0.06 −0.05 0.51 0.48 0.17 0.26 0.08 0.16 0.19 0.19 0.55 1.00(0.67) (0.37) (0.43) (0.62) (0.77) (0.52) (0.63) (0.00) (0.00) (0.07) (0.01) (0.41) (0.09) (0.05) (0.05) (0.00)

(17) fdi expt 0.295 0 1 −0.07 −0.11 −0.11 −0.08 −0.06 −0.09 −0.07 0.52 0.48 0.16 0.24 0.06 0.14 0.16 0.16 0.60 0.96 1.00(0.50) (0.24) (0.27) (0.41) (0.53) (0.33) (0.43) (0.00) (0.00) (0.09) (0.01) (0.55) (0.14) (0.09) (0.09) (0.00) (0.00)

(18) firm sizee 1.411 1 2 0.08 0.09 0.16 0.07 0.20 0.17 0.14 0.50 0.96 0.11 0.26 0.27 0.32 0.12 −0.07 0.27 0.38 0.38 1.00(0.41) (0.37) (0.09) (0.47) (0.03) (0.07) (0.15) (0.00) (0.00) (0.23) (0.01) (0.00) (0.00) (0.23) (0.47) (0.00) (0.00) (0.00)

Authors based on ENESTYC, 2005; INEGI.Notes: standard deviations: a0.675, b2.997, c1.069, d0.697, e0.494. For variables definitions see Tables 1 and 2. p values in parenthesis.

5 ge and E

ciodnbewH

TS

AF


an be cumbersome for the firm; a major barrier for processnnovation in the pharmaceutical industry. In the contextf countries specialized in the manufacturing of genericrugs, development of such products is restricted by theeed to comply with specific parameters and qualities set
y the drug innovator. If firms only reproduce the knowl-dge behind such products, it makes little sense to alloworkers to play around with the technology. Hypothesis5 is supported but H6 requires further scrutiny.
able B1ummary statistics for the pharmaceutical industry in Mexico, 2004.

Mean Standar deviat

R&D in-house Internalc (I) No R&Dd (II) (I)/(II) Internal

Employment 475.7 331.2 1.4 555.2Total salesa 694094.8 433261.5 1.6 1270892

Domestic sales 609320.3 394477.4 1.5 1055332

Export share 0.07 0.08 0.9 0.13

Share of FDI 0.30 0.34 0.9 0.46

Ageb 33.2 27.5 1.2 19.4

Improved process Imp proce No R&Df Imp proc

Employment 492.5 344.3 1.4 589.2

Total salesa 741488.3 427531.3 1.7 1354405.0

Domestic sales 656732.5 375254.1 1.8 1120739.0

Export share 0.1 0.1 0.6 0.1

Share of FDI 0.3 0.3 0.9 0.5

Ageb 33.2 29.2 1.1 20.6

New process Mach & equipg No R&Dh Mach & equip

Employment 655.0 388.3 1.7 804.2

Total salesa 1140099.0 508048.1 2.2 1808071.0

Domestic sales 919528.0 469267.5 2.0 1307236.0

Export share 0.1 0.1 1.4 0.2

Share of FDI 0.3 0.3 0.9 0.5

Ageb 39.2 30.0 1.3 17.7

Improved drug Imp drugi No R&Dj Imp drug

Employment 496.6 324.7 1.5 577.4

Total salesa 738053.8 409433.5 1.8 1328800.0

Domestic sales 654131.2 358097.9 1.8 1101783.0

Export share 0.1 0.1 0.6 0.1

Share of FDI 0.3 0.3 0.9 0.5

Ageb 34.2 26.8 1.3 19.8

New drug drug designk No R&Dl drug design

Employment 526.1 297.4 1.8 592.1

Total salesa 765674.0 403324.4 1.9 1367771.0

Domestic sales 676530.3 356577.6 1.9 1134408.0

Export share 0.1 0.1 0.7 0.1

Share of FDI 0.3 0.3 0.8 0.5

Ageb 34.6 27.1 1.3 19.9

Exploitation rd exploit (I)m No R&D (II)n rd exploit

Employment 475.9 344.3 1.4 560.7

Total salesa 708400 421951.7 1.7 1291759

Domestic sales 626251.9 372290 1.7 1071360

Export share 0.1 0.1 1 0.1

Share of FDI 0.3 0.4 0.7 0.4

Ageb 33.3 27.7 1.2 19.7

Exploration rd exploreo No R&Dp rd explore

Employment 488.1 319.0 1.5 565.6

Total salesa 705485.3 437609.6 1.6 1292822

Domestic sales 620633.8 393435.2 1.6 1073893

Export share 0.1 0.1 1 0.1

Share of FDI 0.3 0.3 1 0.5

Ageb 33.6 27.2 1.2 19.4

uthors, based on ENESTYC 2005, INEGI.irms in sample: 112. Number of firms: c(83); d(29); e(71); f(41); g(21); h(91); i(7

a Thousand Mexican pesos.b Difference between the year in which a firm started operations in current bus


Some comments in relation to estimates associated withthe control variables are pertinent. In relation to foreignownership, the findings here contradict the usual percep-tion that foreign firms are more technologically dynamicthan domestic firms. The choice of performance indica-
tors is important. In terms of R&D, a careful reflectionpoints to the position that countries such as Mexico occupywithin business and innovation strategies of multination-als. Local affiliates maintain low profiles when assisting in
ion Min Max

No R&D Internal No R&D Internal No R&D259.1 1.1 63 3391.5 1158.4694938.1 2394 12127.5 6958020 2297038634741.2 2394 0 6334508 20697990.20 0 0 0.69 10.48 0 0 1 116.6 1 0 74 70

No R&D Imp proc No R&D Imp proc No R&D261.3 1.1 63.0 3391.5 1158.4641430.5 2394.0 12127.5 6958020.0 2297038.0583423.9 2394.0 0.0 6334508.0 2069799.00.2 0.0 0.0 0.6 1.00.5 0.0 0.0 1.0 1.015.1 1.0 0.0 74.0 70.0

No R&D Mach & equip No R&D Mach & equip No R&D386.7 2.2 1.1 3391.5 2852.9914265.9 31859.5 2394.0 6958020.0 6772189.0856102.8 31859.5 0.0 4359928.0 6334508.00.1 0.0 0.0 0.6 1.00.5 0.0 0.0 1.0 1.018.7 16.0 0.0 74.0 72.0

No R&D Imp drug No R&D IImp drug No R&D261.5 1.1 63.0 3391.5 1158.4653886.7 2394.0 7717.9 6958020.0 2297038.0587750.3 2394.0 0.0 6334508.0 2069799.00.2 0.0 0.0 1.0 1.00.5 0.0 0.0 1.0 1.015.9 1.0 0.0 74.0 70.0

No R&D drug design No R&D drug design No R&D238.0 2.2 1.1 3391.5 1158.4631241.8 2394.0 7717.9 6958020.0 2297038.0564394.8 2394.0 0.0 6334508.0 2069799.00.2 0.0 0.0 0.7 1.00.5 0.0 0.0 1.0 1.016.2 1.0 0.0 74.0 70.0

No R&D rd exploit No R&D rd exploit No R&D276.6 1.12 63 3391.5 1158.4665506.5 2394 12127.5 6958020 2297038607670.3 2394 0 6334508 20697990.2 0 0 0.6 10.5 0 0 1 116.1 1 0 74 70

No R&D rd explore No R&D rd explore No R&D249.1 1.12 63 3391.5 1158.4693483.2 2394 7717.9 6958020 2297038623592.3 2394 0 6334508 20697990.2 0 0 0.7 10.5 0 0 1 116.7 1 0 74 70

4); j(38); k(69); l(43); m(80); n(32); o(79); p(33).

iness and the year of the survey, 2004.

ge and E
the exploitation of knowledge generated at the parent loca-tion or elsewhere in the developed world (von Zedtwitzand Ove, 2002). Acquisition of new knowledge, demand-ing more advanced R&D activities, is seldom carried out indeveloping countries.

By contrast, exposure to external competition and largermarket opportunities, via exports, was found to increasethe likelihood that a firm pursues R&D. The strongest effectis associated with new drug designs. In line with Kale andLittle’s (2007), the managing director of a foreign affiliateargued that “Success requires strong commitment of finan-cial and human resources, particularly in research. The goalis to develop a portfolio of products to be launched in exportmarkets over a significant time horizon”. In the case of theMexican industry, strong reliance on the local pharmaceu-tical market inhibit incentives to innovate (Santiago, 2010);management strategies aim to increase productivity andefficiency. In other words, adoption of modern organiza-tional practices may simply contribute to the making ofwhat Cimoli (2000) identifies as “global modern manufac-turing centre”.

Acknowledgements

Earlier versions of this paper benefited greatly fromcomments by two anonymous referees and the editors tothis special issue of SCED; as well as Gabriela Dutrénit;Nobuya Haraguchi; Robin Cowan; Wilfred Dolfsma, BrankaUrem, Jojo Jacob and other members of the research groupon Innovation, global business strategies and host countrydevelopment at UNU-MERIT; Leonel Corona, Javier Jassoand staff of the División de Investigación, Facultad deContaduría, Administración e Informática of the NationalAutonomous University, Mexico. Suggestions by MartinShrolec and other participants at the 7th Annual conferenceof the Globelics network in Dakar, Senegal are appreciated.Fernando Santiago is greatly indebted to the following atINEGI for granting access to the data used in this paper:Gerardo Leyva and Abigail Durán. Special thanks to AdrianaRamírez, Gabriel Romero and Cándido Aguilar also at INEGIfor help in the processing of the data. Maria Fermie helpedin editing and correcting previous versions. Substantial partof the work was carried out while Fernando Santiago was avisiting researcher at the Research and Statistics Branch,UNIDO, and a PhD researcher at UNU-MERIT. Omissionsand errors remain the responsibility of the authors. Theideas expressed in this paper do not reflect those of theorganisations hosting the authors.

Appendix A.

See Table A1.

Appendix B.

See Table B1.

References

Amabile, T., 1997. Motivating creativity in organizations: on doing whatyou love and loving what you do. California Management Review 40,39–58.


Arundel, A., Lorenz, E., Lundvall, B.-A., Valeyre, A., 2007. How Europe’seconomies learn: a comparison of work organization and inno-vation 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 inter-national human resource development among indigenous firms andmultinational affiliates in East and Southeast Asia. Human ResourceDevelopment 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 inHRM and performance research. Human Resource Management Jour-nal 15, 67–94.

Cardinal, L., Hatfield, D., 2000. Internal knowledge generation: theresearch laboratory and innovative productivity in the pharmaceu-tical industry. Journal of Engineering and Technology Management17, 247–271.

Casas, R., 2005. Exchange and knowledge flows between large firms andresearch institutions. Innovation: Management, Policy and Practice 7,188–199.

Caves, R., Hurwitz, M., Whinston, M., 1991. Patent expiration, entry andcompetition in the US pharmaceutical industry. In: Brookings Papersin Economic Activity, pp. 1–48.

Christofides, L., Hardin, Stengos, T., 2000. Corrigendum to “On the calcu-lation of marginal effects in the bivariate probit model”. EconomicsLetters 68, 339.

Christofides, L., Stengos, T., Swidinsky, R., 1997. On the calculation ofmarginal effects in the bivariate probit model. Economics Letters 54,203–208.

Chung-Jen, C., Jing-Wen, H., 2009. Strategic human resource practices andinnovation performance – the mediating role of knowledge manage-ment capacity. Journal of Business Research 62, 104–114.

Cimoli, M. (Ed.), 2000. Developing Innovation Systems: Mexico in a globalcontext. 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. Journalof Health Economics 20 (6), 1033–1057.

Cohen, W., Levinthal, D., 1989. Innovation and learning: the two faces ofR and D. The Economic Journal 99, 569–596.

Cohen, W., Levinthal, D., 1990. Absorptive capacity: a new perspectiveon learning and innovation. Administrative Science Quarterly 35,128–152.

Combs, J., Liu, Y., Hall, A., Ketchen, D., 2006. How much do high-performance work practices matter? A meta-analysis of theireffects 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: thecase of pharmaceutical biotechnology. Research Policy 35, 1–23.

Grant, R., 1996. Prospering in dynamically-competitive environments:organizational capability as knowledge integration. Organization Sci-ence 7, 375–387.

Gray, D., Cundell, S., Hay, D., O’Neill, J., 2004. Learning Through the Work-place: 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 farma-

céutica en México, in Industria Farmacéutica y Propiedad Intelectual:Los países en desarrollo. In: Guzmán, A., Viniegra, G. (Eds.), H. Cámarade Diputados LIX Legislatura, UAM-Iztapalapa, M.A. Porrúa. México,pp. 267–322.

Hara, T., 2003. Innovation in the Pharmaceutical Industry: The Process ofDrug Discovery and Development. Edward Elgar.

Hemmert, M., 1998. Reorganization of R and D in Japanese manufactur-ing firms: preserving competitiveness for the twenty-first century.In: Hemmert, M., Oberländer, C. (Eds.), Technology and Innovation inJapan: Policy and Management for the Twenty-First Century. Rout-ledge, pp. 129–150 (Chapter 8).

Hollis, A., 2002. The importance of being first: evidence from Canadiangeneric pharmaceuticals. Health Economics 11, 723–734.

Johnson, J., Baldwin, J., Diverty, B., 1996. The implications of innovationfor human resources strategy. Futures 28, 103–119.

5 ge and E

K

K

K

K

K

L

L

L

L

L

L

L

L

L

L

L

M

M

M

M


ale, D., Little, L., 2007. From imitation to innovation: the evolution ofR&D capabilities and learning processes in the Indian pharmaceuticalindustry. Technology Analysis & Strategic Management 19, 589–609.

aplinsky, R., 1995. Technique and system: the spread of Japanese man-agement techniques to developing countries. World Development 23,57–71.

ettler, H., Modi, R., 2001. Building local research and development capac-ity for the prevention and cure of neglected diseases: the case of India.Bulletin of the World Health Organization 79, 742–747.

im, L., 1997. Imitation to Innovation: The Dynamics of Korea’s Techno-logical Learning. Harvard University Press, Boston.

im, Y., Kim, L.S., Lee, J., 1989. Innovation strategy of local pharmaceuti-cal firms in Korea: a multivariate analysis. Technology Analysis andStrategic Management 1 (1), 29–44.

abarca, G., 1999. Formación para el Trabajo: Observaciones en AméricaLatina y El Caribe. In: Guillermo Labarca (Ed.), Formación y empresa,CINTERFOR/OIT/CEPAL/GTZ, pp. 19–58 (Chapter 1).

aursen, K., Mahnke, V., 2001. Knowledge strategies, firm types, and com-plementarity in human-resource practices. Journal of Managementand Governance 5, 1–27.

aursen, K., 2002. The importance of sectoral differences in the applica-tion of complementary HRM practices for innovation performance.International Journal of the Economics of Business 9, 139–156.

aursen, K., Foss, N., 2003. New human resource management practices,complementarities and the impact on innovation performance. Cam-bridge Journal of Economics 27, 243–263.

i, Y., Vanhaverbeke, W., Schoemakers, W., 2008. Exploration and exploita-tion in innovation: reframing the interpretation. Creativity andInnovation Management 17, 107–126.

ong, S., Freese, J., 2006. Regression Models for Categorical DependentVariables Using STATA. STATA Press.

orenz, E., Valeyre, A., 2006. Organizational forms and innovative per-formance: a comparison of the EU-15. In: Lorenz, E., Lundvall, B.-A.(Eds.), How Europe’s Economies Learn: Coordinating Competing Mod-els. Oxford University Press, pp. 140–160 (Chapter 6).

orenz, E., Wilkinson, F., 2003. Forum: Organizational change, humanresource management and innovative performance: comparative per-spectives. Cambridge Journal of Economics 27, 239–241.

und, R., 2004a. Some Danish experiences related to the organization ofnew product development. In: Christensen, J., Lundvall, B.-A. (Eds.),Research on Technological Innovation, Management and Policy, 8.Elsevier, Amsterdam, pp. 229–254.

und, R., 2004b. The organization of actors’ learning in connection withnew product development. In: Christensen, J., Lundvall, B.-A. (Eds.),Research on Technological Innovation, Management and Policy, 8.Elsevier, Amsterdam, pp. 129–153.

undvall, B.-A., Johnson, B., Andersen, E., Dalum, B., 2002. National systemsof production, innovation and competence building. Research Policy31, 213–231.

arch, J., 1991. Exploration and exploitation in organizational learning.Organization Science 2, 71–87.

ichie, J., Sheehan, M., 1999. HRM practices, R&D expenditure andinnovation investment: evidence from the UK’s 1990 WorkplaceIndustrial Relations Survey (WIRS). Industrial and Corporate Change 8,211–234.

ichie, J., Sheehan, M., 2003. Labour market deregulation, ‘flexibility’ andinnovation. Cambridge Journal of Economics 27, 123–143.

umford, M., 2000. Managing creative people: strategies and tac-tics 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 diffu-sion of technological knowledge: time for a change? In: Hemmert, M.,Oberländer, C. (Eds.), Technology and Innovation in Japan: Policy andManagement for the Twenty-First Century. Routledge, pp. 115–128(Chapter 7).

OECD, 1998. The OECD Jobs Strategy: Technology, Productivity and JobCreation: Best Policy Practices. OECD, Paris.

Quinn, M., Rubb, S., 2006. Mexico’s labor market: the importance ofeducation–occupation matching on wages and productivity in devel-oping countries. Economics of Education 25, 147–156.

Rothaermel, F., Deeds, D., 2004. Exploration and exploitation alliancesin biotechnology: a system of new product development. StrategicManagement Journal 25, 201–222.

Samstad, J., Pipkin, S., 2005. Bringing the firm back in: Local decision mak-ing and human capital development in Mexico’s maquiladora sector.World Development 33, 805–822.

Santiago, F., 2010. Human resource management practices and learn-ing for innovation in developing countries: pharmaceutical firms inMexico. Doctoral Dissertation, UNU-MERIT/University of Maastricht,The Netherlands.

Seiter, A., 2005. Pharmaceuticals: local manufacturing. Health Nutritionand Population, vol. 5, The World Bank.

Sternberg, R., O’Hara, L., Lubbart, T., 1997. Creativity as investment. Cali-fornia Management Review 40, 8–21.

Tello, M.P., Greene, W., 1996. US managerial strategies and applicationsfor retaining personnel in Mexico. International Journal of Man Power17, 54–94.

Terziovski, M., Morgan, J., 2006. Management practices and strategies toaccelerate the innovation cycle in the biotechnology industry. Tech-novation 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 andD internationalization: four different patterns of managing researchand development. Research Policy 31, 569–588.

White, F., 2002. Capacity-building for health research in developing coun-tries: a manager’s approach. Pan American Journal of Public Health 12,165–172.

Wright, P., Dunford, B., Scott, S., 2001. Human resources and the resourcebased view of the firm. Journal of Management 27, 701–721.

Zúniga, M., Guzmán, A., Brown, F., 2007. Technology acquisition strategiesin the pharmaceutical industry in Mexico. Comparative TechnologyTransfer and Society 5, 274–296.

Websites

Bureau of Labor Statistics, US Department of Labor; Career Guide to Indus-tries, 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 Pharmaceu-tical Manufacturing: Pharmaceutical CGMPs for the 21st Century.US Food and Drug Administration; http://www.fda.gov/Drugs/
DevelopmentApprovalProcess/Manufacturing/QuestionsandAnswersonCurrentGoodManufacturingPracticescGMPforDrugs/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.

http://www.bls.gov/oco/cg/cgs009.htm

http://www.fda.gov/Drugs/DevelopmentApprovalProcess/Manufacturing/QuestionsandAnswersonCurrentGoodManufacturingPracticescGMPforDrugs/default.htm

http://www.fda.gov/Drugs/DevelopmentApprovalProcess/Manufacturing/QuestionsandAnswersonCurrentGoodManufacturingPracticescGMPforDrugs/default.htm

http://www.greatplacetowork.com.mx/best/list-mx.htm

http://www.greatplacetowork.com.mx/best/list-mx.htm

human resource management for learning through knowledge exploitation and knowledge exploration:...

Documents