Choosing among alternative technological strategies:

an empirical analysis of formal sources of

innovation

Pilar Beneito

Departamento de Análisis Económico

(Universitat de València -SPAIN-)

IV ENCUENTRO DE ECONOMIA APLICADA

REUS JUNIO 2001

Abstract

This work aims to fulfil the lack of a detailed conceptual and empirical analysis of the interfirm differences in deciding the composition of their technological efforts. Using data for Spanish firms in the period 1990-96, the study departs from the standard analysis of the determinants of innovative investment, follows with the less analysed question of the determinants of the generate versus import alternative and ends with a completely novel analysis of the firms’ characteristics that lead them to organise research internally as compared to the possibility of contracting R&D services externally. Contrasting with the standard practice, the econometric approach takes into account the existence of non-linearities in the used data.

Key words: R&D; technology imports; discrete choice.

JEL classification: C25; D21; O33;
1 Introduction

Despite widespread acknowledgement of the complexity of the process of innovation, few researchers have taken into account this fact when analysing the determinants of the research endeavours of firms. On the contrary, most previous theoretical and empirical studies of the microeconomic determinants of R&D have treated R&D spending as an homogeneous activity. However, firms may choose among different alternatives in their search for technical advance and, moreover, each firm may combine these alternatives differently to shape its particular technological strategy.

The extent to which the understanding of the factors that underlie technical advancement may help policy makers when designing the research policy depends greatly upon two elements: the relevance of the aspect under scrutiny and the level of disaggregation in the analysis. The present work aims to study the allocative determinants of the formal ways of technical acquisition, which are expected to be the most important means for technical advance. In addition, the analysis distinguishes, in a sequential fashion, three kinds of technology-related decisions: first, the decision to invest in some formal source of innovation (in the present work, R&D and/or technology imports); second, the decision to generate innovations through R&D as compared to the alternative of their acquisition through licensing; and, finally, the reasons for the intrafirm location of R&D activities versus contracting them from outside.

The value added of the work that will be presented in the next pages increases following the above described sequence. The first of the decisions to be studied, the analysis of the determinants of the firms’ decision to invest in innovation, (usually defined in terms of R&D expenditures), is by now a profusely studied topic in industrial organisation, both theoretically and empirically. Among the most well known studies in this area are those of Levin et al. (1985), Cohen et al. (1987), Cohen and Levin (1989) or Cohen (1995). In the case of Spain, some recent studies are those of Paricio (1993), Gumbau (1997) and González and Jaumandreu (1998). The present work follows closely the standard analysis used in the studies to date.

Secondly, the existing empirical studies related to the imports of technology have been more frequently orientated towards the verification of a complementary or substitutive relationship with the own R&D efforts. Some examples are the works of François (1985), Braga and Willmore (1991), Mohnen and Lépine (1991), Siddharthan (1992), Lee (1996) and Katrak (1997). Recently, González and Rodríguez Romero (1999) have analysed such a relationship for the case of Spain. The approach that will be adopted in the present work differs clearly from those mentioned above. This work provides both a conceptual and empirical examination of the factors that lead firms to import technology from abroad considering that these imports are a purchase of technology. Thus the firm will consider this option after taking into consideration the alternative choice of generating own innovations through R&D. The discussion of the determinant factors to be considered are specifically thought in relation to the elements under analysis (the purchase of technology versus the inducement, or generation, of own innovations), instead of being a copy of the set of variables widely used to analyse R&D as a whole. To my knowledge, the only work published that defines the problem in similar terms is that of Link et al. (1983), who analyse the propensity of firms to purchase or induce technical knowledge within the context of the Utterback-Abernathy production process lifecycle. The present work provides a wider view of the problem, where the lifecycle model variables are only a part of the overall analysis.

Finally, the analysis of the relationship between intrafirm and contractual forms of industrial research is practically non existent in the empirical literature on R&D. One exception is Mowery’s (1983) historical case study which, using archival data for three major independent research organisations, constructs a list of client firms of the early years in American manufacturing, (1900-1940), and tests complementarity between internal and external R&D (in his case this amounts to observing the existence of an in-house research facility within a client firm). The author also describes the types of research performed on a contractual basis and studies the relationship between firm size and the integration of the research function into the firm. However, a systematic analysis of the determinants of the internal location of R&D as compared to contracting them from outside remains to be done. The present work aims to fill this gap.

Using recently available data of Spanish manufacturing firms in the period 1990-96, the present work analyses how firms employ different technological strategies, first choosing whether to invest or not in some formal source of innovation, then deciding whether to generate or acquire existing technology and, finally, determining whether to internalise R&D activities or contract them from outside. For the first time, the decisions involved in the sequence above are analysed using the same data source. The econometric technique used is also somewhat novel in this setting, for it will consider the existence of non-linearities in the data, as the corresponding section below will explain.

This study may also be understood as an analysis of the composition of the firms’ investment in technology, where the amount of payments for importing technology, internal R&D expenditures and external payments for R&D services would be the three components of the total amount of resources devoted to the achievement of technical advance.

The first section below discusses the conceptual basis for the subsequent empirical analysis. The next section presents the data and the econometric model to be estimated. Then, the estimation results are discussed, and, to end, the main conclusions and implications of the study are stated.

2 Technological strategies: conceptual basis for the empirical analysis

The prime decision firms undertake in relation to their technological orientation is whether to invest or not in some formal source of innovation. Once this decision has been made, and provided the firm decides to allocate funds to some form of formal technological activities, it still may decide whether to acquire others’ technology directly or whether to make efforts orientated towards the generation of its own technology. Finally, those firms who opt for generating technology, may organise internally their R&D activities or may contract R&D activities from other firms or institutions.

I will call the prime decision in the sequence above the yes/no decision. The second will be referred to as the generate/import decision because the acquisition of technology in this work will be considered to be reflected in the purchase of foreign technology through licensing. This will be so both because technology imports constitute one of the main channels of technological acquisition and because the data available do not allow us to consider other alternatives[1]. The choice between ‘building’ an internal R&D lab and contracting external R&D services will be called the in-house/contract decision.

Interfirm differences in their innovation strategies may be viewed as interfirm differences in the perceived net benefits from their election. Thus, the determinant factors of each choice refer to the evaluation, at the margin, of the specific costs and benefits of each alternative. This will be the starting point of the conceptual analysis of the factors expected to affect each pair of decisions.

2.1 The yes/no decision

I will not elaborate here on the discussion and justification of the determinants of the yes/no decision. As already stated in the introduction, there exists at present a huge body of both empirical and theoretical literature on the determinants of R&D expenditures, which in most of the cases entails a previous discussion of the factors behind the yes/no type response. Although the main interest of the present work does not lie in the analysis of this decision, the empirical analysis of this binary choice is convenient here since it provides a comparative substratum for the subsequent analysis.

Following the existing literature on this topic, I will use those more standard factors that are considered to influence the decision to enrol in R&D activities. These factors are: the economic opportunity (approximated by the (log of the) amount of sales, LSALES, and the size of the market, MARKET, where the firm evolves); the technological opportunity, which may depend not only on the characteristics of the firm’s industrial sector, OPORT, but also on the kind of technology the firm bases its production processes on, ADVC; the firm size, SIZE; the price-cost margin, PCM, which tells us about the financial capabilities of the firm; the appropriability conditions, APROP; the firm’s degree of foreign openness, approximated by the propensity to export, EXPT; the foreign content of the firm’s physical capital, FEQ; the degree of differentiation of the offered product, DIFR; the diversification of the firm’s line of business, DIVERS; whether the firm accounts for a significant market share, SHARE; and the four-firm concentration ratio of the main market where the firm operates, CR-4. The precise definitions and measures of all these variables are presented in the appendix at the end.

2.2 The generate/import decision.

The main feature that distinguishes the generate from the import decision is that the former implies the uncertainty about the innovation results to be obtained whereas the latter means a direct purchase of technology. Moreover, the generation makes the firm owner of its technology whereas the import alternative implies dependence on the technology of others. The costs and benefits of each alternative are closely related to these premises and they will be evaluated by the firm in relative terms to each other.

The costs of the generate alternative are mainly reflected in its direct costs and in the uncertainty regarding the innovation results. As for the direct costs, the firm will compare them with the purchase price of acquired technology. Generally, the generation of technology through R&D will entail some fixed costs that a supplier of technology could spread among all purchasers. This means that the direct costs of generation are greater, in general, than the price of an imported technology.

The likely existence of such fixed costs and the necessary maturity period for a research project to generate innovation results, make the R&D activities show, to some degree, a permanent character that a particular technology purchase would not involve. To account for this relative stability of the R&D activities, the variable STAB, as described in the appendix, has been constructed. This variable has then to be understood as a proxy for the existence of fixed costs associated with the own generation of technology.

The second important kind of costs the generating firm has to face is, as already mentioned, the uncertainty regarding the innovation results. The import alternative is not affected by such a risk. The firm’s willingness to support these risks and the technical facilities for successful generated innovations will positively affect the generate alternative. The firm’s attitude toward risk-taking is thought to be related to the firm’s form of ownership. The more diversified is the ownership of capital, the more risk averse the firm is (see the definition of the OWNR variable in the appendix). Regarding the chances of successful innovation results, they depend on the technological richness of the firm environment (OPORT) and on the nature of the technical basis of the firm’s production processes (ADVC). Moreover, firms may form an expectation of innovation success derived from the observed (or perceived) correlation between other’s R&D activities and the introduced innovations in their closest neibourhood (SUCC).

Albeit, as it would seem to follow from the above stated arguments, the costs of the generation alternative exceed those of the import one, the former entails unquestionable rewards as compared to the latter. Firstly, the output resulting from the own innovation activity is more likely to fit the particular technological requirements of the firm. On the contrary, technology for which an external market exists, is generally much more standardised since it is conceived to match the generalised needs of the potential consumers. Thus, the costs of adaptation have to be added to the direct price of the imported technology[2]. The capabilities of the recipient firm to make such a necessary adaptation play an important role in this setting. In accordance with Bartel and Lichtenberg (1987) and Huffman and Mercier (1991), the portion of skilled workers in the firm is an important measure of such capabilities (see the variable SKILL in the appendix).

Secondly, the generation of own innovations avoids the dependency inherent in the import alternative. This enhances the long-term competitive stance of the generating firm. The relative value each firm places on this aspect will be higher the more interested the firm is in developing an strategy of technical leadership in its market. Within the typology of innovation strategies defined by Freeman (1974), this is called an ‘offensive’ strategy. Firms with an offensive strategy are those highly intensive in R&D, with willingness to assume high risks, with important financial capabilities and which operate in high technology environments. The dummy variable OFFENS, as defined in the appendix, incorporates all this information.

The weight the firm places on the costs of each alternative is lesser the higher the financial capabilities of the firm. The firm’s price-cost margin, PCM, is used in estimation as proxy of them. As stated above, the generate alternative is likely to be more costly than the import option. Thus, the expected effect of the price-cost margin is to increase the probability of a firm choosing to generate its own innovations.

Firms will also perceive differently the costs and benefits derived from each alternative depending both on their own characteristics and on the different configuration of the manufacturing environment in which each of them evolves. Factors such as size, (SIZE), market size (MARKET) and the appropriability conditions, (APROP), are likely to affect the decision. Market size and the appropriability conditions are expected to show positive signs as they positively affect to the firm’s possibilities to exploit the rents derived from the innovation results. It is somewhat more difficult to establish a priori the expected sign of the firm size. Larger firms may show both greater capabilities to face the relative higher cost of the generate alternative and higher possibilities to exploit the results of the innovation process but, on the contrary, other factors such as the inherent bureaucracy of large firms, may limit the organisational autonomy thus lessening the incentive to generate their own innovations.

Finally, the relationship between generation-acquisition of technology may be explained by the product-process lifecycle model of Utterback and Abernathy (1975). This framework may serve as explanatory only in the case of process innovation. According to this model, in the initial phases of the product-process lifecycle, the small production volumes, the rapid and frequent product and process changes derived from the low standardisation of the product, and thus, the high uncertainty regarding the product evolution, make firms reluctant to invest R&D funds into process innovation and to prefer the direct acquisition of technology. In the transition phase, the product standardisation augments, production volumes increase demanding new process technologies, and the price competition is intensified. In this phase, the technological uncertainty is reduced and firms place greater importance on process development to meet cost competition. It is at this point that the own generation of technology is at its peak. Beyond this point, the high product standardisation, the emergence of a dominant process design and the integrated automated process technology, lead to the emergence of a supplier industry, which reduces relatively the acquisition costs. In addition, during this phase, the needed process changes are major and costly and the uncertainty regarding the potential results and their nature is then increased. Again, the purchase of technology is preferred against the own generation.