Assessing measurement errors in the CDM research–innovation–productivity relationships

ABSTRACT

The Crépon-Duguet-Mairesse 1998 article, known as CDM, initiated a structural econometric framework to analyze the relationships among research, innovation and productivity, which has been estimated most generally on the basis of cross-sectional innovation survey-type data. Some econometric implementations of the CDM approach have suggested that such data give useful but imprecise measures of the innovation output (share of innovative sales), and to a lesser degree of the innovation input (R&D). These ‘measurement errors’ may result in attenuation biases of the estimated R&D and innovation impact elasticities in the two basic CDM ‘roots’ relations of R&D to innovation and innovation to productivity, as well as in the extended production function à la Griliches linking directly R&D to productivity. Using a panel of three waves of the French Community Innovation Survey (CIS), we assess these biases and the magnitude of the underlying measurement errors, assuming mainly that they are ‘white noise’ errors. We do so by comparing two pairs of usual panel estimators (Total and Between) in both the cross-sectional and time dimensions of the data (Levels and Differences). We find large measurement errors on innovation output in the innovation–productivity equation, resulting in large attenuation biases in the related elasticity parameter. We also find smaller but sizeable measurement errors on R&D, with significant attenuation biases in the corresponding elasticity estimates, in the R&D–innovation equation and the extended production function. Simulations suggest that the measurement errors on innovation and R&D are unaffected by similar measurement errors on the capital variable.     

Knowledge appropriability, firm size, and growth

Empirical evidence shows that the number of patents per R&D dollar declines with firm size. In this paper, we propose a Schumpeterian growth model that accounts for this evidence. We analyze an economy with firms that engage in cost-reducing innovation resulting from the accumulation of both codified and tacit knowledge: the former occurs through the purchase of patents, while the latter is the result of R&D conducted in-house by firms. We study the relation between knowledge appropriability and market structure, and we show that a shift from patents to in-house research occurs as firm size gets larger. Since innovation statistics concentrate mainly on patents, this process of research reallocation results into an under-estimation of innovative activity and is responsible for the declining ratio of patents to R&D expenditure. Survey data on UK-based firms provide support to our results.     

SCIENCE-BASED R&D IN SCHUMPETERIAN GROWTH

Firm success is often associated with the development of better products. Private firms undertake applied R&D seeking market advantage, by capitalizing on the freely accessible results of basic research. But unpatentable basic research often fails to address applied R&D open problems. What is the role of the incentives in improving the innovative performance of an economy by matching partially motivated public researchers to their mission? Sometimes government-funded research projects are mission-directed, and yet in many cases the public sector academics indulge in carrier-driven research. An innovation system where, as in the United States, basic research is also driven by patents implicitly sets an ex-post incentive to the researchers guided by invisible hand. For a public innovation system – like the European one – designing an incentive scheme to motivate public researchers is of key importance for fostering the performance of the economic system. This paper extends the Schumpeterian multisector growth model with vertical innovation by highlighting a link between the degree of 'targetness' of public research and aggregate innovation. A positive effect of social capital is also proved.     

R&D spillovers and productivity: Evidence from U.S. manufacturing microdata

This paper deals with the estimation of the impact of technology spillovers on productivity at the firm level. Panel data for American manufacturing firms on sales, physical capital inputs, employment and R&D investments are linked to R&D data by industry. The latter data are used to construct four different sets of `indirect' R&D stocks, representing technology obtained through spillovers. The differences between two distinct kinds of spillovers are stressed. Cointegration analysis is introduced into production function estimation. Spillovers are found to have significant positive effects on productivity, although their magnitudes differ between high-tech, medium-tech and low-tech firms. First version received: April 1997/final version received: April 1999     

Firms’ human capital, R&D and innovation: a study on French firms

This article investigates the effects of human capital and technological capital on innovation. While the role of technological capital as measured by research and development (R&D) expenditure has been intensively investigated, few studies have been made on the effect of employee training on innovation. This article explores the relationship between innovation and firm employee training. Our methodological approach contributes to the literature in three ways. We propose various indicators of firm employee training. We build a count data panel with a long time-data series to deal with the issue of firms?? heterogeneity. We propose a dynamic analysis. Using dynamic count data models on French industrial firms over the period 1986?C1992, we find positive and significant effects of R&D intensity and training on patenting activity. Whatever the indicators of training our results show that the firm employee training has a positive impact on technological innovation.     

INNOVATION ACTIVITIES EXPLAINED BY FIRM ATTRIBUTES AND LOCATION

In this paper, innovation activities of a firm are observed as its R&D spending and participation in three categories of innovation systems. The various factors that can influence a firm's innovation efforts are divided into, (i) firm location reflecting the regional milieu and (ii) firm attributes such as corporate structure, nature of the knowledge production, type of industry and a set of specific firm characteristics. The study is based on information about 2094 individual firms, which may be non-affiliated or belong to a group (multi-firm enterprise). The empirical analysis applies a novel data set to examine the influence of location versus a vector of firm attributes. Among innovative firms, the location of a firm does not influence neither the R&D intensity nor the frequency of interaction in horizontal and vertical innovation systems, when controlling the skill composition, physical capital intensity, industry, firm size and market extension. The paper contributes to the literature by observing that innovative firms have similar characteristics irrespective of where they are located, although the share of innovative firms differs between regions.     

Why adopt social enterprise software? Impacts and benefits

This paper explores the performance impacts and benefits of the adoption of Social Enterprise Software (SES). SES forms a nested innovation, given that its adoption requires an already established infrastructure of Information and Communication Technology. To control for induced sample selection, we use a two-step estimation procedure. Based on German firm-level data our results confirm that firms which use business-to-business (B2B) e-commerce applications are more likely to adopt SES. The estimated correlations also provide weak evidence for complementarity between B2B e-commerce and SES. We show that two measures of firm performance, i.e. sales and labor productivity, are highest for firms using SES and B2B e-commerce applications in conjunction.     

Modes of international activities and the innovativeness of firms: an empirical analysis based on the Japanese National Innovation Survey for 2009

We quantitatively examine the factors which account for differences in innovation output depending on the mode of international activities, employing the innovation accounting framework proposed by Mairesse and Mohnen [2001. To be or not to be innovative: An exercise in measurement. NBER Working Paper No. 8644. Cambridge, MA: National Bureau of Economic Research; 2002. “Accounting for Innovation and Measuring Innovativeness: An Illustrative Framework and an Application.” American Economic Review 92 (2): 226–230]. We find that internationally engaged firms use more innovation inputs and generate more innovation outputs. Firms with R&D establishments abroad show the best innovation performance. A significant part of the higher innovation performance of internationally engaged firms can be explained by their greater intra-firm knowledge spillovers, R&D intensity, perceived competitive pressure, and proximity to basic research. However, more importantly, our innovation efficiency analysis suggests that engagement in international activities increases the sales amount of innovative products though it does not necessarily raise the probability that a firm successfully develops a new product or process.     

Technological convergence, R&D, trade and productivity growth

This paper analyses productivity growth in a panel of 14 United Kingdom manufacturing industries since 1970. Innovation and technology transfer provide two potential sources of productivity growth for a country behind the technological frontier. We examine the roles played by research and development (R&D), international trade, and human capital in stimulating each source of productivity growth. Technology transfer is statistically significant and quantitatively important. While R&D raises rates of innovation, international trade enhances the speed of technology transfer. Human capital primarily affects output through private rates of return (captured in our index of labour quality) rather than measured TFP.     

Link between innovation and productivity in Canadian manufacturing industries

Empirical studies commonly use research and development (R&D) to measure innovation and often find, especially in Canada, no strong link between productivity and innovation. In this article, we model innovation as an unobservable latent variable that underlies four indicators: R&D, patents, technology adoption, and skills. We find that these indicators are reasonably good measures of innovation for aggregate manufacturing. However, except for skills, the reliability of the indicators for innovation differs among individual industries. Our innovation indexes, based on the latent variable model, show that most manufacturing industries became more innovative over the 1980–1997 period. The pace of innovation in the electrical and electronic products industry accelerated during the 1990s. In addition, we show that the new measure of innovation has a positive and statistically significant impact on productivity. It takes from 1 to 3 years, depending on the industry, for innovation to generate an impact on productivity.     

PUBLIC R&D AND INDUSTRIAL INNOVATIONS AT THE PROJECT LEVELS: AN EXPLORATION OF TAIWAN'S PUBLIC RESEARCH PROJECTS

This study investigates the role of the ITRI in Taiwan's technological catch-up. The authors examine the relationship between public R&D and industrial innovations in Taiwan using data encompassing 252 ITRI annual research projects and the survey on the characteristics of 5902 cases of transferred technologies within these projects. The authors develop a new index of innovative output to measure the monetary value of patents for research projects. They find that the influences of accumulated R&D stock, high-level R&D personnel, and the intensity of process innovations on project-level R&D productivity to be more pronounced when the monetary value of patents, instead of simple patent counts, is used as the proxy for innovation outputs. (JEL O12 , L63 )     

Firm Size Distribution and Growth*

Abstract Empirical documentation of the sectoral distribution of firm size for a set of European countries reveals substantial differences. We study the relationship between productivity growth at the industry level and size structure. A positive and robust relation is found between average firm size and growth. We ask why size should matter for growth by considering the role of innovation to construct a test based on the differential effect of size on growth according to various indicators of R&D intensity. Our results indicate that larger size fosters productivity growth because it allows firms to take advantage of all the increasing returns associated with R&D. We argue that our test can be interpreted as a test of reverse causality, which lends support to the view that firm size has a causal impact on growth.     

Evidence on the impact of R&D and ICT investments on innovation and productivity in Italian firms

Both research and development (R&D) and information and communication technology (ICT) investment have been identified as sources of relative innovation underperformance in Europe vis-à-vis the USA. In this article, we investigate the R&D and ICT investment at the firm level in an effort to assess their relative importance and to what extent they are complements or substitutes. We use data on a large unbalanced panel data sample of Italian manufacturing firms constructed from four consecutive waves of a survey of manufacturing firms, to estimate a version of the CDM model of R&D, innovation, and productivity [Crépon–Duguet–Mairesse 1998. Research, innovation and productivity: An econometric analysis at the firm level. Economics of Innovation and New Technology 7, no. 2: 115–58] that has been modified to include ICT investment and R&D as the two main inputs into innovation and productivity. We find that R&D and ICT are both strongly associated with innovation and productivity, with R&D being more important for innovation, and ICT investment being more important for productivity. For the median firm, rates of return to both investments are so high that they suggest considerably underinvestment in both these activities. We explore the possible complementarity between R&D and ICT in innovation and production, but find none, although we do find complementarity between R&D and worker skill in innovation.     

Structural Change And Innovation In U.S. Telecommunications

This paper reexamines broadly. from the standpoint of innovation, the arguments for vertical integration in the U.S. telecommunications industry in light of structural change since the breakup of the Bell System. While basic and applied research became the casualty of the 1984 breakup and the 1995 AT&T split, there is no evidence that the pace of innovative activity and productivity has slowed. Evidence from R&D and patent data suggests some acceleration of innovative activity. However. the service segment of the industry ceased to be the center of technological innovation. The source of future innovation seems to lie in the telecommunications and Internet equipment firms and independent software firms. The emergence of the competitive stand-alone software industry, combined with a trend towards open operating systems and customer demand for greater flexibility, and growing substitution of technology alliances for in-house R&D appear to have undermined the case for vertical integration in the telecommunications industry. From the standpoint of business strategy, the question of whether a firm like AT&T, notwithstanding its huge investments in cable facilities, can develop distinctive and sustainable capabilities through horizontal expansion and ubiquity and one-stopshopping marketing alone remains open.     

NON‐SCALE EFFECTS OF INTERNATIONAL TECHNOLOGICAL‐KNOWLEDGE DIFFUSION ON SOUTHERN GROWTH AND WAGES

We develop a dynamic, general equilibrium non‐scale endogenous growth model of North–South technological‐knowledge diffusion by imitation. Countries differ in levels of exogenous productivity, human‐capital levels and R&D capacity. Growth is driven by Northern innovative R&D and the South converges towards the North. Growth is also driven by human‐capital accumulation, scale effects are removed, imitation is only feasible once a threshold distance to the frontier has been attained and is dependent on the South's relative level of employed human capital and on domestic policies promoting R&D. Imitation promotes partial convergence of inter‐country wages and governs the path of intra‐South wage inequality.     

Proximity,networking and knowledge production in Europe: What lessons for innovation policy?

This paper aims at investigating the role of different types of proximity on the technological activity of a region within the context of a knowledge production function, where R&D expenditure and human capital are the main internal inputs. We thus assess to what extent the creation of new ideas in a certain region is enhanced by knowledge flows coming from proximate regions. In particular, we examine in detail different kinds of proximity by combining the usual geographical dimension with the institutional, technological, social and organizational proximity. The analysis is implemented for an ample dataset referring to 276 regions in 29 European countries (EU27 plus Norway, Switzerland) over the last decade. Results show that human capital and R&D are clearly essential for innovative activity with the former being much more effective in driving the production of knowledge. As for the proximity and network effects, we find that technological proximity outperforms the geographic one, whilst a limited role is played by social and organizational networks. As a result, the first policy message is that European regions still need to focus on policies aimed at increasing the endowments of well-educated labour force and therefore their knowledge base. Furthermore, we need innovation policies based on each region's specific innovation potential, due to the existing differences in geographical, cognitive, institutional, social and organizational structures and networks.     

Measuring and Explaining Innovative Capability: Evidence from Southeast Asia

What causes firm‐level product innovation in developing economies? This paper answers this question by emphasizing the role of process improvements that are influential in product innovation. We construct a firm‐level innovative capability score using novel, broad‐based, but detailed data on various process improvement practices obtained from firms in Southeast Asia. We then investigate the factors that may affect innovative capability. We also estimate the effect of the innovative capability score on product innovation controlling for research and development intensity and other firm characteristics. Our empirical investigation identifies a chief executive officer (CEO)'s past experience at a foreign or large firm, and buyer pressure to adopt international standards as key determinants of innovative capability. Novel and unique findings from our examination include: (i) the impact of a CEO's past experience at a foreign or large firm on the innovative capability is larger for local enterprises and small‐ and medium‐sized enterprises; (ii) the impact of share of foreign workers in the upper managerial levels only appears to be significant in foreign firms and larger firms; and (iii) buyer pressure is a more likely contributor to innovative capability compared with capital tie‐ups with multinational enterprises or joint venture buyers that capture vertical technology transfers. Finally, our empirical results show that a firm in Southeast Asia is more likely to achieve product innovation if the firm has had a higher innovative capability.     

R&D spillovers and employment: evidence from European patent data

In this paper we investigate the role of patents in the relationship between R&D activity, spillovers and employment at the firm level. A reduced-form labour demand equation is estimated. Our analysis is based upon a dataset consisting of 879 R&D-intensive manufacturing firms worldwide for which information was collected for the period 2002–2010. We use data from all EU R&D investment scoreboard editions issued every year until 2011 by the JRC-IPTS (scoreboards). Since the innovation output of the industrial strategy of every firm is the number of patents, the main contribution to the existing literature is to investigate also the impact of patents/R&D ratio and patents/spillovers ratio on employment level. The empirical results suggest a significant impact of R&D spillover effects on company employment although the results differ substantially according to the spillover stock, which may considerably affect policy implications.     

Corporate tax policy and heterogeneous firm innovation: Evidence from a developing country

This paper adds to the literature by identifying the causality of corporate tax policy on firm innovation in a developing country. We exploit the China’s 2006 corporate income tax base reform to integrate the tax system between foreign-invested and state/collective-controlled firms as a natural experiment. The difference-in-differences strategy documents a positive effect of corporate tax deduction on firm patenting. The effect is particularly significant if a firm is of larger size or locates in eastern provinces. We also examine possible channels behind the findings, including changes in R&D and capital investment, intangible assets, financial constraints, and new product sales.     

Appropriation Of Returns From Technological Assets And The Values Of Patents And R&D In Japanese High-Tech Firms †

This paper investigates the extent that technological assets contribute to the value of the firm, using the sample of 90 Japanese firms in pharmaceutical, chemical, and electrical equipment industries. We use the firm's R&D expenditures and the number of patents (in stock) as the measures of its technological assets and show that the relative usefulness of these two measures varies across industries. Particularly, Tobin's q is positively related to the technological assets most strongly in the pharmaceutical industry. It is also most sensitive in this industry to the level of patent stock, coinciding with the view that drug patents are more effective than other patents as a means of appropriating returns from innovation. The communications equipment industry is also characterized by its q's dependence on patent stock. In addition, this industry's q is particularly sensitive to the level of net R&D investment in the most recent year, presumably because of the rapid technological progress in this industry.