Alfred Kleinknecht

The Non-Trivial Choice between Innovation Indicators

We discuss the strengths and weaknesses of five alternative innovation indicators: R&D, patent applications, total innovation expenditure and shares in sales taken by imitative and by innovative products as they were measured in the 1992 Community Innovation Survey (CIS) in the Netherlands. We conclude that the two most commonly used indicators (R&D and patent applications) have more (and more severe) weaknesses than is often assumed. Moreover, our factor analysis suggests that there is little correlation between the various indicators. This underlines the empirical relevance of various sources of bias of innovation indicators as discussed in this paper.

Why do firms cooperate on R&D? an empirical study☆

Abstract Past empirical studies of R&D cooperation were usually based on relatively small samples that are likely to be biased towards prominent cases of R&D cooperation. We analyse R&D cooperation in a large sample of firms which is representative of all manufacturing and service industries in the Netherlands. It turns out that R&D cooperation is a much more widespread phenomenon than is generally thought. In contrast to large parts of the literature, we find that firm size, market structure, R&D intensity, and high shares of product-related R&D have little impact on R&D cooperation between firms. R&D cooperation does not typically occur between big, high tech firms which operate in global markets but occurs at least equally frequently between smaller firms in medium and low tech sectors. Finally, we identify certain factors that enhance the probability that a firm will engage in R&D cooperation with certain types of partners.

Measuring R&D in Small Firms: How Much Are We Missing?

The OECD surveys on R & D are considerably biased towards underestimating R & D in small firms. This conclusion is derived from the authors comparison between the official Dutch R & D survey and the findings of his own survey among 3000 firms. Although being based on Dutch observations only, the authors considerations have an impact on other OECD countries, since the official Dutch R & D survey is fully compatible with those in other countries. The findings call into question the results from a number of empirical studies, bearing on subjects such as market structure and R & D, firm size and R & D or R & D and growth.

Employment Growth and Innovation at the Firm Level

We analyze the influence of innovation on growth rates of employment in 859 Dutch manufacturing firms over the period 1983–1988. Whereas the (growth of the) R&D intensity of firms has a slightly negative impact on employment, we find that firms with a high share of product-related R&D (as a proxi of R&D related to industrial activities in an early stage of the life cycle) experienced an above average growth of employment. The same holds for firms which directed their R&D towards information technology. Smaller firms have, ceteris paribus, substantially higher growth rates of employment than their larger counterparts. Against our expectations, R&D cooperation has no significant impact on employment growth. The same holds for activities in the fields of biotechnology and new materials.

Firm Size, Small Business Presence and Sales of Innovative Products: A Micro-econometric Analysis

In our generalized TOBIT analysis we identify a number of variables which have an impact on a firms innovation “output”. Among others we find that larger firms generally have a higher probability of selling some innovative products, although this probability increases less than proportionately with firm size. Given that a firm has some sales of innovative products, the share of such products in a firms total sales tends to be higher in smaller firms. Moreover, a strong small business presence in a sector seems to enhance imitative innovation but has no influence on “true” innovations, whereas market concentration has no influence on innovation “output” in whatever definition. We also find evidence of regional knowledge spill-overs. Furthermore, our results are consistent with Schmooklers hypothesis that the growth of demand enhances innovation. The outcomes about the impact of R&D collaboration and technology transfer on innovation remain ambiguous.

Demand and innovation: Schmookler re-examined

Schmookler’s “demand-pull” hypothesis, implying that innovation is a function of market demand, has been widely accepted. Re-reading Schmookler’s Invention and Economic Growth we discover a statistical inadequacy. Re-estimating his cross-section results in a different way, we find somewhat lower coefficients. Moreover, the direction of causality appears to be far from obvious. An independent cross-sectional test of the Schmookler hypothesis in the Netherlands, based on R&D instead of patent data, shows similar results: the relationship between indicators of demand and innovation turns out to be generally weaker (but in most cases still significant), and the results suggest that there is a simultaneous relationship between demand and innovation. The latter may be explained by the concept of the “innovation multiplier”, which has implications for macro-economic modelling. 1. Re-reading Schmookler Schmookler’s [13] finding of a strong relationship between investment in capital goods users industries and patent applications by capital goods producing sectors has been generally accepted by economists as evidence that patenting is a function of effective demand (“demand pull” hypothesis). Quite a number of case studies have followed SchmookIer’s seminal work, but there have been

New concepts in innovation output measurement

Why do we need new innovation output indicators? - an introduction, Alfred Kleinknecht analyzing innovation output indicators - the US experience, Zoltan J. Acs and David B. Audretsch collecting literature-based innovation output indicators - the experience in The Netherlands, Alfred Kleinknecht et al the Austrian experience with literature-based innovation output indicators, Peter Fleissner et al the Irish experience with literature-based output indicators, D.J. Cogan extracting significant innovations from published sources in Great Britian, Fred Steward testing innnovation indicators for postal surveys - results from a five-country project, Alfred Kleinknecht discussion and conclusions, Alfred Kleinknecht.

Do Regions Matter for R&D?

KLEINKNECHT A. and POOT T. P. (1992) Do regions matter for RD a notable exception is service industries in the four major cities. Moreover, smaller firms do not seem to be more dependent on their regional environment than their larger counterparts. Furthermore, we find no indication that branch plants of larger firms are typically located in the periphery of The Netherlands. However, our finding that firms in more rural areas tend to have relatively higher shares of process-related R&D is consistent with the urban hierarchy and filter-down theory. KLEINKNECHT A. et Poot T. P. (1992) Les regi...

Determinants of Innovation: A Microeconometric Analysis of Three Alternative Innovation Output Indicators

Taking into account the shortcomings of R&D and patents mentioned in the introductory chapter, new indicators have been developed for the output side of the innovation process. Three of these new output indicators will be analyzed in this chapter, using data from The Netherlands. These indicators consist of: (i) a collection of new product announcements in 1989 from a large number of trade journals in The Netherlands; and (ii) the share of innovative products in a firm’s total sales, the latter being subdivided into: products ‘new to the sector’ (that is, not introduced earlier by a competitor); and products ‘new to the firm’ (that is, already known in the sector). While products new to the sector may be conceived of as being ‘true’ innovations, products new to the firm will often be based on imitation of products introduced earlier by competitors. These two new indicators were obtained by asking innovative firms to subdivide their total product range into three types of product, namely: (i) products essentially unchanged during 1990–2; (ii) products incrementally improved during 1990–2; and (iii) products radically changed or newly introduced during 1990–2. Firms were then asked to report the percentages of their 1992 sales which were related to each of the three categories of product named.

More evidence on the undercounting of small firm R&D

Abstract Drawing from the SEO national survey on R&D and innovation in The Netherlands, we argue that there must be an underrecording of small-scale R&D by orders of magnitude in the R&D survey by the Dutch Central Statistical Office. Such underrecording holds for manufacturing and (even more so) for service firms. This finding is confirmed by estimates based on records of firms receiving R&D subsidies which we use as a third independent source. In order to allow for international comparability of R&D data, the survey methods of the Dutch Central Statistical Office are similar to those in other OECD countries. Therefore, our findings in the Dutch manufacturing and service industries are likely to have an impact for the measurement of R&D in other OECD countries. The results have implications for our judgement about the famous “Schumpeterian hypotheses” concerning the influence of firm size and market structure on R&D.