Enterprise and Small Business Principles

The knowledge production function

With regard to theories of innovation, firms are considered exogenous and their per­formance in generating technological change endogenous. For example, in the most prevalent model found in the literature of technological change, the model of the know­ledge production function, formalised by Griliches (1979), firms exist exogenously and then engage in the pursuit of new economic knowledge as an input to the process of generating innovative activity. The most decisive input in the knowledge production function is new economic knowledge. The greatest source generating new economic knowledge is generally considered to be R&D.

When it came to empirical estimation of the knowledge production function, it became clear that measurement issues played a major role. The state of knowledge regarding innovation and technological change has generally been shaped by the nature of the data that were available to scholars for analyses. Such data have always been incomplete and, at best, represented only a proxy measure reflecting some aspect of the process of technological change. The greatest obstacle to understanding the economic role of tech­nological change was a clear inability of scholars to measure it. More recently, Cohen and Levin (1989) warned: ‘A fundamental problem in the study of innovation and technical change in industry is the absence of satisfactory measures of new knowledge and its contribution to technological progress. There exists no measure of innovation that permits readily interpretable cross-industry comparisons.’

Measures of technological change have typically involved one of the three major aspects of the innovative process:

1 a measure of the inputs into the innovative process, such as R&D expenditures, or else the share of the labour force accounted for by employees involved in R&D activities;

2 an intermediate output, such as the number of inventions which have been patented; or

3 a direct measure of innovative output.

These three levels of measuring technological change have not been developed and analysed simultaneously, but have evolved over time, roughly in the order of their pre­sentation. That is, the first attempts to quantify technological change at all generally involved measuring some aspects of inputs into the innovative process. Measures of R&D inputs - first in terms of employment and later in terms of expenditures - were only introduced on a meaningful basis enabling inter-industry and inter-firm compar­isons in the late 1950s and early 1960s.

A clear limitation in using R&D activity as a proxy measure for technological change is that R&D reflects only the resources devoted to producing innovative output, but not the amount of innovative activity actually realised. That is, R&D is an input and not an output in the innovation process. In addition, R&D measures incorporate only efforts made to generate innovative activity that are undertaken within formal R&D budgets and within formal R&D laboratories. The extent of informal R&D is considerable, particularly in smaller enterprises, and not all efforts within a formal R&D laboratory are directed towards generating innovative output in any case. Other types of output, such as imitation and technology transfer, are also common goals in R&D laboratories.

As systematic data measuring the number of inventions patented were made pub­licly available in the mid-1960s, many scholars interpreted this new measure not only as being superior to R&D but also as reflecting innovative output. In fact, the use of patented inventions is not a measure of innovative output, but is rather a type of inter­mediate output measure. A patent reflects new technical knowledge, but it does not indicate whether this knowledge has a positive economic value. Only those inventions that have been successfully introduced in the market can claim that they are innova­tions as well. While innovations and inventions are related, they are not identical. The distinction is that an innovation is ‘a process that begins with an invention, proceeds with the development of the invention, and results in the introduction of a new prod­uct, process or service to the marketplace’ (Edwards and Gordon, 1984, p. 1).

Mansfield (1984, p. 462) has explained why the propensity to patent may vary so much across markets: ‘The value and cost of individual patents vary enormously within and across industries.. . Many inventions are not patented. And in some industries, like electronics, there is considerable speculation that the patent system is being bypassed to a greater extent than in the past. Some types of technologies are more likely to be patented than others.’ The implications are that comparisons between enterprises and across industries may be misleading. According to Cohen and Levin (1989): ‘There are significant problems with patent counts as a measure of innovation, some of which affect both within-industry and between-industry comparisons.’

Thus, even as the US Patent Office has introduced new and superior sources of patent data, such as the new measure of patented inventions from computerisation, the reliability of these data as measures of innovative activity has been severely challenged. For example, Pakes and Griliches (1980, p. 378) warn that ‘patents are a flawed mea­sure (of innovative output); particularly since not all new innovations are patented and since patents differ greatly in their economic impact.’ And in addressing the question, ‘Patents as indicators of what?’, Griliches (1990, p. 1669) concludes that: ‘Ideally, we might hope that patent statistics would provide a measure of the (innovative) output. . . The reality, however, is very far from it. The dream of getting hold of an output indicator of inventive activity is one of the strong motivating forces for economic research in this area.’

Besides the fact that many, if not most, patented inventions do not result in an innovation, a second important limitation of patent measures as an indicator of inno­vative activity is that they do not capture all of the innovations actually made. In fact, many inventions that result in innovations are not patented. The tendency of patented inventions to result in innovations and of innovations to be the result of inventions which were patented combine into what Scherer (1983a) has termed as the propensity to patent. It is the uncertainty about the stability of the propensity to patent across enterprises and across industries that casts doubt upon the reliability of patent meas­ures. According to Scherer (1983): ‘The quantity and quality of industry patenting may depend upon chance, how readily a technology lends itself to patent protection, and business decision-makers’ varying perceptions of how much advantage they will derive from patent rights. Not much of a systematic nature is known about these phenomena, which can be characterised as differences in the propensity to patent.’

Just as for the more traditional measures of technological change, there are also cer­tain limitations associated with the direct measure of innovative activity. In fact, one of the main qualifications is common among all three measures - the implicit assumption of homogeneity of units. That is, just as it is implicitly assumed that each dollar of R&D makes the same contribution to technological change, and that each invention which is patented is equally valuable, the output measure implicitly assumes that innovations are of equal importance. As Cohen and Levin (1989) observe: ‘In most studies, process innovation is not distinguished from product innovation; basic and applied research are not distinguished from development.’ Thus, the increase in the firm’s market value result­ing from each innovation, dollar expended on R&D, and patent, is implicitly assumed to be homogeneous - an assumption which clearly violates real-world observation.

The knowledge production function has been found to hold most strongly at broader levels of aggregation. The most innovative countries are those with the greatest invest­ments to R&D. Little innovative output is associated with less-developed countries, which are characterised by a paucity of production of new economic knowledge. Similarly, the most innovative industries also tend to be characterised by considerable investments in R&D and new economic knowledge. Not only are industries such as computers, pharmaceuticals and instruments high in R&D inputs that generate new economic knowledge, but also in terms of innovative outputs (Audretsch, 1995). By contrast, industries with little R&D, such as wood products, textiles and paper, also tend to produce only a negligible amount of innovative output. Thus, the knowledge production model linking knowledge-generating inputs to outputs certainly holds at the more aggregated levels of economic activity.

Where the relationship becomes less compelling is at the disaggregated micro­economic level of the enterprise, establishment or even line of business. For example, while Acs and Audretsch (1990) found that the simple correlation between R&D inputs and innovative output was 0.84 for four-digit standard industrial classification (SIC) manufacturing industries in the US, it was only about half, 0.40, among the largest US corporations. The model of the knowledge production function becomes even less compelling in view of the recent wave of studies revealing that small enterprises serve as the engine of innovative activity in certain industries. These results are startling, because, as Scherer (1991) observes, the bulk of industrial R&D is undertaken in the largest corporations; small enterprises account only for a minor share of R&D inputs.

As an example of the R&D carried out in different countries, the following com­parison (UNICE Benchmarking Report, 1999) of business expenditures on R&D per capita (in $, based on 1997 data), suggests that the US is a global leader of knowledge production (see Figure 5.1). Figure 5.2 compares knowledge outputs. The number of

00

о

Figure 5.1 International comparisons of business research and development

USA

EU-14

Sweden

Finland

Denmark

Canada

Netherlands

Norway

Ireland

Iceland

Spain

Portugal

Turkey

SHAPE * MERGEFORMAT

R&D expenditure per capita in $

Figure 5.2 International comparisons of patent numbers

2.2 2.2

^■ 2.1 ^^■1.6

I 1.2

0.9 0.9

I 0.6 0.6 0.4

0.1

0.1

I 7

I 4

□ 26

I 5.2

■ 4.7

4.6

4.3

3.8

3.6

■ 3.1 3

Application for patents per 10,000 population (1996)

yearly patent applications is a potential indicator of newly created knowledge that is intended for commercialisation. While the US is historically the largest investor in R&D on the global market, when entrepreneurial activity measurements are based on the number of patent applications submitted (1996 data, per 10,000 population), the US ranks only second after Japan.