1. Introduction
There is a wide variety of views on the strategies, practices & characteristics of firms that make the greatest contribution to innovation and productivity improvement. Internationally, those behaviours that have been found to be positively related to, or a cause of, better firm performance include: research & development (R&D) (e.g., Griliches 1994; Guellec & van Pottelsberghe de la Potterie 2001); human resource management (HRM) practices (e.g., Ichniowski et al. 1997; Lazear 2000; Sels et al. 2006); human capital investment (e.g., Black & Lynch 1996; Addison & Belfield 2004); innovation (e.g., Crepon et al. 1998; Hall & Mairesse 2006); international engagement (e.g., Tybout 2000; Criscuolo et al. 2005); and information & communication technology (ICT) investments (e.g., OECD 2004a; Australian Productivity Commission 2004).
What, perhaps, is interesting given this wealth of evidence is the rather sparse uptake of many of these "high-performance" activities (charts 5 & 6 demonstrate this for New Zealand). Part of the explanation may be that the applicability of such practices will vary by the characteristics of the markets in which firms operate (e.g., the usefulness of patents depend, among other things, on the appropriability characteristics of the knowledge being protected, Levin et al. 1987). The dispersion of practices is also consistent with a resource-based view of the firm, whereby decision-makers within firms deploy resources with idiosyncratic views and "initial" endowments of ability, knowledge, etc (Penrose 1959; Wernefelt 1984). In this view, much emphasis is placed on the market search function, or "entrepreneurial spirit", that takes technologies and tests their worth in localised market conditions (e.g., Hausmann & Rodrik 2002; Baumol 2002). Interpreted in this light, our understanding of the microeconomics of productivity growth provides empirical evidence consistent with the importance of learning-by-doing and path-dependence (e.g., Nelson & Winter 1982; Aghion & Howitt 1998). Two particularly robust results from this literature are the persistence of relative productivity performance over time and the positive contribution of firm turnover to aggregate productivity growth (e.g., Haltiwanger et al. 1999; Bartelsman & Doms 2000; OECD 2004b; Law & McLellan 2005).
From a policy perspective, a narrow view of innovation runs the risk of missing or misattributing the importance of joint strategic decisions, and other activities directed by those decisions, within the firm. Bias can arise in assessing the impact of, say, R&D or ICT investment if innovation outcomes are entirely attributed to these activities.1 Good econometric studies, of course, control for firm fixed effects (of which management quality might be a likely contributor), reducing the risk of coefficient bias. But this doesn't enable the real goal of understanding what truly matters. Knowing that the econometric technique has compensated for some missing variables does not shed much light on the debate around government's role in innovation policy and productivity growth. As Arundel (2005) points out, policy development can tend to coalesce around what is measured (despite the caveats researchers place around their work).
Smith (2006) – in his assessment of the New Zealand innovation system – emphasises the importance of taking a systemic view of innovation policy. Even if the policy guidance dictated by the data is right, risk still arises in policy design if related functions within the firm are not appreciated and accounted for. Would our assessment of best-practice R&D policy be different if we better understood the influence of good marketing or customer engagement on firm innovation outcomes? More generally, from a systems perspective, firms may legitimately choose not to be cutting edge innovators themselves. How should their non-technological practices be structured to enable them to be "fast adopters" of knowledge generated elsewhere in the (world) economy?
Traditionally we have assessed the innovation capability of firms by asking questions such as: "Do you do (Frascati manual) R&D?"; "Do you own a legal monopoly over intellectual property (e.g., a patent)?"; "Do you employ scientists/staff with advanced academic qualifications?" Measures of this nature generally provide a mixed message on the innovative capacity of New Zealand businesses. For example, cross-country comparison of business expenditure on R&D suggests relative weakness, while product & process innovation outcomes suggest performance comparable to the EU average (see charts 1 & 2). The desire to improve New Zealand's innovative capacity motivates much of the domestic debate around government economic policy (e.g., New Zealand Government 2002).
Chart 1: New Zealand’s innovation performance relative to other OECD economies
Source: MED & Treasury (2005)
→ Full size version of Chart 1 [125 kB JPEG]
Chart 2: R&D expenditure (% of GDP) in 2002 or latest year available
Source: OECD (2005)
This paper starts by outlining how research questions stimulated by this debate have resulted in a unique dataset administered by Statistics New Zealand (SNZ). The key concern of the paper is to provide some assessment of the potential "mismeasurement" of innovation arising from a pure focus on technological product & process improvement. To this end, we investigate how broader measurement of innovation changes our understanding of what an innovative New Zealand firm looks like. The ultimate question that interests us is: How do firm practices influence performance (including innovation outcomes)? An answer to this question is the subject of a long-term research programme. The hope is that this paper provides some sense of how policy agencies in New Zealand are working towards the goal of better understanding the economic development process & innovation's role within that process.
Dataset development has a key role in aiding this better understanding. Section 2 of this paper outlines the Business Operations Survey (BOS) that SNZ and policy agencies have constructed in order to begin to address our knowledge gaps. In section 3, we compare and contrast different innovation measures (across both outcomes & practices) within BOS. This is the first research paper to use this data (which was officially released on 22 August 2006), and the analysis is, naturally, exploratory. In section 4, we introduce a panel dataset and begin to explore the evolutionary relationship between innovation outcomes & broader firm practices. We conclude, in section 5, by outlining future work with this data.
Back to Top