Innovation in firms: the OECD innovation microdata project
Latest news: Innovation in Firms: A Microeconomic Perspective published in November 2009.
Innovation surveys were developed to increase our knowledge about innovation in firms beyond what was already covered by R&D surveys, patent data or bibliometric indicators. In order to devise appropriate innovation policies it was felt that more needed to be known regarding the types of innovations, the reasons for innovating (or not), the collaboration and linkages among agents, and the flows of knowledge, and that new quantitative data should be collected on the inputs and outputs of innovation.
In order to harmonise and promote the quality of innovation surveys, the Oslo Manual was developed by the OECD and Eurostat, the first edition of which dates back to 1992. Since then, the Manual has been updated twice to improve its guidelines in view of the acquired experience. While initially it was designed for firms in the manufacturing sector, it was later modified to include the services sector. While initially it mainly dealt with product and process innovations, it was later extended to cover organisational and marketing innovations. In the 3rd version, published in 2005, an appendix is devoted to the discussion of how to frame innovation surveys in developing countries. For the 2005 edition of the Oslo Manual, visit www.oecd.org/sti/oslomanual.
Innovation surveys were first experimented with in several Western European countries but have since been conducted in many other countries including Canada, all EU countries (where the Community Innovation Surveys (CIS) are in their 6th round in 2008), Switzerland, Russia, Turkey, Australia, New Zealand, South Korea, South Africa and most Latin American countries. The notable exception where no official innovation survey based on the Oslo Manual framework exists at this time is the United States.
The OECD innovation microdata project launched in 2006 aims at exploiting microdata (firm-level data) from innovation surveys for economic analysis and circumvents restrictions regarding access to firm-level data (due to confidentiality constraints) by following a decentralized approach. During 2007, research teams from around 20 countries worked in a coordinated way using similar data cleaning methods and econometric models on their national data sets and producing harmonized tabulations with results. The core data used for this work come from recent innovation surveys (notably CIS-4 for European countries) with some countries being able to link these to other national data sources. This project includes two modules:
1. Econometric analysis; and
Innovation survey data have been increasingly used recently in order to explore a number of questions regarding the determinants, the effects and some of the characteristics of innovation. Here is an illustrative list of topics that have been examined using these data:
Except for a few experiences, almost all studies have been conducted at the level of single countries. This allows to draw conclusions valid for that country, but does not tell whether these conclusions are valid for other countries as well. For this project, the goal was to focus on a small set of research themes which countries could address using harmonised statistical techniques. Four research themes were thus identified (which also entailed the compilation of comparable indicators):
The econometric approach allows to estimate functional relationships between the variables which could differ across sub-groups of firms. It will tell for instance if firms spending more on innovation tend to have a higher innovative turnover and increase in productivity, and qualify the relationships across countries or by firm size. As a result of cross-country disparities in data availability some models could not be estimated in all participating countries, or just in a simplified version.
Standard indicators of innovation, such as the share of firms having developed a new product or a new process in a country are extremely informative and are compiled in a standard way by survey producers. However, in order to understand the innovation process and to conduct more refined cross-country comparisons, it is necessary to compile more sophisticated indicators, which entails crossing several variables or performing the calculations for a particular subset of firms. Microdata-based indicators allow to reflect the diversity of firms. Some firms innovate, others do not. Among those which innovate, the innovation performance is skewed (some are highly innovative, other are less so). Firms differ also in the types of innovation that they perform (product, process, organisational, marketing).
Microdata-based indicators allow to differentiate firms by their size, by their industry etc. They also allow to cross variables, which can be extremely informative on the innovation process, as it allows to identify particular categories of firms with several characteristics. For instance, one would expect that firms doing new-to-market product innovation and doing intramural R&D are in a way “more innovative” than firms doing new-to-the firm innovation based on other firms’ R&D. This module aims at developing such indicators and computing them for the participating countries.
The book Innovation in Firms: A Microeconomic Perspective was published in November 2009. A summary of the project and its outputs also appeared in the 2008 edition of the OECD Science, Technology and Industry Outlook, available at www.oecd.org/sti/outlook.
Links to innovation survey questionnaires and reports
Canada (survey of innovation 2005):
Australia (2005 innovation survey):
New Zealand (Business Operations Survey 2005):