Tuesday, September 26, 2006

Concurrent workshops

• Session B–1 The role of public sector actors in innovation
• Session B–2 Specialised surveys: Developing countries, remote regions, special topics
• Session B–3 Global knowledge flows – The diffusion and location of knowledge

Session B–2 Specialised surveys: Developing countries, remote regions, special topics

• Science, technology and innovation for sustainable development
• What drives productivity growth in Tanzania : Technology or institutions?
• Specialised R&D surveys: Design and application
• Measuring SERVERD – Pie in the sky or substantive activity?

Biographical Note

Michael Bordt is Chief of the Human Resources and Intellectual Property Section in the Science, Innovation and Electronic Information Division at Statistics Canada. He is also the editor of Statistics Canada's Innovation Analysis Bulletin. In his current position he is responsible for surveys of intellectual property commercialization (one in universities and one in federal departments and agencies), developing new statistics on: human resources in science and technology, commercialization, sub–provincial S&T indicators and business dynamics of technology companies. In his previous position at Statistics Canada, he was Chief of the Environmental Information and Spatial Accounts Section in the National Accounts and Environment Division.

Johanne Boivin has a bachelor's degree in mathematical economics and a master's degree in economics from Université Laval. She also worked at the university as a research assistant. She joined Agriculture and Agri–food Canada as a market development analyst in 2003. In 2005, she accepted an assignment as an analyst at Statistics Canada, where she works with data from the Bioproducts Development Survey.

Julio M. Rosa has a master's degree in economics from the Université du Québec à Montréal and is currently writing his doctoral thesis, which is about organizational complementarity in research and development. He has been working at Statistics Canada for more than five years as an analyst specializing in innovation and science and technology.

Abstract

A broadening of the scope of the international STI framework may be required to better measure activities in relation to their contribution to sustainable development. The particular situation of developing countries and frontier regions (including Canada 's North) will also be considered. The two concepts are closely related since most economic development issues are bound by sustainable development principles. In both cases, developing regions and sustainable development, it is necessary to understand the intent of the R&D, the technology or the product.

The intent may not be evident from the industrial sector of the performer. In terms of sustainable development, the Frascati Manual recommends the classification of R&D into socio–economic objectives and one of these objectives includes "control and care of the environment". In Canada , this classification is applied only in classifying federal government R&D. There are no socio–economic objectives covering developing areas.

The paper will highlight Canada 's experience in measuring R&D for developing countries; measuring R&D activities that support sustainable development: the case of bioproducts development sector and the results of a project to tabulate the extent of R&D for sustainable development in Canada 's industrial R&D survey.

Indicators will be proposed that are both derived from existing data, for example by linking technology or industry surveys with R&D and innovation surveys. For example, Statistics Canada conducts a comprehensive survey of the environment industry. As well, indicators that require new specific surveys will be proposed.

Related S&T frameworks, such as field of science and occupational classifications, will be reviewed in terms of their capacity to distinguish "sustainable development".

Top of Page

Biographical Note

Pierre Mohnen is Professor at the Faculty of Economics and Business Administration of Maastricht University and Professorial Fellow at UNU–MERIT (Maastricht Economic Research on Innovation and Technology). He is also Associated Fellow at CIRANO, the Montreal–based Center for Interuniversity Research and the Analysis on Organizations. He has a BA and an MA in economics from the Catholic University of Louvain and a PhD in economics from New York University. He was Professor of Economics at the University of Québec in Montréal (UQAM) from 1984 to 2001.

His research areas are the economics of production, applied econometrics, productivity and innovation. He does work on the effectiveness of R&D tax incentives, the determinants of innovation, the relationship between R&D, innovation, and productivity, competition and productivity, complementarities in innovation policies, the economics of intellectual property rights, and the estimation of informational rents in public contracts.

Micheline Goedhuys and Norbert Janz (UNU–MERIT)

Abstract

How important are R&D, technology and innovation for growth in a developing country? Some authors like Lederman and Saenz (2005) show that innovation has a positive effect on economic growth, stronger than the effect of institutions. Others, like Banerjee and Duflo (2004) emphasize the role of institutions, like low investor protection, excessive government interventions, or family arrangements in the splitting of work and revenue, in explaining productivity growth in developing countries. We shall address this controversy for one of the least developed countries, Tanzania.

We do so by using micro establishment data from the Investment Climate Survey conducted by the World Bank and covering the years 2000–2002. These surveys reveal a wide range of information on the sampled individual establishments: inputs, outputs, ownership structure, environmental variables, such as infrastructure, business–government relations, legal environment, labor relations, free trade policies, industry structure, intellectual property rights, financial conditions, and also some questions regarding R&D efforts, training, and innovation output. After a serious effort of cleaning and careful reconstruction of certain variables from various pieces of information, we end up with a cross–section of 340 observations (to be double–checked) with some lagged variables that can be used as instruments.

We first estimate a Cobb–Douglas production function where output, measured by value added, is regressed on labor, the capital stock, various measures of innovation (a training dummy or the percentage of trained people, an R&D dummy or the R&D/sales ratio, a product innovation dummy, a process innovation dummy, the number of products introduced in the last three years, a licensing dummy), various uses of ICT (E–mail, internet, website), and also controlling for the age of the firm, capacity utilization, the importance of foreign ownership, the importance of temporary employment, access to electricity, belonging to a business association, subcontracting, the skill composition and level of education of the employees, and industry dummies.

In a second stage we run quantile regressions to find out whether the determinants of growth, and in particular the relative importance of innovation and environmental variables, differ for different quantiles of the conditional distribution of the dependent variable. We shall also explore the existence of some complementarities and do some instrumental variables regression to control for possible endogeneities. Sensitivity results will also be conducted with respect to alternative measurements of the variables, or alternative specifications such as level versus growth regressions.

Preliminary results suggest that R&D is at best positively significant for large R&D performers. For the greatest bunch of R&D performers, R&D has no above normal rate of return. Innovation output variables also turn out non–significant. Significant and robust are access to credit, belonging to a business association, and possession of own electricity generator. Out first results seem to suggest that the environmental business climate is more important for fostering growth in a country like Tanzania than innovation.

For the Blue Sky indicators conference and for the NEPAD initiative, this paper suggests the usefulness of alternative data sources like the Investment Climate Survey of the World Bank for most African countries that have not yet conducted an innovation survey. The preliminary results perhaps also indicate that besides collecting data on innovation, it is equally important to collect data on the environmental conditions in which firms operate.

Top of Page

Biographical Note

Peter Mortensen is Head of Department and Carter Bloch assistant professor at the Danish Centre for Studies in Research and Research Policy. Both are active in S&T measurement and survey design, and were draft writers for the recent Oslo Manual revision.

Abstract

The OECD has in recent years promoted specialized surveys on biotechnology, based on the need for more detailed information and indicators within this area. This same need is present in other areas, and the demand for more detailed, specialized indicators can be expected to increase in the future.

While specialized surveys can in many cases be undertaken by universities and other organizations, there are a number of benefits from linking these surveys to national R&D statistics. For example, R&D surveys can be used to determine target populations for specialized surveys, and these surveys can often be integrated with general surveys, either directly or through a linking of data afterwards.

This paper discusses two specialized surveys that have been conducted in Denmark, surveys of R&D within information and communication technologies, and surveys of R&D activities that are related to Greenland. Both surveys offer a number of insights in terms of their methodology that can be applied to other areas. Furthermore, the data can be used to develop a number of specialized indicators of use both for research and policy.

The central importance of information and communications technologies (ICT) for innovation and growth are very widely stressed. Towards measuring the role of ICTs, a large number of surveys have been conducted on both firms' and households' usage of ICTs. However, there have been almost no attempts to measure ICT R&D activities for business enterprises. ICT R&D is vital for growth and innovation both in ICT–related and non–ICT sectors and also provides an important base for the overall diffusion of ICTs in the economy.

The ICT R&D survey covers ICT R&D activities in both the public and private sector, thereby allowing a matching of R&D in firms and public research institutions, with a number of insights for policy. Within the business enterprise sector, non–ICT sectors are also surveyed, providing information on which non–ICT firms conduct ICT R&D and in what ICT research areas.

Among the topics in the survey are a two–dimensional classification of types of ICT R&D activities, R&D cooperation, commercialization activities and barriers to ICT R&D and private–public cooperation. In addition, the survey is linked to the general R&D survey, allowing the construction of a number of other ICT related indicators.

Greenland , which is an autonomous territory of Denmark, has begun establishing an R&D base in the last decade, with the establishment of the University of Greenland and increases in business R&D. There is thus an increasing need for data on Greenland–related R&D activities, both performed by Danish and Greenlandic institutions and enterprises, and on the linkages between these activities.

The Greenland–related R&D survey is conducted biannually and is integrated with the general Danish R&D survey, isolating R&D activities that are related to Greenland and where these activities take place.

Top of Page

Biographical Note

Michael Kahn is executive director of the Knowledge Systems research programme of the Human Sciences Research Council (HSRC). He is responsible for the Centre for Science, Technology and Innovation Indicators (CeSTII) and the GIS and Surveys units.

Abstract

All industrialized and even industrializing economies now exhibit a dominance of services in the composition of their GDP. The services sectors are dynamic and innovative, and frequently show growth rates in excess of the manufacturing sector. Services include 'knowledge intensive service activities' or 'KISAs' that are both part of the innovation cycle and innovative in and of themselves. However KISAs as intangible aspects of company assets, especially the area of R&D expenditure are felt to present difficulties for accounting purposes, let alone the production of meaningful and reliable S&T indicators.

This paper argues that the problem of measuring services expenditure on R&D (SERVERD) is perhaps overstated. Working from the Frascati definition of R&D, a set of company case studies is presented in order tease out the various dimensions of SERVERD that are present both in knowledge intensive business services (KIBS) and firms outside the service sector. Case studies presented include banking, insurance, clinical trials, leisure, and mineral extraction. It is argued that the solution to the problem of computing SERVERD lies both in the co–generation of a mutual understanding between measurer and respondent of what counts as SERVERD, and a clear understanding of the value chain contributing to the R&D project under scrutiny. To this end a core set of questions that can form the basis of semi–structured interviews are identified.

As an effective mechanism for collating SERVERD it is suggested that survey agencies identify one or two exemplar firms per sub–sector; understand the firm(s); benchmark the scale of firm SERVERD in relation to a reliable company accounting measure (revenue, EBIDTA, COGS) and use this ratio to impute comparable values for other similar firms.

The implication of this approach is that conventional 'knock and drop' survey technique is inadequate to the task of measuring SERVERD. Greater resources for measurements are required in order that the competent national authority performing the measurement is enabled to get closer to the responding entities. This proposed resolution of the measurement problem is not definitional per se, but instrumental. A raft of case studies drawn across sub–sectors and across countries may serve to provide a basis for a common approach to measuring SERVERD for KIBS and the broader area of knowledge intensive services activities.