Elsevier

Research Policy

Volume 36, Issue 2, March 2007, Pages 260-273
Research Policy

Relative efficiency of R&D activities: A cross-country study accounting for environmental factors in the DEA approach

https://doi.org/10.1016/j.respol.2006.11.004Get rights and content

Abstract

This paper applies the production framework associated with the data envelopment analysis (DEA) method to evaluate the relative efficiency of R&D activities across countries. R&D capital stocks and manpower are treated as inputs while patents and academic publications are considered as outputs. A three-stage approach, which involves using DEA for evaluating efficiency and using Tobit regressions for controlling the external environment, is applied to 30 countries in recent years. The results show that less than one-half of the countries are fully efficient in R&D activities and that more than two-thirds are at the stage of increasing returns to scale. Most countries have a more significant advantage in producing SCI cum EI publications than in generating patents.

Introduction

R&D activity is a well-organized process of knowledge creation, production, diffusion, and application. It entails innovation in scientific technology, in management measures, and in social and political systems, etc. OECD (2003) defined the investment in knowledge as the sum of R&D expenditure, expenditure for higher education and investment in software. Since R&D investment is one of the most crucial elements in promoting scientific and technological progress, any country that uses the resources inefficiently could bear a penalty in the form of achieving a much slower progress. Furthermore, if R&D resources are not used effectively, additional investment may be of little help in stimulating progress. However, the relevant literature has focused primarily on efforts at new R&D investment and comparatively little attention has been given to the effective use of the resources, particularly at the national level, once they are in place. This is a potentially important omission, since the very conditions responsible for scientific and economic backwardness may operate through the poor management of R&D activities. Understanding the nature of R&D efficiency/inefficiency is important for designing policies to improve resource allocation.

In this paper, we attempt to fill in this gap by examining the efficiency of national R&D activities. We propose a three-stage approach, which involves applying data envelopment analysis (DEA) for estimating efficiency and Tobit regressions for controlling the external environment. Following Pakes and Griliches (1984) and Griliches (1990), this paper considers R&D to be a production process and regards each country as a decision-making unit (DMU) conducting R&D. By setting up an inter-country R&D innovation production framework and using the DEA model and Tobit regression iteratively, our three-stage approach can identify and separate the intrinsic technical inefficiency in the R&D process from the external effects stemming from the operating environment, which differs substantially from country to country.1 The sample in this paper consists of thirty (30) countries that engage in R&D activities intensively. In addition, slack and advantage analyses provide detailed assessments on each country's R&D resource allocation.

There has been a large amount of literature devoted to discussing the effects of R&D investment on raising productivity and profits at the firm and industry levels. Mansfield, 1980, Mansfield, 1988, Terleckyj (1982), Griliches (1986), Meliciani (2000), Hartmann (2003), Timmer (2003), and Gonzalez and Gascon (2004) provided evidence from many industries in various countries. Feller (1990) and Adams and Griliches (2000) emphasized the importance of the productivity of basic research in universities.

Only in recent years have a few examples in the literature discussed R&D efficiency by using quantitative approaches with regard to R&D at the firm level, Zhang et al. (2003) applied the stochastic frontier analysis (SFA) approach to the R&D efforts of Chinese firms to examine the difference in efficiency among various types of ownership. As for academic research, Korhonen et al. (2001) and Cherchye and Vanden Abeele (2005) applied the DEA technique to evaluate the efficiency of university R&D in Finland and the Netherlands, respectively.

The rest of the paper is organized as follows. Section 2 presents a summary of the overall methodology and discusses the R&D production framework and the DEA model of the efficiency measure. Section 3 gives a description of the data management and the hypotheses tested in the paper. Section 4 presents the empirical results of the three-stage approach. A comparison of the relative efficiency scores for Stage I and Stage III as well as advantage analysis are also performed in this section. The final section provides a summary and the conclusion.

Section snippets

An overall summary of the methodology

This paper sets up an inter-country innovation production framework for R&D activities in 30 countries. Each country is regarded as a DMU that employs R&D manpower and physical resources as inputs to produce countable outputs. Inspired by Fried et al. (1999), we propose a three-stage approach to analyze the relative efficiency of R&D production. In the first stage, the input-oriented DEA model is applied to estimate the inter-country efficiency frontier of R&D activities. Technical efficiency

Data and hypotheses

A total of 30 countries form the sample used in this study. Twenty-three (23) of them are OECD members and seven (7) are non-OECD economies. Quantitative R&D input and output data for these countries are collected and processed. Official data compiled and released by international organizations and governments, such as OECD, UNESCO, the World Intellectual Property Organization, and the U.S. Patent and Trademark Office, etc., are used. A list of countries and detailed data sources are given in

Empirical results of the three-stage approach

The efficiency of an R&D unit to transform inputs into outputs is influenced by its technical proficiency, which is mostly under its control, and external operating environment, which is usually beyond its control. The purpose of the three-stage approach is to obtain a measure of net technical efficiency after controlling for the exogenous features of the environment.15

Implications of the findings and concluding remarks

Large amounts of resources have been devoted to scientific R&D in many countries. However, if R&D resources are not used effectively, additional investment may be of little help in stimulating scientific and technological progress. This paper proposes a three-stage approach to evaluate, in a production framework, the relative technical efficiency of R&D across 30 countries, mainly by controlling for country variations in external environments. The commonly used non-parametric DEA technique is

Acknowledgements

This paper is part of a joint research project carried out when the first author was a visiting scholar at Western Michigan University, under the financial support of the Fulbright Program, Grant No. 68428298, USA, and the National Science Council, Grant No. 42115F, Taiwan. The authors would like to thank two anonymous referees of the Journal and E. Alvi, D. Alexander, M. Higgins, J. Kimmel, and seminar participants at Western Michigan University and at the 2005 Taipei Conference on Efficiency

References (32)

  • R.D. Banker et al.

    Some models for estimating technical and scale inefficiency in data envelopment analysis

    Management Science

    (1984)
  • A. Charnes et al.

    Evaluating program and managerial efficiency: an application of data envelopment analysis to program follow through

    Management Science

    (1981)
  • T. Coelli et al.

    An Introduction to Efficiency and Productivity Analysis

    (1998)
  • Eaton, J., Kortum, S., 1995. Trade in ideas: patenting and productivity in the OECD, NBER Working Papers Series, No....
  • M.J. Farrell

    The measurement of productive efficiency

    Journal of the Royal Statistical Society, Series A

    (1957)
  • H.O. Fried et al.

    Incorporating the operating environment into a nonparametric measure of technical efficiency

    Journal of Productivity Analysis

    (1999)
  • Cited by (0)

    View full text