Transcript R-D, Innovation and Productivity: Some Thoughs about an

Innovation and Productivity in France:
A firm-level analysis for Manufacturing
and Services (1998-2000 and 2002-2004)
Jacques Mairesse
Stéphane Robin
CREST-ENSAE,
UNU-MERIT, and NBER
BETA
University of Strasbourg 1
Innovacio a nivell empresarial
XREAP, Universitat de Barcelona, 1de juliol de 2008
OUTLINE
• Bird-eye view of the « CDM » model
• Model and empirical strategy
• Data
• Specification of the model
• Results
• Comparisons and further analyses
• Conclusion
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
2
The « CDM » model
with Bruno Crepon and Emmanuel Duguet
• Brings together the three main fields of
investigation in the econometrics of research and
innovation
• Proposes a “simple” framework articulating
innovative and productive activities
• Takes advantage of the innovation survey
information
• Uses estimation methods appropriate to the
specification of the model and nature of data
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
3
The « CDM » model
Diversification
Market share
Demand Pull
Technology Push
R&D
Size
Industry
Knowledge Capital
Innovation
Physical Capital
Skills
Jacques Mairesse
Stephane Robin
Productivity
Barcelona, 1 July 2008
4
Past and on-going work (1)
•
•
•
•
•
•
Crépon, B., E. Duguet and J. Mairesse (1998), “Research and Development,
Innovation and Productivity: An Econometric Analysis at the Firm Level”,
Economics of Innovation and New Technology, 7(2), 115-158.
Mairesse, J. and P. Mohnen (2001), “To Be or Not to Be Innovative: An Exercise in
Measurement”, STI Review. Special Issue on New Science and Technology
Indicators, OECD, 27, 103-129.
Mairesse, J. and P. Mohnen (2002), “Accounting for Innovation and Measuring
Innovativeness: An Illustrative Framework and an Application”, American Economic
Review, Papers and Proceedings, 92(2), 226-230.
Mairesse, J. and P. Mohnen (2005), “The Importance of R&D for Innovation: A
Reassessment Using French Survey Data”, The Journal of Technology Transfer,
special issue in memory of Edwin Mansfield, 30(1-2), 183-197.
Hall, B.H. and J. Mairesse (2006),"Empirical Studies of Innovation in the
Knowledge Driven Economy", Introduction to a special issue on: “Empirical studies
of innovation in the knowledge driven economy”, Economics of Innovation and
New Technology, 15(4/5), 289-299.
Mohnen, P., J. Mairesse and M. Dagenais (2006), “Innovativity: A Comparison
across Seven European Countries”, special issue on: “Empirical studies of
innovation in the knowledge driven economy”Economics of Innovation and New
Technology, 15(4/5), 391-413.
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
5
Past and on-going work (2)
•
•
•
•
•
•
Kremp, E., and J. Mairesse (2003), “Knowledge Management, Innovation and
Productivity: A Firm Level Exploration Based on the French CIS3 Data”, in D.
Foray and F. Gault eds., “Measuring Knowledge management in the Business
Sector”, OECD.
Mairesse, J. and P. Mohnen (2004), “Intellectual Property in Services: What Do We
Learn from Innovation Surveys ?”, in Patents, Innovation, and Economic
Performance, OECD Conference Proceedings, OECD, Paris, 227-245.
Griffith R., E. Huergo, J. Mairesse and Bettina Peters (2006), "Innovation and
Productivity across Four European Countries", Oxford Review of Economic Policy,
22(4), 483-498.
Hall, B. H., F. Lotti and J. Mairesse (2007) "Employment, Innovation and
Productivity: Evidence from Italian MicroData", Industrial and Corporate Change,
Forthcoming.
Harrison, R., J. Jaumandreu, J. Mairesse and Bettina Peters, (2005), "Does
Innovation Stimulate Employment? A Firm-Level Analysis Using Comparable Micro
Data from Four European Countries", Mimeo, January.
Kremp, E., J. Mairesse and P. Mohnen (2005), “The Importance of R&D and
Innovation for Productivity: A Reexamination in Light of the 2000 French Innovation
Survey” , Draft, November.
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
6
Special issue on
Empirical studies of innovation
in the knowledge driven economy
in Economics of Innovation and New Technology,
B. Hall and J. Mairesse, guest eds. (2006)
• Benavente, J-M., “The Role of Research and Innovation in
Promoting Productivity in Chile”.
• Heshmati, A. and H. Lööf, “Knowledge Capital and Heterogeneity
in Firm Performance. A Sensitivity Analysis”.
• Jefferson, G., B. Huamao, G. Xiaojing and Y. Xiaoyun, “R&D
performance in Chinese industry”.
• Van Leeuven G., and L. Klomp, “On the Contribution of Innovation
to Multi-Factor Productivity”.
• Duguet, E., “Innovation Height, Spillovers and TFP Growth at the
Firm Level: Evidence from French Manufacturing for Company
Performance”.
• ….
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
7
OTHER STUDIES (very very incomplete)
•
•
•
•
•
•
•
Klomp, L. and G. Van Leeuwen G. (2001), “Linking Innovation and Firm
Performance: A New Approach”, Journal of the Economics of Business, 8(3), 343364.
Lööf, H. and A. Heshmati (2002), “Knowledge Capital and Performance
Heterogeneity: A Firm-Level Innovation Study”, International Journal of Production
Economics, 76(1), 61-85.
Lööf, H., A. Heshmati, R. Apslund and S.O. Nås (2002), “Innovation and
Performance in Manufacturing Firms: A comparison of the Nordic Countries”,
mimeo.
Criscuolo, C. and J. Haskel (2003), “Innovations and Productivity Growth in the
UK: Evidence from CIS2 and CIS3”, CeRiBa discussion paper.
Galia, F. and D. Legros (2003), “Research and Development, Innovation, Training,
Quality and Profitability: Econometric Evidence from France”, mimeo.
Janz, N., H. Lööf, and B. Peters (2004), “Firm level Innovation and Productivity: Is
There a Common Story across Countries?”, Problems and Perspectives in
Management, 2, 184-204.
Parisi, M., F. Schiantarelli and A. Sembenelli (2006), “Productivity, Innovation and
R&D: Micro Evidence for Italy”, European Economic Review, 50, 2037–2061.
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
8
Objectives of the present study
• The present study (re-)investigates the links
between R&D, innovation and productivity in the
French Manufacturing and Services industries
• Estimating a variant of the “CDM” econometric
framework on the last two waves of the French CIS
for the periods (1998-2000) and (2002-2004).
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
9
Empirical strategy
• Our model is a variant of the one estimated by Griffith,
Huergo, Peters & Mairesse (2006), which is itself a
convenient simplification of the original CDM model.
• We try to improve on Griffith and al.(2006):
• by including indicators of “demand pull / technology push” that are
available only in the French CIS survey
• by estimating process and product innovation simultaneously
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
10
Data: the French CIS3 and CIS4
• Our data comes from the 3rd and 4th waves of the Community
Innovation Survey (CIS3 and CIS4).
• CIS : harmonised survey carried out by national statistical agencies
in all 27 EU Member States under the co-ordination of Eurostat (core
questionnaire + country-specific questionnaire).
• CIS3 covers the period 1998-2000, and CIS4 the period 2002-2004.
Both surveys gives information on:
• R&D activities, product and process innovation,
• intellectual property protection, effects of innovation,
• abandoned innovations, hampering factors.
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
11
Differences between CIS3 and CIS4
• In CIS3: information about investment in physical capital and quality
of labour force. This information is not available in CIS4.
• In CIS4, sampling is based on firms with 10+ employees (vs 20+
employees in CIS3).
• CIS3: representative of manufacturing industry only, whereas CIS4:
representative of both manufacturing and services.
• We will compare:
• manufacturing industry in France from 1998-2000 to 2002-2004.
• manufacturing and services industries in the recent period (2002-2004).
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
12
Comparison of CIS3 and CIS4
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
13
Comparison of manufacturing and services
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
14
Specification of the model
• Stage 1: R&D equations (Generalized Tobit model)
Pr(Continuous R&D)=(International market, Appropriability conditions, firm

size, Demand pulled / Technology pushed, 2-digit industry dummies)
Log(R&D intensity)=f(International market, Cooperation, Public funding,

Appropriability conditions, Demand pulled / Technology pushed,

Sources of information, 2-digit industry dummies)
• Stage 2: innovation production function (bivariate Probit model)
Pr(Product innovation)=(Appropriability conditions, firm size,

Predicted Log-R&D intensity, 2-digit industry dummies)
 Pr(Process innovation)=(Appropriability conditions, firm size,

Predicted Log-R&D intensity, 2-digit industry dummies)
• Stage 3: productivity equation (linear model)
Log(Labour Productivity)=f(predicted probabilities of: product only, process only,
product and process innovation, firm size, 2-digit industry dummies)
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
15
Main results: stage 1
We report
standard1 errors
in brackets.
Jacquesmarginal
Mairesse effects, with robustBarcelona,
July 2008
Stephane Robin
Significance:
* 10%, ** 5%, *** 1%
16
Main results: stage 2
Pred log(R&D intensity)
Product Innovation Equation
(1) CIS3 Manufacturing (2) CIS4 Manufacturing
0.53 (0.03)***
0.55 (0.02)***
(3) CIS4 Services
0.13 (0.01)***
Pred log(R&D intensity)
Process Innovation Equation
(1) CIS3 Manufacturing (2) CIS4 Manufacturing
0.30 (0.02)***
0.42 (0.02)***
(3) CIS4 Services
0.09 (0.01)***
Rho (correlation coeff.)
Log Likelihood
Wald test of H0: “j=0”
0.47 (0.03)***
-3609.21
2037.61***
0.42 (0.02)***
-5796.54
2669.12***
0.56 (0.02)***
-6625.10
2407.60***
We report marginal effects, with robust standard errors in brackets.
Significance: * 10%, ** 5%, *** 1%
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
17
Main results: stage 3
Productivity equation, dependent variable: Log of labour productivity
We report marginal effects, with robust standard errors in brackets.
Significance: * 10%, ** 5%, *** 1%
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
18
Robustness checks
• We replicated our analysis on CIS2 (1994-1996), w/ some
changes in the specification of the model.
• The results we obtain are consistent with those obtained on
CIS3 and CIS4.
• We also estimated our model on pooled CIS3 - CIS4 data
(1717 manufacturing firms), including a time effect and a
time*industry effect.
• The analysis on the pooled data confirms our previous
results.
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
19
Lagged analysis and growth of LP
• We estimated the model with CIS4 dependent variables, using both
CIS4 variables and lagged (CIS3) explanatory variables.
• In addition, we implemented two alternative specifications involving
the growth of labour productivity instead of its level.
• As before, we find that product innovation (on its own or combined w/
process innovation) is associated with a higher labour productivity.
Process innovation has no significant effect.
• However, we fail to find any significant effect of innovation on the
growth of labour productivity.
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
20
Conclusion
• We estimated a three-stage econometric model linking research,
innovation, and labour productivity using recent French data (CIS3
and CIS4).
• We compare two periods (1998-2000 and 2002-2004) and two
sectors in the recent period (manufacturing and services)
• We find that product innovation (on its own or combined with process
innovation) has a positive impact on labour productivity.
• The effect of process innovation appears to be weak.
• These results hold for the previous period (CIS2: 1994-1996) and are
robust when we use a matched CIS3-CIS4 sample
Jacques Mairesse
Stephane Robin
Barcelona, 1 July 2008
21
“Innovation is not the product
of logical thought”
(Albert Einstein, 1879-1955)