Abstract
Competition can be good or bad for innovation by firms. On the one hand it stimulates firms to innovate in order to escape competition, on the other hand it hampers firms to reap additional profits from innovation. The recent literature has embraced a model that describes an inverted-U shape relationship between competition and innovation at the industry-level. With the Price Cost Margin and Profit Elasticity as measures of competition, we find evidence supporting this prediction using industry data from the Dutch National Accounts. Moreover, we test the non-linear relation at the micro-level, with special attention for the role of the distribution of technology within industries. We find evidence that there is a threshold for this ‘technology spread’ at which the (marginal) effect of competition on innovation activity by firms turns from positive to negative.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Additional argumentation and an overview of related papers can be found in Aghion and Griffith (2008).
Our results serve as a complementary country specific analysis in the project ‘Market incentives to innovation’ of the Working Party for Industrial Analysis (WPIA) of the OECD.
Chapter 3 of Aghion and Griffith (2008) presents a different model that also results in an inverted-U relationship but under quite different assumptions. See also Vives (2004) and Rauch (2008) for other models giving rise to an inverted-U. Each of these models has its own implications for empirical testing, so that our results do not necessarily relate to these models as well.
By construction the leader does not innovate in an unlevelled industry (Aghion et al. 2005, p. 713). Or, more precisely, it can innovate but the laggard will take over the old leading technology. Since innovation occurs step-by-step in the model, the new situation is the same as the old, so the leader has no incentive to innovate. (The leader does not seem to take into account the possibility of catching-up of the laggard, which could provide an incentive to stay one step ahead. In fact, an increase in competition decreases the threat of a laggard catching up. A model in which such strategic innovation decisions are considered is presented in Aghion and Griffith 2008, chapter 3, but in the context of an incumbent firm versus a potential entrant.)
In this context it is interesting to look at the non-parametric evidence summarized in figure II of Aghion et al. When we look at the spline it looks more like an M than an inverted-U. In particular, if one disregards the outer tails of the M—which have a relatively low number of observations according to the figure—the resulting shape is convex rather than concave.
We focus on the identification from a cross-sectional or panel data set. Similar caveats apply to the identification of the inverted-U by industry from a time-series, however. One then needs a sufficiently long time-series with observations for the industry distributed well over the inverted-U. That is, periods of low and high competition must have occurred.
The semi-log specification accounts for the fact that the R&D intensity is skewed.
Industry dummies are not included explicitly in X since we use within-regression. In principle, this implicitly takes account of industry-specific effects (i.e. the industry dummies drop out in the within-transformation). Note that, if included, the coefficients on the industry dummies are identified through observations where firms have switched industry only, so that the pertinent industry dummy has some time variation.
This measure for spread is calculated at the 3-digit level. Although not pursued in this study, it is also possible for the macro regression to interact competition with a measure of spread at the P42 level.
As mentioned in footnote 3 there are other models yielding an inverted-U shape relation.
Technically speaking, this is true if Cov(W it ,u it ) is small relative to Cov(W it ,α i ).
The selection of industries corresponds to what in the Dutch Growth Accounts is called the ‘commercial sector’ (NACE codes 01 to 67, 72–74, 804, and 85–93), see CBS (2009). Research and development (NACE 73) is excluded since it is highly atypical in this context. We also exclude Mining and quarrying (NACE 10–14) because the data do not allow to calculate meaningful competition measures for this industry.
At the firm-level, costs of capital are proxied by total depreciation costs.
An innovator is a firm that has developed a product or a process innovation, or has a project aimed at this that is still ongoing or that has been abandoned at an earlier stage. The innovation can have been developed by the firm itself, in cooperation, or entirely by third parties.
We use deflators for gross output, value added, and intermediate inputs from the EUKLEMS-database (see www.euklems.net). Deflators for capital and labour cost are derived at the 2-digit level from labour costs and user-cost of capital figures in current and constant prices (unpublished estimates made for purposes of the Dutch Growth Accounts, recalculated to the 2-digit level, which is the lowest level available for these price indices). In all cases base year 2000 was chosen.
Because the various specifications for the firm-level regressions do not have exactly the same number of observations we use the largest estimation sample, which is for the specification with 1-PCM in year t as competition measure.
In section 6.3 we present results using also non-R&D performers to see whether our results for the R&D firms extend to all firms.
We have experimented with instrumental variable techniques (using lagged values of competition as instruments for current values), but these exercises did not yield satisfactory results. Most coefficients turned insignificant in this case.
We also estimated the industry regression making this sample split, but the number of observations appeared too low to produce satisfactory results.
As in the macro regressions we also estimated specifications with lagged competition (and distance-to-frontier) variables but in this case we could not identify any significant effects. This could be due to the fact that the timing assumption is not appropriate. When using instrumental variable estimation with lagged competition and distance-to-frontier as instruments, the results are again mostly insignificant. Thus, we cannot rule out the possibility that the firm-level results are subject to endogeneity bias. On the other hand, as argued above, taking account of the fixed-effect through within-estimation mitigates the effect of endogeneity if the correlation of the endogenous variable(s) with the fixed effect is relatively large compared to the correlation with the idiosyncratic error. Moreover, for the competition measures at the industry-level, it can be argued that simultaneity of competition and firm-level R&D, as well as possible feedback effects of R&D on competition, is less of an issue.
Note that in the second column, the effect of competition is positive when it is above (rather than below) the threshold since we find that β 2 > 0. Because the implied threshold is negative, the DTF is always above the threshold. Moreover, it can be argued that for columns (2) and (4), the implied threshold is in fact infinity because β 2 does not differ significantly from 0.
There is no fixed-effect equivalent of the Heckman model available. Instead we first apply the within transformation to our model to get rid of the fixed effects, and subsequently estimate the transformed equation with the Heckman procedure.
References
Aghion P, Griffith R (2008) Reconciling theory and evidence. MIT, Cambridge
Aghion P, Bloom N, Blundell R, Griffith R, Howitt P (2005) Competition and innovation: an inverted-U relationship. Q J Econ 120:701–728
Ahn S (2002) Competition, Innovation and Productivity Growth: A Review of Theory and Evidence, OECD Economics Department Working Papers, No. 317, OECD Publishing, doi:10.1787/182144868160
Amemiya T (1984) Tobit models: a survey. J Econ 24:3–62
Askenazy P, Cahn C, Irac D (2008) Competition, R&D, and the cost of innovation, Working Paper No. 2008-32, Paris School of Economics
Boone J (2008) Competition: theoretical parameterizations and empirical measures. J Inst Theoretical Econ 164:587–611
Boone J, van Ours J, van der Wiel H (2007) How (not) to measure competition, working paper, CENTER
Bos JWB, Kolari JW, van Laroen RCR (2009) Competition and innovation: evidence from financial services, Tjalling C. Koopmans Research Institute, Discussion Paper 09-16
CBS (2009) De Nederlandse Groeirekeningen 2008, CBS, Voorburg/Heerlen, http://www.cbs.nl/NR/rdonlyres/F2630959-6267-4A67-AF0D-D8A8EDEE3007/0/2008p40pub.pdf (in Dutch)
Dixit A, Stiglitz J (1977) Monopolistic competition and optimum product diversity. Am Econ Rev 67(3):297–308
Friesenbichler KS (2007) Innovation and market concentration in Europe’s mobile phone industries: evidence from the transition from 2G to 3G’, WIFO Working Papers, no. 306, November 2007
Griffith R, Redding SJ, van Reenen J (2004) Mapping the two faces of R&D: productivity growth in a panel of OECD industries. Rev Econ Stat 86(4):883–895
Heckman JJ (1979) Sample selection bias as a specification error. Economet-rica 47(1):153–161
Jorgenson D, Griliches Z (1967) The explanation of productivity change. Rev Econ Stat 34(3):249–283
Lind JT, Mehlum H (2010) With or without U? The appropriate test for a U-shaped relationship. Oxf Bull Econ Stat 72(1):109–118
Mundlak Y (1963) Estimation of Production and Behavioral Functions from a Combination of Cross-Section and Time Series Data. In: Christ C et al (eds) Measurement in economics: studies in mathematical economics and econometrics in memory of Yehuda Grunfeld. Stanford University Press, Stanford
Nickell SJ (1996) Competition and corporate performance. J Polit Econ 104(4):724–746
Olley SG, Pakes A (1996) The dynamics of productivity in the telecommunications equipment industry. Econometrica 64(6):1263–1297
Poldahl A, Tingvall PG (2006) Is there really an inverted-U shaped relation between competition and R&D? Econ Innovat New Tech 15(2):101–118
Polder M, Veldhuizen E, van den Bergen D, van der Pijll E (2009) Micro and macro indicators of competition: comparison and relation with productivity change, CBS Discussion Paper 09024, http://www.cbs.nl/NR/rdonlyres/B399FD48-78D2-42AD-B30C-C4CB6E14049B/0/200924x10pub.pdf
Puhani P (2000) The Heckman correction for sample selection and its critique. J Econ Surv 14(1):53–68
Rauch F (2008) An explanation of the inverted-U relationship between competition and innovation, Working Paper No. 0813, Department of Economics, University of Vienna
Salop S (1977) The noisy monopolist: imperfect information, price dispersion, and price discrimination. Rev Econ Stud 44:393–406
Scherer F (1967) Market structure and the employment of scientists and engineers. Am Econ Rev 57:524–531
Tanriseven M, van Rooijen-Horsten M, van den Bergen D, de Haan M (2007) Research and Development Statistics, Paper prepared for the Intellectual Capital Congress at the Inholland University of Professional Education, the Netherlands, Haarlem, 3–4 May, 2007
Tingvall PG, Karpaty P (2008) Service-sector competition, innovation and R&D, SSE/EFI Working Paper No. 702, October 2008
United Nations (2008) System of National Accounts, European Commission, International Monetary Fund, Organisation for Economic Co-operation and Development, World Bank, Brussels/Luxembourg, New York, Paris, Washington D.C.
Van Rooijen-Horsten M, Tanriseven M, de Haan M (2007) R&D sattelite accounts in the Netherlands: A progress report, paper prepared for the OECD Working Party on National Accounts, Paris 3–5 October 2007
van Rooijen-Horsten M, van den Bergen D, Tanriseven M (2008) Intangible capital in the Netherlands: a benchmark, Discussion Paper 08001. Statistics Netherlands, Voorburg/Heerlen
Vives X (2004) Innovation and competitive pressure, CEPR Discussion Paper 4369
Wooldridge JM (2002) Econometric analysis of cross section and panel data. MIT, Cambridge
Acknowledgements
We thank Dirk van den Bergen and George van Leeuwen for helpful discussions, Leo Soames for comments on the draft version, and Danny Pronk and Murat Tanriseven for their help with respect to the R&D data.
Author information
Authors and Affiliations
Corresponding author
Additional information
The views expressed in this paper are those of the authors and do not necessarily reflect any policy by Statistics Netherlands. The empirical work includes data that are the results of calculations by the authors, and do not necessarily equal officially published statistics.
Appendix. Response and coverage
Appendix. Response and coverage
Rights and permissions
About this article
Cite this article
Polder, M., Veldhuizen, E. Innovation and Competition in the Netherlands: Testing the Inverted-U for Industries and Firms. J Ind Compet Trade 12, 67–91 (2012). https://doi.org/10.1007/s10842-011-0120-7
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10842-011-0120-7