Abstract
Alternative models of productivity predict a range of its determinants besides that of research and development (R&D). We investigate the robustness of R&D vis-à-vis a dozen productivity determinants in a panel of 16 Organisation for Economic Co-operation and Development countries through panel cointegration, bootstrap simulations and extensive sensitivity tests. Domestic knowledge stocks, international knowledge diffusion and human capital remain robust across all measures. The cross-country differences in accumulated knowledge stocks and human capital appear to explain productivity differences across countries.
Similar content being viewed by others
Notes
We acknowledge that it is not always convincing to lump all other determinants of productivity except for the three forms of R&D capital stocks as non-R&D determinants. For example, it is hard to segregate ICT from its knowledge content, and similar arguments may apply in other cases. However, for the sake of convenience, and without any prejudice, we denote them as “non-R&D” determinants throughout.
Of course, productivity has been separately modelled as a function of a range of other variables, like cross-border flow of people, structural composition of the economy, to name but a few. However, our focus here is on those studies that augment R&D capital stocks by other (non-R&D) determinants.
We acknowledge that there are other channels of knowledge spillover, e.g., international student flows, indirect trade-related knowledge spillovers, information technology, telephone traffic, to name but a few.
For further details please refere to OECD Compendium of Productivity Indicators 2012, OECD Publishing. http://dx.doi.org/10.1787/9789264188846-en.
As expected, the R&D and non-R&D determinants shown in Eq. (3) are positively correlated except that employment rate (E) shows negative correlations with inward FDI (−0.107) and inward FDI-weighted foreign knowledge stocks (−0.019).
It is well known that a higher dimensional system produces more stable cointegrating relationships and that the joint inclusion of all relevant variables also makes estimates robust to potential misspecification bias—e.g., the exclusion of relevant variables.
It is often argued that R&D intensity measures capture cross-country differences in R&D activity. However, Khan and Luintel (2006) illustrate that intensity measures fail to capture the full extent of disparity in R&D activity across sample countries. Instead, they show that the mean levels of R&D activity better capture such differences; hence we use the mean levels of S b it , S p it and H it to capture cross-country heterogeneity.
As stated above, Pedroni’s panel estimates are essentially the mean of the country-specific FMOLS estimates of Phillips and Hansen (1990).
We do not report the results of panel unit root tests to conserve space, however, they are available in an earlier version of this paper (Cardiff Business School Working Paper Number E2012/15) at http://business.cardiff.ac.uk/research/working-papers/robustness-rd.
Since the computation of P m incorporates ICT in its measure of capital input, strictly speaking, there is no need to augment the benchmark model by ICT. We nevertheless report it for the sake of completeness as the other two measures of domestic productivity (P ec and P a) do not incorporate ICT in their measure of capital input. The significance of ICT in Tables 2 and 3 implies that the parameters of ICT and other constituent components of capital input (capital services) differ statistically in explainingy P m, which is not implausible.
For brevity, we do not report these results. Yet another group is the ICT and Services Sector (ICT and SER), which is addressed in Sect. 6 (last paragraph). R&D and human capital remain robust to the joint use of these two variables as well. We thank an anonymous referee for suggesting these specifications.
References
Aghion P, Howitt P (1992) A model of growth through creative destruction. Econometrica 60(2):323–351
Arrow K, Kurz M (1970) Public investment, the rate of return, and optimal fiscal policy. Johns Hopkins Press, Baltimore
Barro R, Sala-i-Martin X (2003) Economic growth. MIT Press, Cambridge, MA
Bassanini A, Scarpetta S (2002) Does human capital matter for growth in OECD countries? A pooled mean-group approach. Economics Letters 74(3):399–405
Basu S, Fernald J, Oulton N, Srinivasan S (2004) The case of the missing productivity growth: or, does information technology explain why productivity accelerated in the United States but not in the United Kingdom? In: Gertler M, Rogoff K (eds) NBER macroeconomics annual 2003. MIT Press, Cambridge, MA
Beck T, Levine R (2002) Industry growth and capital accumulation: does having a market- or bank-based system matter? J Financ Econ 64:147–180
Benhabib J, Spiegel M (2005) Human capital and technology diffusion. Chapter 13. In: Aghion P, Durlauf S (eds) Handbook of economic growth, vol 1A. Elsevier, Amsterdam, pp 935–966
Bernard A, Jensen B (1999) Exceptional exporter performance: cause, effect, or both? J Int Econ 47(1):1–25
Coe D, Helpman E (1995) International R&D spillovers. Eur Econ Rev 39(5):859–887
Coe D, Helpman E, Hoffmaister A (2009) International R&D spillovers and institutions. Eur Econ Rev 53(7):723–741
Council of Economic Advisors (2001) Annual report of the council of economic advisors, in economic report of the president, January
Devereux M, Smith G (1994) International risk sharing and economic growth. Int Econ Rev 35(3):535–550
Diamond D (1984) Financial intermediation and delegated monitoring. Rev Econ Stud 51(3):393–414
Eaton J, Kortum S (2002) Technology, geography, and trade. Econometrica 70(5):1741–1779
Engelbrecht HJ (1997) International R&D spillovers, human capital and productivity in OECD economies: an empirical investigation. Eur Econ Rev 41(8):1479–1488
Funk M (2001) Trade and international R&D spillovers among OECD countries. Southern Econ J 67:725–736
Gengenbach C, Palm FC, Urbain JP (2010) Panel unit root tests in the presence of cross-sectional dependencies: comparison and implications for modelling. Econ Rev 29(2):111–145
Goncalves S, White H (2005) Bootstrap standard errors for linear regression. J Am Stat Assoc 100:970–979
Gordon R (2000) Does the “New Economy” measure up to the great inventions of the past? J Econ Perspect 14(4):49–74
Gramlich E (1994) Infrastructure investment: a review essay. J Econ Lit 32(3):1176–1196
Greenwood J, Jovanovic B (1990) Financial development, growth, and the distribution of income. J Polit Econ 98(5):1076–1107
Greenwood J, Hercowitz Z, Krusell P (1997) Long-run implications of investment-specific technological change. Am Econ Rev 87(3):342–362
Griffith R, Harrison R, Van Reenen J (2006) How special is the special relationship? Using the impact of U.S. R&D spillovers on U.K. firms as a test of technology sourcing. Am Econ Rev 96(5):1859–1875
Griliches Z (1988) Productivity puzzles and R&D: another non-explanation. J Econ Perspect 2(4):9–21
Griliches Z, Mairesse J (1990) R&D and productivity growth: comparing Japanese and U.S. manufacturing firms. In: Hulten C (ed) Productivity growth in Japan and the United States. University of Chicago Press, Chicago
Grossman G, Helpman E (1991) Innovation and growth in the global economy. MIT Press, Cambridge, MA
Grossman H, Lucas R (1974) The macro-economic effects of productive public expenditures. Manch Sch Econ Soc Stud 42(2):162–170
Hadri K (2000) Testing for stationarity in heterogeneous panel data. Econom J 3(2):148–161
Hall B, Mairesse J (1995) Exploring the relationship between R&D and productivity in French manufacturing firms. J Econom 65(1):263–293
Hlouskova J, Wagner M (2006) The performance of panel unit root and stationarity tests: results from a large-scale simulation study. Econom Rev 25(1):85–116
Im KS, Pesaran H, Shin Y (2003) Testing for unit roots in heterogeneous panels. J Econom 115(1):53–74
Jankowski JE (2001) Measurement and growth of R&D within the service economy. J Technol Transfer 26(4):323–336
Kao C (1999) Spurious regression and residual-based tests for cointegration in panel data. J Econom 90(1):1–44
Kao C, Chiang MH, Chen B (1999) International R&D spillovers: an application of estimation and inference in panel cointegration. Oxf Bull Econ Stat (special issue): 691–709
Keller W (2004) International technology diffusion. J Econ Lit 42(3):752–782
Keller W, Yeaple S (2009) Multinational enterprises, international trade, and technology diffusion: a firm-level analysis of the productivity effects of foreign competition in the United States. Rev Econ Stat 91(4):821–831
Khan M, Luintel K (2006) Sources of knowledge and productivity: how robust is the relationship? OECD science, technology and industry working papers 2006/6, Paris
Kunsch H (1989) The jackknife and the bootstrap for general stationary observations. Ann Stat 17(3):1217–1261
Lee G (2006) The effectiveness of international knowledge spillover channels. Eur Econ Rev 50(8):2075–2088
Levine R (2002) Bank-based or market-based financial systems: which is better? J Financ Intermed 11(4):398–428
Levine R (2005) Finance and growth: theory and evidence. In: Aghion P, Durlauf SN (eds) Handbook of economic growth, vol 1. Elsevier, Amsterdam, pp 865–934
Lichtenberg FR, van Pottelsberghe de la Potterie B (1998) International R&D spillovers: a comment. Eur Econ Rev 42(8):1483–1491
Liu R, Singh K (1992) Moving blocks jackknife and bootstrap capture weak dependence. In: Lepage R, Billard L (eds) Exploring the limits of bootstrap. Wiley, New York
Lucas R (1993) Making a miracle. Econometrica 61(2):251–272
Luintel K, Khan M (1999) A quantitative reassessment of the finance-growth nexus: evidence from a multivariate VAR. J Dev Econ 60(2):381–405
Luintel K, Khan M (2004) Are international R&D spillovers costly for the US? Rev Econ Stat 86(4):896–910
Luintel K, Khan M (2009) Heterogeneous ideas production and endogenous growth: an empirical investigation. Can J Econ 42(3):1176–1205
Luintel K, Khan M, Arestis P, Konstantinos T (2008) Financial structure and economic growth. J Dev Econ 86(1):181–200
Maddala G, Wu S (1999) A comparative study of unit root tests with panel data and a new simple test. Oxf Bull Econ Stat 61(S1):631–652
Nadiri I, Mamuneas T (1994) The effects of public infrastructure and R&D capital on the cost structure and performance of U.S. manufacturing industries. Rev Econ Stat 76(1):22–37
O’Mahony M, Vecchi M (2005) Quantifying the impact of ICT capital on output growth: a heterogeneous dynamic panel model. Economica 72(288):615–633
OECD (2007) OECD science, technology and industry scoreboard 2007. OECD, Paris
OECD (2008) OECD compendium of productivity indicators 2008. Paris. http://www.oecd.org/dataoecd/6/3/40605524.pdf
Pasaran MH (2007) A simple panel unit root test in the presence of cross-section dependence. J Appl Econom 22(2):265–312
Pasaran MH, Shin Y (2002) Long-run structural modelling. Econom Rev 21(1):49–87
Pedroni P (1999) Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxf Bull Econ Stat 61(special issue): 653–670
Pedroni P (2001) Purchasing power parity tests in cointegrated panels. Rev Econ Stat 83(4):727–731
Pedroni P (2004) Panel cointegration: asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis. Econom Theory 20(3):597–625
Phillips P, Hansen B (1990) Statistical inference in instrumental variable regression with I(1) processes. Rev Econ Stud 57(1):99–125
Romer P (1989) Human capital and growth: theory and evidence. NBER working paper no. 3173
Romer P (1990) Endogenous technological change. J Polit Econ 98(5):S71–S102
Schreyer P (2000) The contribution of information and communication technology to output growth: a study of the G7 countries. OECD science, technology and industry working paper no 2000/2. Paris
Timmer M, Ypma G, Van Ark B (2003) IT in the European Union: driving productivity divergence? GGDC Research Memorandum GD-67
Van Ark B, Mekla J, Mulder N, Timmer M, Ypma G (2002) ICT investment and growth accounts for the European Union, 1980–2000. Final report on ICT and growth accounting for the DG Economics and Finance of the European Commission, Brussels
Zhu L, Jeon BN (2007) International R&D spillovers: trade, FDI, and information technology as spillover channels. Rev Int Econ 15(5):955–976
Acknowledgments
We thank Andrea Bassanini and Colin Webb for their help in providing data.
Author information
Authors and Affiliations
Corresponding author
Additional information
The views expressed in this paper are those of the authors, and not necessarily those of the Bank of England, the Cardiff Business School, or the World Intellectual Property Organization. The usual disclaimer applies.
Data Appendix
Data Appendix
Data on multifactor productivity (P m) are obtained from the OECD (2008). German data are available from 1991 only; we extrapolated pre-1991 data using the growth rate of the TFP series obtained from Timmer et al. (2003). However, dropping Germany from the sample does not change our results (see Sect. 6). Two further measures of TFP are used for robustness. The first one is our own measure of TFP (P a) calculated as: log P a = log GDP − 0.3 log K − 0.7 log L, where K is total net physical capital stock and L is total employment level. The second measure of TFP (P ec) is published by the EC. Domestic business-sector R&D capital stocks (S b) are calculated from business-sector R&D expenditure (\(E_{b}^{RD}\)), using the perpetual inventory method. Initial stock, \(S_{0}^{b}\), is calculated as:
where δ denotes the depreciation rate, g is the average annual growth rate of \(E_{b}^{RD}\) over the sample, and \(E_{b,o}^{RD}\) is the initial value of \(E_{b}^{RD}\). This method of computing capital stocks requires making assumptions about the average life of capital stocks and depreciation rates, which do not always capture the complexity of different types of capital assets and the different depreciation rates affecting them. The issues of taxes on capital assets and the price of capital further complicate the matter. However, this method is widely used in the literature on the grounds of cost and convenience, and we do the same. All R&D capital stocks are computed using 15, 10 and 5 % depreciation rates. Public-sector R&D expenditure (\(E_{p}^{RD}\)) is the total R&D expenditure of the government and higher education sectors. Public-sector R&D capital stocks (S p) are generated from public-sector R&D expenditure (\(E_{p}^{RD}\)), applying the same approach as in Eq. (7). Due to the lack of R&D deflators, R&D expenditure data are converted to constant prices by the GDP deflators. Initial capital stocks, \(S_{o}^{b}\) and \(S_{0}^{p}\), are generated for the earliest year for which R&D expenditure data are available (their availability ranges from the late 1960s to early 1980s).
We compute four different measures of foreign R&D capital stocks using bilateral imports, bilateral R&D collaborations and stocks of bilateral inward and outward FDI as weights, following the approach suggested by Lichtenberg and van Pottelsberghe (1998). In this framework, the import ratio-weighted foreign R&D capital stock (\(S^{fm}\)) is:
where, Y j denotes the GDP of country j and M ij is the imports of country i from country j; throughout, ‘t’ denotes time subscript. We use bilateral capital import ratios which include chemicals and related products (SITC 5), manufactured goods classified chiefly by material (SITC 6), machinery and transport equipment (SITC 7), and miscellaneous manufactured articles (SITC 8). Agro-industries and raw materials (SITC 0–4) are excluded. The bilateral R&D collaboration-weighted foreign knowledge stock (S fc) is:
where TP i is country i’s total patent applications and PC ij is its joint patent applications with countries J, both filed at the EPO. Data on patent applications are obtained from the EPO. We compute 15X23 matrixes of bilateral patent cooperation coefficients for each sample country. Likewise, foreign R&D capital stocks based on inward (S fI) and outward (S fO) FDI stocks are computed as:
where K j is country J’s capital stock, generated from non-resident fixed capital formation using the perpetual inventory method at an 8.0 % depreciation rate. FDI ij is country i’s FDI stock originating in country j; FDO ij is country J’s FDI stocks originating in country i. Data are expressed in constant 2,000 prices using the GDP deflator. The relevant weights for all foreign knowledge stocks are computed using 3-year moving averages to avoid yearly fluctuations. Human capital (H) is proxied by the average years of schooling of the 25–64 age group. Bassanini and Scarpetta (2002) kindly provided data for the period up to 2000; we extrapolate the last four observations. We acknowledge that this is only a rough measure of human capital, but we do not have any suitable alternative measures. Data on ICT investment consist of non-resident investment in hardware, communications equipment and software. They are expressed as a percentage of GDP. High-technology exports (X h) and imports (M h) are expressed as a percentage of total exports and imports, respectively. We follow the OECD’s (2007) definitions of high-technology items of trade, which include: pharmaceuticals (ISIC.2423); office, accounting and computing machinery (ISIC.30); radio, TV and communications equipment (ISIC.32); medical, precision and optical instruments (ISIC.33); and aircraft and spacecraft (ISIC.353). Service sector (SER) is measured as the value added of the service sector relative to GDP. The service sector consists of ISIC Rev.3 industries from 50 to 90. The proxy for the business cycle is the rate of employment (E).
Stocks of public infrastructure (Z) is generated from government’s fixed capital formation (I gov) using the perpetual inventory method (Eq. 7). I gov is converted to constant 2000 PPP US dollars using the fixed capital formation deflator. Measures of Z based on 3, 5 and 8 % depreciation rates are generated. Data on stocks of inward (F I) and outward (F O) FDI are published by the United Nations Conference on Trade and Development (UNCTAD) in current US dollars. They are converted to constant PPP dollars using GDP deflator and PPP exchange rates. Banking sector development is proxied by the ratio of private-sector credit by deposit money banks and other financial institutions to GDP (P K). Two measures of capital market development are the stock market capitalization-to-GDP ratio (S MC) and the stock market total value traded-to-GDP ratio (S MV). They are well-known measures of financial sector development (see Beck and Levine 2002; Luintel et al. 2008).
Data | Sources |
Multifactor productivity and ICT | Multifactor productivity database (OECD) |
Total factor productivity, capital stocks | AMECO database (European Commission) |
R&D expenditure | Research and development database (OECD) |
Human capital | Bassanini and Scarpetta (2002) |
High-technology exports and imports, and total exports and imports | STAN bilateral database (OECD) |
Stocks of total inward and outward foreign direct investment | UNCTAD’s foreign investment database |
Private-sector credit by deposit money banks and other financial institutions to GDP, stock market capitalization to GDP and stock market total value traded to GDP | World Bank |
Service sector value added | STAN indicators database (OECD) |
GDP, GDP deflator, total employment level, employment rate, PPP exchange rate, government fixed capital formation and its deflator, non-resident fixed capital formation and its deflator, investment | Analytical database (OECD) |
Bilateral imports | International trade by commodities statistics database (OECD) |
Patent applications at the European Patent Office | Patent database (OECD) |
Rights and permissions
About this article
Cite this article
Luintel, K.B., Khan, M. & Theodoridis, K. On the robustness of R&D. J Prod Anal 42, 137–155 (2014). https://doi.org/10.1007/s11123-013-0360-0
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11123-013-0360-0