Corresponding author: Han-Sol Lee ( 1042185141@pfur.ru ) © 2021 Non-profit partnership “Voprosy Ekonomiki”.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits to copy and distribute the article for non-commercial purposes, provided that the article is not altered or modified and the original author and source are credited.
Citation:
Lee H-S, Moseykin YN, Chernikov SU (2021) Sustainable relationship between FDI, R&D, and CO2 emissions in emerging markets: An empirical analysis of BRICS countries. Russian Journal of Economics 7(4): 297-312. https://doi.org/10.32609/j.ruje.7.77285
|
This paper empirically analyzes sustainable relations between inward FDI (IFDI), outward FDI (OFDI), the R&D expenditure ratio and CO2 emissions based on balanced panel data from the BRICS (namely, Brazil, Russia, India, China and South Africa) countries for the period 2003–2017. Generally, the results confirm a negative effect of IFDI and a positive effect of OFDI on the R&D expenditure ratio, both with statistical significance. Further exploration of the IFDI, OFDI and R&D impacts on CO2 emissions was based on an assumption that innovation development mitigates environmental pollution. The research outcome revealed positive associations between IFDI and the R&D expenditure ratio with CO2 emissions, showing the connection of environmental pollution to growth-focused national economic strategies. Based on these results, we recommend the following policies: (1) rethinking domestic industries protectionism trends and research support to enhance FDI spillover effects, (2) the drafting of New Development Bank specific environment-friendly investment programs aimed at innovation activities, and (3) looking into further easing the green technologies from developed countries.
foreign direct investment, FDI, research and development, R&D, environment pollution, panel analysis.
The modern economy is fueled by the flows of cross-country capital movements to obtain global competitiveness amongst differently endowed countries (
BRICS (namely, Brazil, Russia, India, China and South Africa), the leading developing countries, are also making efforts to increase the stability of their economies, but at the same time they are largely dependent on investment flows and innovative cooperation to achieve their goals. BRICS countries are striving for structural transformations that will ensure the modernization of their economies and the development of modern knowledge-intensive industries (
BRICS nations’ leaders have regularly confirmed that joint work in the field of science, technology and innovation has remained a priority area of cooperation between them. It would seem that BRICS states have great potential due to their complementary scientific basis, and their goal of technical development, as well as huge markets for high-tech products (
Research and development expenditure (R&D) of BRICS, 2003–2017 (% of GDP).
Country | 2003 | 2008 | 2013 | 2017 |
Brazil | 0.999 | 1.129 | 1.196 | 1.263 |
Russia | 1.286 | 1.044 | 1.025 | 1.107 |
India | 0.719 | 0.859 | 0.706 | 0.666 |
China | 1.120 | 1.446 | 1.998 | 2.145 |
South Africa | 0.760 | 0.888 | 0.725 | 0.832 |
In terms of FDI flows, the overall heterogeneous nature of the BRICS shows a general momentum toward the inflow side (
Foreign direct investment (FDI) of BRICS, 2003–2017 (billion U.S. dollars at current prices).
Country | FDI direction | 2003 | 2008 | 2013 | 2017 |
Brazil | Total | 10.352 | 76.832 | 90.855 | 90.226 |
Net Inflows | 10.123 | 50.716 | 75.211 | 68.885 | |
Net Outflows | 229.000 | 26.115 | 15.644 | 21.341 | |
Russia | Total | 17.653 | 130.446 | 155.725 | 65.314 |
Net Inflows | 7.929 | 74.783 | 69.219 | 28.557 | |
Net Outflows | 9.724 | 55.663 | 86.507 | 36.757 | |
India | Total | 4.920 | 62.663 | 29.918 | 51.056 |
Net Inflows | 3.682 | 43.406 | 28.153 | 39.966 | |
Net Outflows | 1.238 | 19.257 | 1.765 | 11.090 | |
China | Total | 66.357 | 228.277 | 363.899 | 304.377 |
Net Inflows | 57.901 | 171.535 | 290.928 | 166.084 | |
Net Outflows | 8.456 | 56.742 | 72.971 | 138.293 | |
South Africa | Total | 1.336 | 7.765 | 14.752 | 9.508 |
Net Inflows | 783.000 | 9.885 | 8.233 | 2.059 | |
Net Outflows | 553.000 | –2.120 | 6.520 | 7.449 |
The studied period of time has also been marked by a growing pressure for updating the development of global capital markets to new economic realities — the threat of climate change, water scarcity, general natural resource depletion and other factors related to human activity (
Country | 2003 | 2008 | 2013 | 2017 | ||||||||
Billion tons | Ratio (%) | Billion tons | Ratio (%) | Billion tons | Ratio (%) | Billion tons | Ratio (%) | |||||
World | 27.176 | 100 | 31.946 | 100 | 34.987 | 100 | 35.696 | 100 | ||||
BRICS | 7.759 | 28.55 | 11.349 | 35.53 | 14.400 | 41.16 | 14.804 | 41.47 | ||||
Brazil | 0.318 | 1.17 | 0.380 | 1.19 | 0.495 | 1.41 | 0.4846 | 1.36 | ||||
Russia | 1.525 | 5.61 | 1.637 | 5.12 | 1.619 | 4.63 | 1.646 | 4.61 | ||||
India | 1.059 | 3.90 | 1.463 | 4.58 | 2.033 | 5.81 | 2.457 | 6.88 | ||||
China | 4.452 | 16.38 | 7.375 | 23.09 | 9.797 | 28.00 | 9.751 | 27.32 | ||||
South Africa | 0.404 | 1.49 | 0.495 | 1.55 | 0.456 | 1.30 | 0.466 | 1.31 |
The economies of BRICS countries have had diverse development paths over the last two decades, predominantly based on import substitution and active industrial policy ideology. It can currently be stated that BRICS productive competencies have a slightly greater technological content and complexity compared to other developing countries (
The remaining paper is structured as follows. Section 2 covers the literature review on inward and outward FDI’s impacts on innovation development and environmental pollution. Section 3 presents the data and model specifications. Panel analyses are carried out in Section 4. Lastly, in conclusion, we provide policy directions related to FDI and innovation activities for the BRICS nations.
Positive impact of FDI on national economies can be direct and indirect: the former is measured by FDI’s contribution to market size expansion or economic growth, while the latter is achieved via spillover effects leading to production, technology and managerial innovations, which are used as a basis for economic development. The spillover effects of FDI are discussed in a variety of studies, but the results are mixed depending on the period, the proxy use for innovation, country of research, and the direction of the FDI (inward or outward).
It is evident that a multitude of studies have previously explored the relationship between inward FDI and productivity development.
In addition, there are studies which investigate the association of outward FDI with productivity growth.
Meanwhile, innovation development is estimated by various proxies other than productivity growth to verify its association with FDI.
In contrast, in a case study of Czech manufacturing firms,
Meanwhile, a growing awareness of environmental issues has led scholars to demonstrate a linkage between innovation and CO2 reductions in recent years. The impacts seem to be applied differently depending on the country of research, but are much positively sufficient in economically developed areas. For example,
On the other hand, multiple studies demonstrate the nexus between FDI and environment pollution to test whether the former is a factor in the degradation of the environment of host countries as a trade-off for economic growth, but the results of different studies do not reach common ground (
To conclude, the impact of (inward and outward) FDI would not appear to have any universal common ground and its results vary, depending on the country, period, industry and methodology. The same goes for the relationship of (inward and outward) FDI, innovation development and CO2 emissions. In this respect, our study is focused on the BRICS countries with fresh datasets by employing three different types of econometric techniques (which are FE, FE-GLS and FE2SLS-GLS).
The empirical analysis employed in this study utilizes country-level panel data of BRICS countries for the period of 2003–2017. Detailed data descriptions and data sources are shown in Appendix A. A descriptive data analysis for every variable — which are RnD, IFDI, OFDI, Ln(CO2), Capital, Human, Growth and TFP — is provided in Table
Variables | Mean | Std. dev. | Min. | Max. | Skewness | Kurtosis |
RnD | 1.097159 | 0.374005 | 0.665840 | 2.145120 | 1.301147 | 4.135171 |
IFDI | 0.023867 | 0.011676 | 0.002295 | 0.045543 | 0.019128 | 2.081229 |
OFDI | 0.012270 | 0.010231 | –0.007391 | 0.037735 | 0.741123 | 2.846303 |
Ln(CO2) | 7.177878 | 1.092386 | 5.760503 | 9.192213 | 0.497954 | 2.088286 |
Capital | 28.17224 | 10.18719 | 14.63039 | 46.66012 | 0.548989 | 1.701193 |
Human | 2.555156 | 0.455518 | 1.826568 | 3.403041 | 0.419059 | 2.225992 |
Growth | 4.117771 | 4.032594 | –7.827749 | 13.63582 | –0.314371 | 3.085264 |
TFP | 1.594437 | 3.141662 | –7.226635 | 10.52120 | –0.287746 | 3.546523 |
To construct a theoretical model,
I = f (L, K, IFDI, OFDI), (1)
where L and K denote labor and capital inputs and IFDI and OFDI denote inward and outward FDI, respectively. Our study assumes that both IFDI and OFDI are positive and important factors of innovation development. For estimations, the effects of FDI are measured both in year t in Eq. (2) (
RnDit = β 0 + β 1 IFDIit + β 2 OFDIit + β 3 Humanit + β 4 Capitalit +
+ β5 Growthit + γit + εit, (2)
RnDit = β 0 + β 1 IFDIit– 1 + β 2 OFDIit– 1 + β 3 Humanit + β 4 Capitalit +
+ β5 Growthit + γit + εit, (3)
where: RnD (a proxy for innovation development) — R&D expenditure ratio (% of GDP); IFDI — net inflows (U.S. dollars)/GDP (U.S. dollars); OFDI — net outflows (U.S. dollars)/GDP (U.S. dollars); Human — human capital index (based on years of schooling and returns of education); Capital — gross capital formation ratio (% of GDP); Growth — per capita GDP growth (annual %) (considering that the economic development is related to absorptive capability). In addition, i is an index for a country and t is an index for an year. γit represents country-fixed effects and εit is an error term. Our key variables are IFDI and OFDI, while the other remaining variables are utilized as control variables.
This study additionally derives Eq. (4) to verify the sustainable relationship between R&D, FDI and CO2 emissions. To consider the period necessary for absorbing innovation development through RnD and FDI, one-year-lagged variables are also tested according to Eq. (5). The equations are as follows:
Ln(CO2)it = β0 + β1 RnDit + β2 IFDIit + β3 OFDIit + β4 TFPit + γit + εit, (4)
Ln(CO2)it = β0 + β1 RnDit–1 + β2 IFDIit–1 + β3 OFDIit–1 +
+ β4 TFPit–1 + γit + εit (5)
where CO2 — a natural logarithm of CO2 emissions (million tons); TFP — growth (%). RnD, IFDI and OFDI are used as key independent variables. TFP is included as a control variable by employing the notion of the environmental Kuznets curve (EKC) (
In terms of methodology, at first this study attempted to apply country fixed effects models in the beginning. But country fixed effects models had cross-section correlations for all units as the p-value of the BP-LM test was below 0.05. Thus, to handle cross-section dependence, this study further employs fixed effects (FE) with GLS weights, cross-section Seemingly Unrelated Regression (SUR) (
As shown in Table
Variables | RnD | IFDI | OFDI | Ln(CO2) |
RnD | 1.000000 | |||
IFDI | 0.388752*** | 1.000000 | ||
OFDI | 0.095951 | 0.092755 | 1.000000 | |
Ln(CO2) | 0.650668*** | 0.270314** | 0.076729 | 1.000000 |
Dependent variable | RnD | Ln(CO2) | ||||
Eq. (2) | Eq. (3) | Eq. (4) | Eq. (5) | |||
RnD | – | – | 1.230577 | – | ||
RnD(–1) | – | 1.280438 | ||||
IFDI | 1.238627 | – | 1.203332 | – | ||
IFDI(–1) | – | 1.188551 | – | 1.260611 | ||
OFDI | 1.983311 | – | 1.013518 | – | ||
OFDI(–1) | – | 1.885459 | – | 1.018947 | ||
Human | 2.493244 | 2.509757 | – | – | ||
Capital | 2.069807 | 2.253906 | – | – | ||
Growth | 2.041140 | 1.891945 | – | – | ||
TFP | – | – | 1.095110 | – | ||
TFP(–1) | – | 1.089001 |
The panel analysis is carried out by using Eviews (ver. 11). We applied GLS weights (cross-section SUR) to FE and FE–2SLS regression analyses to handle cross-section dependence. As shown in Tables
Dependent variable: RnD | ||||||
(1) FE | (2) FE-GLS | (3) FE2SLS-GLS | (4) FE | (5) FE-GLS | (6) FE2SLS-GLS | |
IFDI | –4.894617** (1.949985) |
–2.903959*** (0.822258) |
–9.752749*** (2.006414) |
– | – | – |
IFDI(–1) | – | – | – | –4.359677** (1.777275) |
–3.490788*** (0.713831) |
–10.299990*** (2.291407) |
OFDI | 2.014982 (2.656647) |
2.346997* (1.181310) |
7.933830* (4.277609) |
– | – | – |
OFDI(–1) | – | – | – | 2.511971 (2.507802) |
2.512381*** (0.831396) |
3.117150 (6.410033) |
Human | 0.332282** (0.128733) |
0.321836*** (0.054693) |
0.391716*** (0.079459) |
0.313774** (0.131008) |
0.359010*** (0.043833) |
0.396215*** (0.073605) |
Capital | 0.023553*** (0.006879) |
0.014661*** (0.002976) |
0.014489*** (0.004954) |
0.021704*** (0.006903) |
0.017632*** (0.002344) |
0.019966*** (0.004108) |
Growth | –0.009277 (0.007064) |
–0.012334*** (0.002854) |
–0.006256 (0.004706) |
–0.016104** (0.006529) |
–0.013543*** (0.002057) |
–0.013942*** (0.003752) |
Constant | –0.285118 (0.390809) |
–0.046905 (0.164213) |
–0.149674 (0.244079) |
–0.176250 (0.422350) |
–0.208345 (0.124845) |
–0.205794 (0.228272) |
BP-LM test stat (Prob) | 52.36147 (0.0000) |
7.002344 (0.7252) |
5.647129 (0.8440) |
51.02263 (0.0000) |
4.362508 (0.9295) |
6.218088 (0.7966) |
Prob (F-statistics) | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 |
Adjusted R2 | 0.845797 | 0.941000 | 0.910864 | 0.876536 | 0.968328 | 0.940732 |
N | 75 | 75 | 70 | 70 | 70 | 65 |
OFDI presents positive coefficients at 10% significance levels in (2) FE-GLS and (3) FE2SLS-GLS regression models. However, the coefficient signs of 1 year-lagged OFDI present mixed outcomes: in (5) FE-GLS regression model, it shows a positive coefficient at the 1% significance level; while, in (4) FE and (6) FE2SLS-GLS regression model, its statistical significance disappears. This means that outward FDI contributes to innovation development to some extent, although its positive effects could not offset the negative effects from inward FDI. The result is in line with those in previous studies (
In terms of control variables, the human capital index and the gross capital formation ratio consistently show a statistically positive relationship to the R&D expenditure ratio as expected from Eq. (1). Meanwhile, it is worth noting that Growth shows a statistical negative association with RnD. This implies that income growth of the BRICS nations does not increase national R&D expenditure. There may be different explanations for this result, depending on the way it is approached. As noted above, BRICS is not a homogenous set of countries, and in the last decade Brazil, Russia, and South Africa have suffered a drop in GDP per capita, while India and China enjoyed a rather stable growth of this indicator. This might imply that in some countries the spending on R&D could not keep up with the rate of growth, while in others it suffered from anti-crisis cost cutting policies. However, another possible justification for this outcome is that innovation spending does not necessarily grow in parallel with national per capita wealth because the latter is situational and cannot be precisely predicted annually, while R&D spending in the BRICS nations is largely government-based, and is subject to long-term budget-planning.
Table
The relationships between IFDI, OFDI, RnD and Ln(CO2) in BRICS, 2003–2017.
Dependent variable: Ln(CO2) | ||||||
(7) FE | (8) FE-GLS | (9) FE2SLS-GLS | (10) FE | (11) FE-GLS | (12) FE2SLS-GLS | |
RnD | 0.668296*** (0.117383) |
0.659407*** (0.030634) |
0.754554*** (0.073808) |
– | – | – |
RnD(–1) | – | – | – | 0.590310*** (0.112700) |
0.550580*** (0.046211) |
0.606109*** (0.069616) |
IFDI | 3.534820* (2.056413) |
2.329441*** (0.602115) |
15.50278*** (3.984779) |
– | – | – |
IFDI(–1) | – | – | – | 3.948807** (1.932821) |
2.740672*** (0.677559) |
12.875840*** (2.892095) |
OFDI | –3.064185 (2.845193) |
–2.503244*** (0.846378) |
1.980097 (11.18935) |
– | – | – |
OFDI(–1) | – | – | – | –3.340740 (2.671809) |
–3.150517*** (0.994707) |
8.190162 (7.482044) |
TFP | –0.002046 (0.007835) |
0.004283* (0.002499) |
–0.010560 (0.008307) |
– | – | – |
TFP(–1) | – | – | – | –0.001513 (0.007180) |
0.004987* (0.002679) |
–0.010919 (0.008294) |
Constant | 6.401147*** (0.148659) |
6.422693*** (0.035712) |
5.976517*** (0.201843) |
6.502002*** (0.139770) |
6.561761*** (0.055685) |
6.143336*** (0.135671) |
BP-LM test stat (prob) | 54.32741 (0.0000) |
8.050668 (0.6239) |
8.317877 (0.5978) |
40.35681 (0.0000) |
4.476235 (0.9233) |
13.33697 (0.2054) |
Prob (F-statistics) | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 |
Adjusted R2 | 0.978863 | 0.997018 | 0.978477 | 0.982666 | 0.996397 | 0.977323 |
N | 75 | 75 | 70 | 70 | 70 | 65 |
However, the negative relation between R&D expenditures and CO2 emissions in BRICS leaves a lot of room for discussion and further research. The highest potential is within the possible connection between the industrial output of BRICS nations and planned R&D budgets. As R&D expenses tend to be funded by BRICS governments, their increase would be politically more justified in times of industrial growth, leading to more export and taxation income. However, due to the largely non-emission-friendly character of the BRICS industries, the growth of industrial output leads to a rise in CO2 emissions, leading to a consecutive relation between R&D expenses and the latter. However, due to the diverse character of the BRICS nations, this hypothesis needs to be researched on a more detailed basis per country.
The coefficient sign of OFDI and TFP does not match between FE, FE-GLS and FE2SLS-GLS typed regression models. Their statistical significance only appears in FE-GLS typed models regardless of year-lagging. It indicates the absence of statistical robustness in outcomes.
This paper conducted two strands of studies based on panel datasets from BRICS nations for the 2003–2017 time period by employing FE, FE-GLS and FE2SLS-GLS regression analyses. Firstly, the spillover effects of inward and outward FDI on innovation development (measured by R&D expenditure ratio) were examined. The results of this panel data analysis disprove spillover effects from FDI inflows, while simultaneously supporting them with regard to FDI outflows. It is likely that FDI inflows (in the case of mergers and acquisitions) worsen the R&D activities of host countries’ firms as foreign firms relocate important R&D activities to their headquarters (
Secondly, taking into account growing environmental challenges and sustainable development goals, the authors further investigated a nexus of (inward and outward) FDI and R&D expenditure ratio to CO2 emissions, including TFP growth as a control variable. Contrary to the authors’ assumption regarding the mitigation effects of innovation development on CO2 emissions based on the existing literature (
The above results from empirical analysis allow us to draw a few policy implications. First, the empirical results above — the negative association of FDI inflows with the R&D expenditure ratio — point to the need for rethinking traditional protectionism patterns in domestic industries and research areas. The statistical relationship between a level of protectionism or privatization and FDI spillover effects has been proven in
Second, it is empirically demonstrated that FDI inflows increase CO2 emissions and this can be explained by a lack of BRICS’s colossal program to modernize domestic basic industries up to EU sustainability and emission standards. Thus, it is important for the policymakers to consider the possibility of launching a specific environment-friendly international investment program. The latter could be formed through the structure of BRICS-associated New Development Bank and would fund innovation projects specifically dealing with the environmental restructuring of BRICS industry.
The last point concerns the recently drafted European Green Deal, which aims to turn the EU into the first emission-free region and at the same time prepares a set of economic fines on products coming from emission-based economies (BRICS nations included). However, draining financial resources from developing countries in such a way would obviously fail to create the sufficient incentive or capacity in the BRICS industries to invest in actual modernization. The results of empirical analysis confirmed the resource-based view of internationalization in terms of FDI outflows. In this respect, reaching an agreement with the BRICS nations on relieving the intellectual property regulations in green technologies transfer seems to be a much more efficient activity on the part of the EU in moving toward an emission-free planet.
Variables | Descriptions | Sources |
RnD | R&D ratio (% of GDP) |
|
IFDI | FDI net inflow (U.S. dollars) / GDP (U.S. dollars) |
|
OFDI | FDI net outflow (U.S. dollars) / GDP (U.S. dollars) |
|
Ln(CO2) | CO2 emission (million tons) |
|
Capital | Gross capital formation (% of GDP) |
|
Human | Human Capital Index (based on years of schooling and returns to education) |
|
Growth | GDP per capita growth (annual %) |
|
TFP | TFP growth (%) |
|