Examining the FDI–Growth Nexus in BRICS: Panel Data Evidence

Abstract

The purpose of this article is to examine the relationship between foreign direct investment (FDI) and economic growth in Brazil, Russia, India, China and South Africa (BRICS) economies, which are considered to be the fastest-growing economies and dominant players in the global investment landscape. In order to assess the relationship between the dependent variable (economic growth) and explanatory variables (FDI inflows and other growth determinants), we analyse a 32-year panel data starting from 1987 to 2018 using feasible generalised least squares (FGLS) method. The article found a significant positive FDI impact on economic growth in BRICS. However, exports, human capital and inflation (macroeconomic instability) exert a negative impact on economic growth of BRICS, whereas domestic investments exert a positive impact on growth.

JEL Codes: F21, F43, C23, O47

Keywords

FDI Economic growth GDP Panel data BRICS

I. Introduction

Over the past few decades, the global economy has witnessed a paradigm shift, having moved away from isolated national economies, towards an interconnected global economic system. Owing to globalisation, the barriers to international trade and investment have reduced to a considerable extent in integrated world economy order. One of the most important factors that decides a country’s integration with the global economy is foreign direct investment (FDI). FDI occurs when a firm invests directly in facilities to produce or market a product in a foreign country (Hill, 2011). World over, the opening up of economies and increase in the rate of globalisation have led to the belief that FDI is an important instrument of development for any economy and a source of aggregate demand and real economic growth for both developing and developed nations, as FDI augments the level of investment or capital stock in the host country. Moreover, FDI creates additional employment opportunities due to the creation of new economic activities and helps in the transfer of intangible assets like technological know-how and managerial skills to the host country, and stimulates the procurement of other benefits such as new processes, products and technologies. Therefore, the host countries are sure to benefit from FDI through its spillover effects, which may take the form of technology transfers, introduction of new processes and transfer of managerial skill. Since FDI helps the economies in their developmental effort, many countries have gradually removed the restrictions pertaining to FDI. According to the United Nations, about 80% of the 1,440 changes made globally in the laws governing FDI over the period from 2000 to 2013 have created an enabling environment for FDI. In order to attract FDI, both developing and developed countries have increasingly offered incentives in the form of tax holidays, infrastructural subsidies, etc., to multinational enterprises (Alfaro et al., 2009).

Although researchers have found mixed evidence from time to time, regarding the role of FDI in accelerating growth, the countries have increasingly made efforts to attract FDI globally. The Brazil, Russia, India, China and South Africa (BRICS) nations, which are considered to be the fastest emerging economies due to their increasing growth rates, have also attracted huge FDI inflows over the past few decades. It, therefore, interests us to undertake the present study to ascertain the FDI–growth linkage in these economies, given their growth trajectory and dominance in the international investment arena. In addition to FDI, the study includes other explanatory variables, such as gross capital formation (domestic investment), labour force, exports, financial development, human capital and economic stability, in order to gain a better insight on promoting growth in BRICS.

II. Review of Literature

The impact of FDI on the economic growth of host nations has been extensively studied by the researchers around the globe. However, the growth literature has produced conflicting evidences regarding the FDI–growth relationship. Many researchers have empirically proven the positive impact of FDI on growth (Bouchoucha & Ali, 2019; de Mello, 1999; Hansen & Rand, 2006; Tiwari & Mutascu, 2011; Yao, 2006). Feridun and Sissoko (2011) examined the linkages between FDI and gross domestic product (GDP) per capita in Singaporean economy and found a unidirectional causal nexus from FDI to GDP. Dritsaki et al. (2004) found a long-run relationship among FDI, exports and economic growth in Greece. A bidirectional causal nexus was found between exports and economic growth, and a unidirectional causality was found from FDI to economic growth. Szkorupová (2014) found that both FDI and exports have a positive impact on the economic growth of Slovakia. Mencinger (2003) investigated the FDI–growth nexus in a set of eight European Union countries, and the results revealed a negative causal relationship between FDI and growth. Herzer (2012) also contradicted the earlier studies, revealing a positive growth impact of FDI, and found that the effect of FDI on growth is negative on average. However, the author has also mentioned that the growth impact of FDI may not continue to be negative in future as well. In this regard, it is suggested to policymakers to undertake economic reforms, for improving the political and economic environment, in order to attract FDI in a better manner. Another strand of literature argues that FDI has neither positive nor negative impact on economic growth of host countries (Akpan & Eweke, 2017; Ledyaeva & Linden, 2006; Lian, & Ma, 2013). On evaluating these studies spanning over various time periods, it can be concluded that the varied results found could be due to the difference in research methods adopted and different sample countries used, which possess unique country characteristics. Some studies propounded that FDI does not have an independent influence on growth, but the growth impact of FDI depends upon the host nations’ special characteristics (Borensztein et al., 1998; Cao & Jariyapan, 2012; Carkovic & Levine, 2002). In the growth literature, there is ample evidence suggesting the role of financial development in the economic growth of a country. Well-developed financial markets lead to rise in growth rates due to better allocation of capital and decrease in transaction costs. Beck et al. (2000) and King and Levine (1993a, 1993b) found in their empirical studies that efficient financial systems lead to economic growth and development. Alfaro and Charlton (2007) found that FDI does not affect the growth independently but in presence of absorptive capacities existing in host economies like financial development and availability of human capital. In order to reap the harvest of FDI, it is recommended to policymakers to develop the financial markets in their countries (Alfaro et al., 2004, 2009). Other researchers also revealed that growth performance of FDI is positive for such countries, which have efficient financial systems in place (Alfaro & Chauvin, 2020; Hermes & Lensink, 2003; Kelly, 2016; Ljungwall & Li, 2007). However, Kholdy and Sohrabian (2005) found that FDI has no impact on economic growth, even in those countries which have well-developed financial systems. It is also widely held that economic stability is good for the economic growth of a country. In Middle East and North Africa (MENA) countries, Jallab et al., (2008) found a positive impact of FDI on growth in the presence of macroeconomic stability. Alguacil et al. (2011) also found economic stability to significantly influence the FDI–growth relationship. Abdelmalki et al. (2012) investigated the role of economic stability in the FDI–growth relationship and found that the positive growth impact can be seen only when the inflation does not exceed a specific level. The macroeconomic instability proxied by inflation has a negative effect on both investments and growth (Bleaney, 1996). Prüfer and Tondl (2008) revealed a strong association between FDI and productivity growth in Latin America, which depends on an efficient legal system and economic stability. Mehic et al. (2013) found the macroeconomic stability to be a strong driver of growth in European countries.

III. Model, Data and Estimation Methods

Model Specification

In order to specify our model, we begin with the basis production function as follows:

Y = f (K,L)

(1)

where Y is the GDP growth, K denotes the gross capital formation (domestic investment) and L specifies the labour measured by the labour force.

We have expanded the above-mentioned function (Equation [1]) according to the new growth theory following Barro and Sala-i-Martin (1995), who propound that there are channels in addition to those specified in Equation (1) (labour and capital), which can help in promoting growth of an economy. Given such a focus, Mankiw (2004) argues that foreign trade affects the growth of a country, and Grossman and Helpman (1991) found that exports exert a positive influence on growth in different ways. We, therefore, expand Equation (1) to include exports (denoted by X) as follows:

Y = f (K, L, X)

(2)

Ögütçü (2002) states that FDI is a major driver of growth for developing economies. Evans et al. (2002) studied the contribution of financial development and human capital to economic growth in a set of 82 countries and found both variables to be important for growth. Mehic et al. (2013) found macroeconomic stability to be a main driver of growth. Bengoa and Sanchez-Robles (2003) revealed that the FDI receiving country requires a threshold level of human capital and economic stability to reap the growth benefits of FDI. Thus, Equation (2) is expanded by adding the variables such as FDI, financial development (liquid liabilities as proxy), human capital and macroeconomic stability (inflation as proxy) as follows:

Y = f (K, L, X, FDI, LL, HC, INF)

(3)

where FDI denotes foreign direct investment inflows, LL denotes liquid liabilities, HC stands for human capital and INF is inflation.

Data

The present study uses secondary data for the BRICS countries over the period from 1987 to 2018. The data for the dependent variable and independent variables have been taken from the World Investment Reports published by the World Bank annually. The dependent variable for the study is gross domestic product growth (GDPG) and the explanatory variables are FDI inflows, gross capital formation (domestic investment), labour force, exports, financial development, human capital and economic stability. Liquid liabilities is the proxy for financial development (Kholdy & Sohrabian, 2005), gross enrolment ratio for human capital (Abbas & Mujahid-Mukhtar, 2001) and inflation is the proxy for macroeconomic instability (Abdelmalki et al., 2012). The operational definition of variables used in the present study is as follows:

Economic Growth

The GDPG (economic growth) refers to annual percentage growth rate of GDP at market prices, based on constant local currency. Aggregates are based on constant 2010 US dollars. The GDP is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources.

Foreign Direct Investment Inflows

FDI inflows are the net inflows of investment to acquire a lasting management interest (10% or more of voting stock) in an enterprise operating in an economy other than that of the investor. It is the sum of equity capital, reinvestment of earnings, other long-term capital and short-term capital, as shown in the balance of payments. The data for FDI inflows in this study show net inflows (new investment inflows less disinvestment) in the reporting economy from foreign investors divided by GDP.

Gross Capital Formation

Gross capital formation (gross domestic investment) consists of outlays on additions to the fixed assets of the economy plus net changes in the level of inventories. Fixed assets include land improvements (fences, ditches, drains, etc.); plant, machinery and equipment purchases; and the construction of roads, railways and the like, including schools, offices, hospitals, private residential dwellings, and commercial and industrial buildings. Inventories are stocks of goods held by firms to meet temporary or unexpected fluctuations in production or sales, and ‘work in progress’.

Labour Force

Labour force comprises people who are aged 15 and older who supply labour for the production of goods and services during a specified period. It includes people who are currently employed and people who are unemployed but seeking work as well as first-time jobseekers.

Exports

Exports of goods and services represent the value of all goods and other market services provided to the rest of the world. They include the value of merchandise, freight, insurance, transport, travel, royalties, license fees and other services, such as communication, construction, financial, information, business, personal and government services. They exclude compensation of employees and investment income (formerly called factor services) and transfer payments.

Financial Development

The liquid liabilities used as proxy for financial development is also known as broad money or M3. The liquid liabilities are the sum of currency and deposits in the central bank (M0), plus transferable deposits and electronic currency (M1), plus time and savings deposits, foreign currency transferable deposits, certificates of deposit and securities repurchase agreements (M2), plus travellers checks, foreign currency time deposits, commercial paper and shares of mutual funds or market funds held by residents.

Human Capital

Gross enrolment ratio (proxy for human capital) is the ratio of total enrolment, regardless of age, to the population of the age group that officially corresponds to the level of education shown. Secondary education completes the provision of basic education that began at the primary level and aims at laying the foundations for lifelong learning and human development, by offering more subject- or skill-oriented instruction using more specialised teachers.

Macroeconomic Stability

Inflation (proxy for macroeconomic instability) is measured by the consumer price index, and it reflects the annual percentage change in the cost to the average consumer of acquiring a basket of goods and services that may be fixed or changed at specified intervals, for instance, yearly.

Estimation Methods

The first step for conducting any sort of econometric analysis is to perform a descriptive/summary statistical analysis of the data, the main emphasis of which is on data description rather than analysing and interpreting it. Hence, it is handy and practical in the sense that it covers description, summarisation and presentation of raw data. Next, the bivariate correlation analysis is attempted in order take a decision whether the linear relationship exists between the dependent and independent variables. In the present study, the pair-wise correlation is used to ascertain the degree of relationship between GDPG and individual independent variables. This analysis conveys whether the value of coefficient of correlation is close to ‘0’ or significantly different from ‘0’. When the value of coefficient is significantly different from ‘0’, we can say that it is significant, and vice versa.

After conducting the correlation analysis, the unit root test is performed to check the stationarity of variables. To check whether the panel data series contains a unit root or not, the Levin–Lin–Chu (LLC) unit root test is performed because it is suitable for panel data (Levin et al., 2002). In this context, the following hypotheses are tested:

H_o: Panels have unit root/panels are not stationary.

H_a: Panels have no unit root/panels are stationary.

After performing the LLC test, the panel data regression analysis is performed in order to assess the relationship between the dependent and explanatory variables. In order to arrive at the appropriate model for the study, four different panel models are performed, namely pooled ordinary least squares (OLS) regression, least squares dummy variable (LSDV) model, and fixed-effects and random-effects models. A simple linear regression technique that uses a panel data arrangement is known as pooled OLS regression. It is a straightforward method for estimating a panel data model and assumes that no differences exist among the data matrices of the cross-sectional dimension (Asteriou & Hall, 2007). The fixed-effects least-squares dummy variable (LSDV) model allows for the heterogeneity across cross-sections. It allows each subject to have its own intercept value; however, the slope coefficients of explanatory variables are assumed to be constant both across cross-sections and over time. The fixed-effects (within estimator) model involves within-group transformation of data, which is carried out to simplify the matters by avoiding the need to estimate many dummy variable parameters (Brooks, 2008). In this model, for each subject or entity, the values of the dependent and explanatory variables are expressed as deviations from their respective average or mean values (Gujarati & Porter, 2009). An alternative to fixed-effects model is random-effects model, also known as error components model. The random-effects model, like fixed-effects model, proposes different intercepts for each subject, which is constant over a period of time. However, the difference is that in random-effects model, for each individual unit, it is held that the intercepts arise from a common intercept term, β₁ (which is the same across individual units and over time) plus a random variable, ε_i, that differs across units but remains constant over time (Brooks, 2008).

In order to choose an appropriate model, the F-test has been applied to choose between pooled OLS and LSDV models (Gujarati & Porter, 2009). Next, Breusch–Pagan Lagrange Multiplier (LM) test (Breusch & Pagan, 1980) is used to decide whether random-effects model is the best or the pooled OLS regression test is the best. Lastly, the Hausman test is performed to choose between the fixed-effects and random-effects models (Hausman, 1978). On evaluating all these models, authors found that the fixed-effects model is appropriate for this study. The econometric model specification, showing the relationship between the dependent variables and explanatory variables under the fixed-effects method, is as follows:

{\ddot{GDPG}}_{i t} = β_{1} {\ddot{K}}_{i t} + β_{2} {\ddot{L}}_{i t} + β_{3} {\ddot{X}}_{i t} + β_{4} {\ddot{FDI}}_{i t} + β_{5} {\ddot{LL}}_{i t} + β_{6} {\ddot{HC}}_{i t} + β_{7} {\ddot{INF}}_{i t} + {\ddot{u}}_{i t}

where i denotes country, t denotes time and u_it is the error term. The double dots in the above models indicate the demeaned or mean corrected values of the variables.

In order to check whether our model fulfils the underlying assumptions of serial correlation, cross-sectional dependence, heteroscedasticity and multicollinearity, certain diagnostic tests have been performed before final conclusions are derived on the basis of fixed-effects model.

IV. Empirical Results and Discussion

In light of the estimation methods discussed in the previous section, the discussion on the results obtained are presented in this section.

Descriptive Statistics

The starting point in any type of analysis is to check the behaviour of data. Hence, the mean score of each dependent and independent variable together with their variability in terms of standard deviation are computed.

The summary statistics presented in Table 1 reveal that economic growth (GDPG) of BRICS on average is 4.22%, indicating healthy growth during the reference period. The minimum and maximum average economic growth is −14.53 and 14.23, respectively. The variation of economic growth in terms of dispersion is 4.657, indicating the sharp variability in the GDPG. The average FDI for the reference period is recorded at 1.86, and the variability in terms of σ is equal to 1.53. The mean for gross capital formation is 26.84 with a standard deviation of 9.36. For labour force, the mean value is 2.42 with a deviation of 2.9. Exports have a mean of 21.33 and variability equal to 9.32. The mean score of explanatory variable, liquid liabilities, is recorded at 61.23, and variability in terms of dispersion is 43.07. Human capital has a mean score of 87.29, with standard deviation at 36.75. The analysis of the explanatory variable, inflation, reveals that its mean score is recorded at 189.97, and variability in terms of dispersion is 663.72.

Table 1.

Descriptive Statistics.

Variables	N	Mean	Std. Deviation	Min.	Max.
1. Dependent GDPG	160	4.22	4.65	–14.53	14.23
2. Independent FDI	160	1.86	1.53	–0.21	6.18
Gross capital formation	160	26.84	9.36	14.83	47.81
Labour Exports Liquid liabilities Human capital Inflation	160 160 160 160 160	2.42 21.33 61.23 87.29 189.97	2.90 9.32 43.07 36.75 663.72	2.85 5.60 3.11 34.71 –1.40	7.85 62.32 218.00 231.87 4,273.38

Source: Authors’ calculation using STATA-14 Software.

Bivariate Correlation Analysis

To check whether a linear association exists between the dependent and independent variables, bivariate correlation analysis has been performed. The pair-wise correlation results between dependent and independent variables are exhibited in Table 2. The results reveal that there is positive correlation between FDI and GDPG, which is significant at the 1% level. Hence, it is concluded that there is a significant linear relationship between FDI and GDP growth. Gross capital formation and labour have a significant positive correlation with GDPG at 1% and 5%, respectively. The analysis also reveals that there is positive correlation between liquid liabilities and GDPG, which is significant at the 1% level. Moreover, human capital and inflation have a negative significant correlation with GDPG at the 1% level of significance.

Table 2.

Bivariate Correlation Results between Independent Variables and GDPG.

Independent Variables	r-Value	p-Value	Null Hypothesis
FDI	0.368	0.000	Rejected
Gross capital formation	0.540	0.000	Rejected
Labour	0.181	0.021	Rejected
Exports	–0.103	0.193	Accepted
Liquid liabilities	0.499	0.000	Rejected
Human capital	–0.371	0.000	Rejected
Inflation	–0.291	0.000	Rejected

Source: Authors’ calculation using STATA-14 Software.

Levin–Lin–Chu Unit Root Test

As already stated, it is a precondition to check the stationarity of variables before conducting any financial econometric analysis. If the mean and variance of a variable remain constant over a period of time, it is said to be stationary. In the present study, the Levin–Lin–Chu (LLC) test has been performed to check the presence of unit root for the variables under study.

Table 3 exhibits the results of the Levin–Lin–Chu test for the unit root; the results show that the variables, viz. GDPG, FDI, gross capital formation, liquid liabilities and inflation are stationary at level as the null hypothesis of non-stationarity is rejected in each case. However, in the case of other variables, that is, labour, exports and human capital, the null hypothesis is not rejected at level but at the first difference, so these are stationary at first difference. After checking the stationary nature of data series, panel data analysis is attempted next, the results of which are discussed in the following section.

Table 3.

Levin–Lin–Chu Unit Root Test.

Variable	t-Statistics	p-Value	Stationary At
GDPG	–3.94	0.000	Level
FDI	–1.66	0.048	Level
Gross capital formation	–1.43	0.075	Level
Labour Exports Liquid liabilities Human capital Inflation	–1.54 –4.50 –6.16 –2.42 –2.19	0.026 0.000 0.000 0.007 0.014	1st Difference 1st Difference Level 1st Difference Level

Source: Authors’ calculation using STATA-14 Software.

Panel Data Analysis

As stated in the earlier section, the four panel data models, namely pooled OLS regression, least squares dummy variable model, fixed-effects model and random-effects model have been analysed, but only the fixed-effects model has been found appropriate for the present research study.

F-test/Wald Test and LM Test

To choose between OLS regression and fixed-effects LSDV model, F-test/Wald test is conducted, whereas Breusch–Pagan LM test (Breusch & Pagan, 1980) is conducted to select between OLS regression and random-effects model. In this regard, if the null hypothesis is not rejected in either case, pooled OLS regression is the appropriate model. The null hypothesis is as under:

H₀: No difference exists in the cross-sectional units, or

H₀: µ₁ = µ_{n − 1} = 0 (all dummy parameters, except for one for dropped are all 0)–F-test

H₀: σ²_µ = 0 (individual (or time) specific variance components are 0—LM-test.

The results of LM test presented in Table 4 show that the p-value > 0.01, indicating that there is no significant random effect. The results also reveal that both F-test and Wald test are significant at the 1% level, which results in rejection of null hypothesis (i.e., no difference exists in the cross-sectional units). Therefore, there is a significant fixed effect and the data are not poolable, indicating that pooled OLS model is not suitable when compared to the fixed-effects model. To ensure the appropriateness of the model, we conducted Hausman test, the results of which are presented next.

Table 4.

F-test/Wald-test and LM Test.

Model	Test	Value	Decision
Fixed	F-test p-Value	7.88 0.000	H is Rejected
Random	Wald-test p-Value	2.96 0.006	H is Rejected
Random	LM test p-Value	0.00 1.000	H is Accepted

Source: Authors’ calculation using STATA-14 Software.

Hausman test

Hausman test is used to decide between the two models, that is, the fixed-effects model and random-effects model (Hausman, 1978). The null hypothesis is that there is no substantial difference between the fixed-effects and random-effects estimator. In case the null hypothesis is rejected, it is concluded that the random-effects model is not appropriate as the random effects may be correlated with the explanatory variables.

The results of the Hausman specification test presented in Table 5 depict that the χ² value is significant at the 1% level of significance; thus, the null hypothesis is rejected (p-value < 0.01). Since null hypothesis of no correlation is rejected, one can conclude that individual effects (µ_i) are significantly correlated with other regressors, so the fixed-effects model is suitable. However, before drawing any conclusion about the growth effect of FDI in the BRICS nations the basis of fixed-effects model, it is important to assure that the identified model fulfils the underlying assumptions of linear regression model, so the following section presents the results of requisite diagnostic tests.

Table 5.

Hausman Test.

	Coefficient
χ² test value	35.9
p-Value	0.000

Source: Authors’ calculation using STATA-14 Software.

Diagnostic Tests

Fixed-effects model has been found to be appropriate for the present research study. However, to ensure that this model is valid and reliable, it has to satisfy the following underlying assumptions:

Serial Correlation: An assumption of classical linear regression model (CLRM) states that the covariance between the different disturbances, over time or cross-sectionally, is equal to 0. In case the disturbances are found to be correlated with one another, they are said to be ‘autocorrelated’ or ‘serially correlated’, violating a key assumption of CLRM. In this context, the Wooldridge test for autocorrelation in panel data is used to test the serial correlation, and the following null hypothesis is tested:

H_o: No serial correlation exists.

The results of the test presented in the Table 6 reveal that the p-value < 0.01, so we reject the null hypothesis. Hence, it is deduced that the panel data suffers from the issue of serial correlation.

Table 6.

Wooldridge Test for Autocorrelation.

H₀: No First-order Autocorrelation
F (1, 4) Prob. > F	22.6 0.008

Source: Authors’ calculation using STATA-14 Software.

Cross-sectional Dependence: An issue inherent in almost all the panel data studies is the possibility of individual units, that is, the cross-sections, being interdependent. Econometric literature focused on the panel-data models substantiates that the cross-sectional dependence may be due to some common factors or unobservable components. In this regard, the Pesaran’s CD Test is used to check cross-sectional dependence of residuals, and the following null hypothesis is tested:

H_o: Residuals are cross-sectionally uncorrelated.

The results of the test presented in Table 7 reveal that p-value < 0.01, so null hypothesis is rejected. Hence, it is concluded that the model suffers from the problem of cross-sectional dependence.

Table 7.

Test for Cross-sectional Dependence.

H₀: Residuals Are Cross Sectionally Uncorrelated
Pesaran’s Test Statistics p-Value Average absolute value of the off-diagonal elements	2.87 0.004 0.210

Source: Authors’ calculation using STATA-14 Software.

Heteroscedasticity: In the context of panel data, the deviation from the assumption of homoscedasticity (constant variance) is likely due to the error variances that are particular to the cross-sectional unit. It may happen that the error process is homoscedastic within the units, but its variance differs across the cross-sectional units: a condition known as group-wise heteroscedasticity. When we conduct regression in the presence of heteroscedasticity, we have confidence intervals, which are unusually large due to which the t- and F-tests may give inaccurate results. In the present study, modified Wald statistic is applied to test group-wise heteroscedasticity. The null hypothesis specifies that σ²_i = σ² for i = 1, ... , N_g, where N_g denotes the number of cross-sectional units.

The results presented in Table 8 reveal that p-value < 0.01, so null hypothesis is rejected. Therefore, it is deduced that the model suffers from heteroscedasticity.

Table 8.

Modified Wald Test for Group-wise Heteroscedasticity.

H₀: σ(i)² = σ² for all i
χ² (5) Prob. > χ²	98.74 0.000

Source: Authors’ calculation using STATA-14 Software.

Multicollinearity: ‘No multicollinearity’ is an important assumption to be proved for all regression models because if this problem exists, regression analysis will give absurd results, and standard errors for coefficients will be highly inflated. Therefore, it is important to test the presence of multicollinearity. The multicollinearity occurs when variance inflation factor (VIF) value is greater than 10 (Gujarati & Sangeetha, 2007).

The results of collinearity test presented in Table 9 clearly reveal that the VIF value is less than 10, indicating that the model does not suffer from the problem of multicollinearity.

Table 9.

Test of Collinearity.

Variables	VIF	1/VIF
FDI	1.59	0.630
Gross capital formation	3.08	0.324
Labour Exports Liquid liabilities	1.65 1.01 2.94	0.607 0.986 0.339
Human capital	1.55	0.643
Inflation Mean VIF	1.57 1.91	0.638

Source: Authors’ calculation using STATA-14 Software.

Feasible Generalised Least Squares Model

The results of the Hausman test revealed that fixed-effects panel estimates are preferred over the random-effects estimates; however, the diagnostic test has identified that the model suffers from the problem of autocorrelation, cross-sectional dependence and heteroscedasticity. Hence, the fixed-effects panel estimation may yield misleading and absurd results. In this situation, feasible generalised least squares (FGLS) model is normally recommended (Reed & Ye, 2011). The FGLS produces autocorrelation, cross-sectional dependence and heteroscedasticity consistent estimation, provided T (time period) is greater than N (cross sections) (Parks, 1967). Obviously, the FGLS model is suitable for present data set. Hence, the FGLS model is employed in the present study to understand the relationship between the dependent and explanatory variables. The results of the FGLS regression model presented in Table 10 reveal that the FGLS model fits the data well at the 0.01 significance level, as the χ² value of the model is 122.65, which is significant at 1%.

Table 10.

Cross-sectional Time-series FGLS Regression.

Dependent Variable: GDPG	Coefficients	St. Err.	z	P>\|z\|	[95% Conf.	Interval]
FDI	0.568	0.226	2.51	0.012	0.124	1.012
Gross capital formation	0.203	0.051	3.91	0.000	0.101	0.304
Labour	–2.81	1.23	–0.23	0.819	–2.68	2.12
Exports	–0.074	0.029	–2.49	0.013	–0.133	–0.015
Liquid liabilities	0.004	0.011	0.45	0.653	–0.016	0.026
Human capital	–0.021	0.009	–2.29	0.022	–0.039	–0.003
Inflation	–0.001	0.0005	–2.10	0.036	–0.002	–0.00007
Constant	1.151	1.629	0.71	0.480	–2.042	4.345
Number of obs	160.000		χ ²		122.65
Prob > χ²	0.000		Akaike crit. (AIC)		870.31

Source: Authors’ calculation using STATA-14 Software.

The results reveal that for 1 percentage point increase in FDI, the GDPG is expected to increase by 0.568 percentage points, holding all other variables constant (p < 0.05), indicating that FDI inflows exert a significant positive impact on economic growth of BRICS economies at the 5% significance level. Our results lend support to the studies, claiming FDI to be a driver of growth (de Mello, 1999; Hansen & Rand, 2006; Yao, 2006; Tiwari & Mutascu, 2011). FDI is believed to influence growth through spillovers in the form of technology and knowledge. Gross capital formation (i.e., domestic investment) also exerts a positive impact (0.203) on growth at the 1% level of significance, implying that both FDI and domestic investment contribute positively to the growth of BRICS economies. It is found that human capital exerts a negative impact on growth at the 5% significance level, which seems contrary to theoretical expectations; however, the reason of this could be the heterogeneity of countries in the panel (Pelinescu, 2015). Inflation used as proxy for economic instability exerts a negative effect on growth at the 5% significance level, supporting the results of previous studies that found the negative effect of inflation on growth (Abdelmalki et al., 2012; Bleaney, 1996). It is found in our study that exports have a negative impact on economic growth of BRICS at the 5% level of significance. Our results are in favour of FDI-led growth as opposed to export-led growth hypothesis (a long debated issue in growth literature).

V. Conclusion

Due to the immense surge in FDI flows over the past three decades, the impact of FDI on growth is extensively studied by researchers from time to time in order to guide the policymakers in decisions regarding stimulation or restriction of FDI. However, mixed evidence is found in the growth literature in this regard. The present study is, therefore, aimed at assessing the growth effect of FDI and select macroeconomic variables in BRICS economies. The BRICS economies is selected meticulously because of their growth trajectory and dominance in the global investment landscape, having attracted huge inward FDI over the past decades. The study employed the FGLS method over the period from 1987 to 2018. The results reveal a significant positive impact of both FDI and domestic investment on economic growth in host countries. The BRICS countries should focus on facilitating both domestic investment and FDI to reap the growth benefits, they should make efforts to attract FDI flows by offering incentives to foreign investors and maintain an enabling environment for both domestic and foreign investors. Moreover, it is suggested that the policymakers should make efforts to maintain economic stability by containing inflation in order to achieve sustainable economic growth.

Footnotes

Declaration of Conflicting Interests

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

ORCID iDs

Bashir Ahmad Joo

Sana Shawl

References

Abbas

, & Mujahid-Mukhtar

(2001). The role of human capital in economic growth: A comparative study of Pakistan and India. The Pakistan Development Review , 39(4), 451–473.

Abdelmalki

, Gbakou

, Jallab

M. S.

, & Sandretto

(2012). Does macroeconomic stability condition the impact of FDI on economic growth and emergence? An empirical study of countries in Africa, Latin America, the Caribbean and Asia. Developing Worlds , 2(158), 101–114.

Akpan

E. S.

, & Eweke

G. O.

(2017). Foreign direct investment and industrial sector performance: Assessing the long-run implication on economic growth in Nigeria. Journal of Mathematical Finance , 7, 391–411. https://doi.org/10.4236/jmf.2017.72021

Alfaro

, & Charlton

(2007). Growth and the quality of foreign direct investment: Is all FDI equal? (CEP Discussion Paper, Issue 830). London School of Economics and Political Science. https://doi.org/10.2139/ssrn.981163

Alfaro

, & Chauvin

(2020). Foreign direct investment, finance, and economic development. In Spatareanu

(Ed.), Encyclopedia of international economics and global trade (Vol. 1, pp. 231–258). World Scientific Publishing Co. Pte. Ltd.

Alfaro

, Chanda

, Kalemli-Ozcan

, & Sayek

(2004). FDI and economic growth: The role of local financial markets. Journal of International Economics , 64, 89–112. https://doi.org/10.1016/S0022-1996(03)00081-3

Alfaro

, Kalemli-Ozcan

, & Sayek

(2009). FDI, productivity and financial development. The World Economy , 32(1), 111–135. https://doi.org/10.1111/j.1467-9701.2008.01159.x

Alguacil

, Cuadros

, & Orts

(2011). Inward FDI and growth: The role of macroeconomic and institutional environment. Journal of Policy Modeling , 33, 481–496. https://doi.org/10.1016/j.jpolmod.2010.12.004

Asteriou

, & Hall

S. G.

(2007). Applied econometrics (2nd ed.). Palgrave Macmillan.

10.

Barro

R. J.

, & Sala-i-Martin

(1995). Economic growth . McGraw-Hill.

11.

Beck

, Levine

, & Loayza

(2000). Finance and the sources of growth. Journal of Financial Economics , 58, 261–300.

12.

Bengoa

, & Sanchez-Robles

(2003). Foreign direct investment, economic freedom and growth: New evidence from Latin America. European Journal of Political Economy , 19(3), 529–545. https://doi.org/10.1016/S0176-2680(03)00011-9

13.

Bleaney

M. F.

(1996). Macroeconomic stability, investment and growth in developing countries. Journal of Development Economics , 48, 461–477.

14.

Borensztein

, Gregorio

J. De

, & Lee

(1998). How does foreign direct investment affect economic growth? Journal of International Economics , 45, 115–135.

15.

Bouchoucha

, & Ali

(2019). The impact of FDI on economic growth in Tunisia: An estimate by the ARDL approach. MPRA , 1, 1–23.

16.

Breusch

T. S.

, & Pagan

A. R.

(1980). Lagrange multiplier test and to model applications specification in econometrics. The Review of Economic Studies , 47(1), 239–253.

17.

Brooks

(2008). Introductory econometrics for finance (2nd ed.). Cambridge University Press.

18.

Cao

, & Jariyapan

(2012). Foreign direct investment, human capital and economic growth of people’s republic of China using panel data approach. Journal of Economics , 16(1), 30–42.

19.

Carkovic

M. V.

, & Levine

(2002). Does foreign direct investment accelerate economic growth? (Working Paper). Department of Finance, University of Minnesota.

20.

de Mello

L. R.

(1999). Foreign direct investment-led growth: Evidence from time series and panel data. Oxford Economic Papers , 51, 133–151.

21.

Dritsaki

, Dritsaki

, & Adamopoulos

(2004). A causal relationship between trade, foreign direct investment and economic growth for Greece. American Journal of Applied Sciences , 1(3), 230–235.

22.

Evans

A. D.

, Green

C. J.

, & Murinde

(2002). Human capital and financial development in economic growth: New evidence using the translog production function. International Journal of Finance and Economics , 7, 123–140.

23.

Feridun

, & Sissoko

(2011). Impact of FDI on economic development: A causality analysis for Singapore; 1976–2002. International Journal of Economic Sciences and Applied Research , 4(1), 7–17.

24.

Grossman

G. M.

, & Helpman

(1991). Innovation and growth in the global economy . MIT Press.

25.

Gujarati

D. N.

, & Sangeetha

(2007). Basic econometrics (4th ed.). McGraw-Hill. https://doi.org/10.1126/science.1186874

26.

Gujarati

D. N.

, & Porter

D. C.

(2009). Basic econometrics (5th ed.). McGraw Hill Inc.

27.

Hansen

, & Rand

(2006). On the causal links between FDI and growth in developing countries. The World Economy , 29(1), 21–41. https://doi.org/10.1111/j.1467-9701.2006.00756.x

28.

Hausman

J. A.

(1978). Specification tests in econometrics. Econometrica , 46(6), 1251–1271.

29.

Hermes

, & Lensink

(2003). Foreign direct investment, financial development and economic growth. The Journal of Development Studies , 38, 1–42.

30.

Herzer

(2012). How does FDI really affect developing countries’ growth? Review of International Economics , 20(2), 396–414.

31.

Hill

C. W. L.

(2011). International business: Competing in the global marketplace (8th ed.). McGraw-Hill.

32.

Jallab

M. S.

, Benoît

, Gbakou

, & Sandretto

(2008). Foreign direct investment, Macroeconomic instability and economic growth in MENA countries (Working Paper No. 0817). Groupe d'Analyse et de Théorie Economique Lyon St-Étienne, Université de Lyon.

33.

Kelly

(2016). Does financial sector development enhance the relationship between FDI and economic growth? A comparative study of East African countries. Journal of Reviews on Global Economics , 5, 145–153.

34.

Kholdy

, & Sohrabian

(2005). Financial markets, FDI, and economic growth: Granger causality tests in panel data model . EFA 2005 Moscow Meetings. https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.676085.

35.

King

R. G.

, & Levine

(1993a). Finance, entrepreneurship, and growth theory and evidence. Journal of Monetary Economics , 32, 513–542.

36.

King

R. G.

, & Levine

(1993b). Finance and growth: Schumpeter might be right. The Quarterly Journal of Economics , 108(3), 717–737.

37.

Ledyaeva

, & Linden

(2006). Foreign direct investment and economic growth: Empirical evidence from Russian regions (BOFIT Discussion Papers No. 17). https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.1002361

38.

Levin

, Lin

, & Chu

C. J.

(2002). Unit root tests in panel data: Asymptotic and finite-sample properties. Journal of Econometrics , 108, 1–24.

39.

Lian

, & Ma

(2013). FDI and economic growth in western region of China and dynamic mechanism: Based on time-series data from 1986 to 2010. International Business Research , 6(4), 180–186. https://doi.org/10.5539/ibr.v6n4p180

40.

Ljungwall

, & Li

(2007, August). Financial sector development, FDI and economic growth in China (Working Paper No. E2007005). China Center for Economic Research.

41.

Mankiw

N. G.

(2004). Principles of economics . Thomson, South-Western.

42.

Mehic

, Silajdzic

, & Babic-Hodovic

(2013). The impact of FDI on economic growth: Some evidence from Southeast Europe. Emerging Markets Finance and Trade , 49(1), 5–20. https://doi.org/10.2753/REE1540-496X4901S101

43.

Mencinger

(2003). Does foreign direct investment always enhance economic growth? KYKLOS , 56(4), 491–508.

44.

Ögütçü

(2002). Foreign direct investment and regional development: Sharing experiences from Brazil, China, Russia and Turkey (OECD Paper, Paris). https://www.oecd.org/cfe/regionaldevelopment/2489844.pdf

45.

Parks

R. W.

(1967). Efficient estimation of a system of regression equations when disturbances are both serially and contemporaneously correlated. Journal of the American Statistical Association , 62, 500–509.

46.

Pelinescu

(2015). The impact of human capital on economic growth. Procedia Economics and Finance , 22, 184–190. https://doi.org/10.1016/s2212-5671(15)00258-0

47.

Prüfer

B. P.

, & Tondl

(2008). The FDI-growth nexus in Latin America: The role of source countries and local conditions (CentER Discussion Paper Series No. 2008-61, pp. 1–34). https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.1154914

48.

Reed

W. R.

, & Ye

(2011). Which panel data estimator should I use? Applied Economics , 43(8), 985–1000. https://doi.org/10.1080/00036840802600087

49.

Szkorupová

(2014). A causal relationship between foreign direct investment, economic growth and export for Slovakia. Procedia Economics and Finance , 15, 123–128. https://doi.org/10.1016/S2212-5671(14)00458-4

50.

Tiwari

A. K.

, & Mutascu

(2011). Economic growth and FDI in Asia: A panel-data approach. Economic Analysis and Policy , 41(2), 173–187. https://doi.org/10.1016/S0313-5926(11)50018-9

51.

Yao

(2006). On economic growth, FDI and exports in China. Applied Economics , 38, 339–351. https://doi.org/10.1080/00036840500368730