Does FDI Impact the Economic Growth of BRICS Economies? Evidence from Bayesian VAR

Abstract

This paper examines the dynamic relationship between foreign direct investment (FDI), economic growth, and trade openness in BRICS countries. Our research aims to address a significant gap in the literature by focusing on this crucial group of emerging nations, given their substantial contribution to the global economy. Annual data for these economies from 1991 to 2020 were collected from various secondary sources. This study employed the Bayesian VAR framework to investigate the panel data. The Pedroni residual cointegration test was used to check the existence of a long-run relationship between FDI and economic growth. The results provided evidence that foreign direct investment (FDI) does exhibit a substantial correlation with economic growth in the short run. However, no long-run relationship was found in the case of BRICS economies. This research contributes to methodological innovation by introducing the Bayesian VAR framework, offering a deeper understanding of the dynamic interactions among these key variables. The incorporation of this framework yields estimates that are both stable and reliable, which is certainly a novelty of this paper. The findings of this study have implications suggesting that policymakers from these emerging economies should establish mechanisms that will monitor the short-term impacts of FDI and adjust policies accordingly to maximize economic gains. The government should tailor policies to the specific circumstances of each country for sustainable economic development.

1. Introduction

Foreign direct investment (FDI) and economic growth have increased substantially in recent decades, especially in emerging nations (Khan and Mehboob 2014; Irshad and Qayed 2022). Developing nations are competing for foreign direct investment (FDI) in an era of rising global inequality (Banday et al. 2021). This is because foreign direct investment (FDI) is perceived as a significant driver of economic growth since it improves technology, brings employment opportunities, expands trade, boosts commerce, and facilitates global market integration (Atique et al. 2004; Tiwari and Mutascu 2011). When a person or corporation from one nation invests money in the economy of another country, they are undertaking a foreign direct investment, or FDI (Hansen and Rand 2006; Herzer 2012). This involves the infusion of foreign-based funds into a commercial venture that operates in a different country of origin from the investor (Johnson 2006). The ‘source’ country is the nation from which the funds are invested, while the ‘host’ or ‘receiver’ country is the nation that ultimately receives the investment inflows. Foreign direct investment (FDI) is defined by the World Bank as investments made outside of an investor’s home economy with the intention of acquiring a long-term stake in or effective control of a business enterprise. FDI has proven to be the most preferred choice over the other modes of investment due to its stable nature (Nistor 2014). The governments of developing countries have now realized the significance of increased FDI inflows. They try to attract foreign investments through numerous government policies and the adoption of bilateral investment treaties (Herzer 2012). Host nations open up their economies and provide various incentives to attract more and more foreign direct investment into their countries (Hansen and Rand 2006; Bhandari et al. 2007). This is mainly carried out to reap the benefits of incoming investment by the recipient country. Therefore, it can be argued that FDI serves as one of the essential components for the process of economic integration as it strengthens ties and the long-term advantages of cooperation between multiple countries (Mehic et al. 2013). It is regarded as the main source of economic sustainability for developing nations (Omri 2014).

Incoming foreign direct investment (FDI) is seen as a vital source of capital by developing nations since it boosts a wide range of macroeconomic parameters, including domestic investment, exports, and the incorporation of new technology into the private sector, which ultimately contributes to economic expansion and increased productivity (Akinmulegun 2012; Mahmoodi and Mahmoodi 2016). The investment inflows generate employment opportunities and stimulate the economy by either the opening up of a local manufacturing unit or the acquiring of an existing one. The extent of these benefits, however, is still being determined since they are greatly dependent on the quality of the technology supplied by foreign companies and the adaptability of the recipient nation (Rakhmatullayeva et al. 2020).

The present research study aims to understand the inter-relationship between FDI inflows, economic growth, and trade openness in the BRICS economies, as these nations have the potential to overtake established economies largely owing to their size, high population, and geographic regions (O’Neill 2001). The emerging economies came into existence in the year 2001, when American economist Jim O’Neill created the acronym BRIC, formed by four countries: Brazil, Russia, India, and China. Later, South Africa was added to the BRIC group. The official process for South Africa’s inclusion into the organization started in August 2010. On 24 December 2010, South Africa formally joined the BRIC nations after receiving a formal invitation from the group. To reflect South Africa’s inclusion in the expanded group, the name of the organization was changed to BRICS—with an “S” at the end. At the 15th BRICS summit held in Johannesburg, South Africa, from 23–24 August (2023), the member nations decided to expand their economic collaboration by admitting six new members, namely Argentina, Egypt, Ethiopia, Iran, Saudi Arabia, and the UAE. The new membership will be effective from 1 January 2024. With this expansion, the BRICS economies will constitute nearly 30% of the world’s GDP and a 46% share of the global population. Their combined share in per-day oil production will increase from 20.4% to 43.1%. By the end of year 2022, the original BRICS members accounted for 18.28% of the global export of goods and services. With the inclusion of the six new economies, the BRICS alliance will contribute a 20.58% share of global exports. Among the 11 members, China is the leading contributor with 12% of global exports, followed by India at 2.5%. Clearly, the BRICS countries have taken on a significant role as the producers of products and services in the global economy for a variety of reasons.

According to the latest World Economic Outlook (2023) report released by the International Monetary Fund (IMF), China and India will dominate global GDP growth in the next five years. Based on the projections made by the WEO report in

Table 1. World economy growth contribution in 2023–2028.

Country	Share in World Economy Growth (2023–2028)
China	22.6
India	12.9
USA	11.3
Indonesia	3.6
Germany	2.1
Turkey	2.1
Japan	1.8
Brazil	1.7
Egypt	1.7
Russia	1.6
Bangladesh	1.6
Vietnam	1.6
UK	1.5
France	1.5

Source: International Monetary Fund (IMF) WEO report (April 2023).

below, China will account for 22.6% of global GDP growth, India for 12.9%, and the United States for 11.3%. It is worth noting that China’s contribution to global growth is still likely to be enormous compared to that of the United States and India. Of all countries, the BRICS (Brazil, Russia, China, India, and South Africa) nations together are expected to contribute almost 40% of the world’s growth through the year 2028. The share of global GDP contributed by BRICS countries has increased from 18% in 2010 to 26% in 2021. Interestingly, the population growths of BRICS countries have slowed relative to the global average, while their share of global GDP has been growing, resulting in BRICS countries’ per capita GDP growing faster than the global per capita GDP.

Further, Figure 1 depicts the share of global GDP using purchase-power parity from 1992 to 2028. The data show a continuous decline in the share of G7 nations, consisting of Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States. The BRICS nations have shown a consistent rise in their share of the global GDP since 1992 and have outpaced the G7 nations since 2020.

The future predictions for 2023–2028 will see an increase in BRICS nations’ GDP growth share while seeing a declining trend for the G7 nations. This is why investors from around the globe are currently drawn to the BRICS nations. These nations are characterized by massive populations, promising consumer markets, rapid economic expansion, vast land masses, etc., which make them the investors’ first choice of destination. By the middle of this century, the combined economies of the BRICS countries are projected to have grown to be the biggest in the world. Together, the BRICS nations have shown continuous growth and formed a strong economic power (Nistor 2015).

Additionally, the trends in FDI flows in BRICS nations show that these countries have remained a major attraction for investors in recent times.

Table 2. FDI Inflows in BRICS countries (2022).

Sno.	Country	FDI Inflows in USD Millions
1.	Brazil	91,502
2.	China	180,167
3.	India	49,916
4.	Russia	−43,133
5.	South Africa	8803

Source: World Bank Database—www.worldbank.org (accessed on 24 September 2023).

below presents the total flow of foreign direct investment in BRICS countries, namely Brazil, Russia, India, China, and South Africa. The latest data for these nations were procured for 2022 from the World Bank Database.

China has received the highest investment among all the five BRICS countries, with a total inflow of funds amounting to USD 180,167 million. China is followed by Brazil and India with FDI inflows equal to USD 91,502 and 49,916 million, respectively. The investment received by South Africa is considerably low as compared to the other BRICS nations, amounting to only USD 8803 million. Russia remains an exception with negative incoming inflows in 2022 due to the prolonged Ukraine-Russia war, which caused the FDI flows in this region to follow a downward trajectory. Subsequently, Figure 2 below shows the cumulative FDI inflows in BRICS countries from 1991 to 2020. Over the past thirty years, China has received more FDI than the other four nations combined, while South Africa has attracted the least.

Thus, studying the nexus between FDI, economic growth, and trade openness in BRICS nations is critically important. These nations collectively represent a significant portion of the global economy, and understanding the dynamics of FDI, economic growth, and trade openness within them will provide valuable insights into broader economic trends.

Further, this study is organized into Section 2, which provides a concise review of the existing literature, and Section 3, which describes the data and methods used in the study. The outcomes of the study are addressed in Section 4. Finally, the findings are summarized in Section 5.

2. Literature Review

Investment of any kind, including FDI, is believed to boost the host country’s economic growth (Samad 2009; Pegkas 2015). Foreign direct investment (FDI) is widely regarded as the most viable option for boosting economies in emerging markets because it provides access to a pool of trained managers, boosts the competitiveness of local firms, and broadens the scope of their exports (Borensztein et al. 1998; Zhang 2001; Strat et al. 2015; Iqbal et al. 2018). The effect of FDI on economic development from 1970–2001 is examined by Atique et al. (2004). The findings traced that the impact of FDI was higher on the growth rate under the export-promotion trade regime in contrast to the import-substitution trade regime. Likewise, Makki and Somwaru (2004) analyzed the impact of foreign direct investment (FDI) on economic growth in 66 developing nations. The results indicated a positive correlation between FDI and GDP growth rates. Similar results were reported by Lee (2007) for 11 OECD countries for the period 1971–1999. The study finds a significant and robust long-term relationship between inbound foreign direct investment and the growth productivity of the host nations. Several other studies have also examined the long-term relationship between FDI and economic growth (De Mello 1999; Hansen and Rand 2006; Narayanamoorthy et al. 2008; Tiwari and Mutascu 2011; Saidi et al. 2023). Naveed and Shabbir (2006) studied the impact of foreign direct investment (FDI) and trade openness on gross domestic product (GDP). This study found that trade openness positively affects GDP, while FDI negatively affects GDP. The study by Alam et al. (2022) pointed out that FDI can impact economic growth, affecting both demand and supply in the case of India.

Another major study conducted by Mehic et al. (2013), using the Prais–Winsten regression technique for seven southeast European countries for the period 1998–2009, revealed that FDI is a crucial factor that enhances economic growth. In contrast, the impact of domestic capital formation on economic growth is relatively weaker. Other factors that were also helpful in determining economic growth included trade openness and macroeconomic stability, denoted by the inflation rate. Likewise, Khan and Mehboob (2014) evaluated the impact of FDI on gross domestic product (GDP) using balanced panel data from 1992–2010 in 59 countries. The study emphasized the importance of liberal policies for the free flow of capital across nations. Using a growth model structural theory, Omri (2014) tried to evaluate the effect of both foreign and domestic investment on economic growth in the MENA countries. An indirect relationship was observed between the countries’ financial development and economic growth via foreign investment. The exchange rate and degree of openness also contributed significantly to the development of MENA countries. Sunde (2017) recognized the correlation between FDI, exports, and economic growth in the South African economy and concluded that a strong macroeconomic environment will help to attract more investment from other countries. Other variables, like trade openness and foreign exchange, also favorably impacted economic growth (Fadhil and Almsafir 2015; Alam et al. 2022). Similarly, Yimer (2023) examined the growth effects of FDI on Africa from 1990 to 2016 using the error correction model, keeping in mind the heterogeneity of each country. The study revealed that the impact of FDI on growth differs for each country. Shell and Zheng (2015) used the globalization and institutional quality indexes to study their impact on international capital inflows from 1970–2010. The results showed that both variables have a significant positive relationship with capital inflows. The research conducted by Kim et al. (2018) pointed out the role of regional vs. global savings in financing domestic investment, leading to increased international capital flows from 1980 to 2014 in 141 nations worldwide. Likewise, Athari and Adaoglu (2019) explored the various determinants of capital inflows for a sample of 135 countries across the globe at different stages of economic development, emphasizing the role of institutional quality, accountability, and corporate ethics. In another study, Athari et al. (2020) investigated the role of a country’s competitiveness on the cross-border flow of capital by using the global competitiveness index at different stages of economic development from 2006–2014 for 133 countries. The findings revealed a strong positive relationship between competitiveness and capital flows, although, the degree of its impact varies across different development stages and country risk levels.

The study by Asafo-Agyei and Kodongo (2022) highlighted the impact of FDI on the economic growth of sub-Saharan African countries. The research suggested that for foreign direct investment (FDI) to contribute to economic growth significantly, countries must possess a minimum capacity to absorb the growth-enhancing benefits of FDI. Further, the study by Nupehewa et al. (2022) also investigated the causal relationship between foreign direct investment (FDI) and the economic growth of seven regions, covering 117 countries, from 2010 to 2020 using the Granger causality approach and panel VAR/block exogeneity test. The empirical findings revealed a bidirectional causality between FDI and economic growth globally and in the Asian region. In contrast, the causality is unidirectional in the American region. Similar results were reported by Kumari et al. (2023) for time series data from 1985–2018 in the case of India. The study revealed a bidirectional causality from FDI to economic growth, whereas there is no bidirectional causality between trade openness and economic growth. For countries in southeast Asia (SEA) and south Asia (SA), the study by Rao et al. (2023) investigated the interrelationship among foreign aid, foreign direct investment (FDI), and economic growth from 1980–2016. The results indicated a negative association between foreign aid and both FDI and economic growth. Conversely, FDI was found to have a positive influence on economic growth.

On the other hand, some research studies show a negative impact of FDI on economic growth, while others find no meaningful association between the two variables. While looking at developing economies in the 1970s and 1980s, Saltz (1992) discovered that foreign direct investment (FDI) has a negative impact on GDP growth. Similarly, incoming investment from overseas has a detrimental impact on the economy, according to a study by Mencinger (2003) that investigated the eight EU member states during 1994–2001. Herzer (2012) contributed to the existing literature by examining the impact of foreign direct investment on the growth of 44 developing economies with the help of heterogeneous panel co-integration techniques. The findings showed a negative relationship between the two variables, largely owing to country-specific differences in the growth effects of foreign direct investment. Using GMM and OLS methodology, Anyanwu and Yameogo (2015) examined the correlation between FDI and economic growth in west African countries from 1970 to 2010. They discovered a negative correlation between the two variables. The results were consistent with the previous studies by Mencinger (2003) and Herzer (2012). In recent times, the study by Alvarado et al. (2017) for Latin American countries concluded that incoming investment from overseas is not an appropriate vehicle of growth and development except in the high-income group nations. Additionally, the results of the study by Saidi et al. (2023) suggested that the quality of governance in developing countries does not significantly affect foreign direct investment (FDI) and economic growth.

Furthermore, using data from 72 countries between 1960 and 1995, Carkovic and Levine (2005) showed a weak association between FDI and GDP growth. Likewise, Herzer and Klasen (2008) for 28 developing countries from 1970–2003; Nath (2009) for 13 transition economies of central and eastern Europe and the Baltic region during the period 1991–2005; Louzi and Abadi (2011) for Jordan for the period 1997–2006; Tekin (2012) for the case of the least-developed countries during the period 1970–2009; Belloumi (2014) for the case of Tunisia during the period 1970–2008; and more recently, Mahembe and Odhiambo (2016) for the case of low-income SADC countries during the period 1980–2012 reported a weak or no significant impact of FDI on economic growth.

The findings of the study by Choi and Baek (2017) implicated that the gradual increase in foreign direct investments (FDI) had little impact on productivity over the decades. The economic growth increased over the years, but not as much as the increase in incoming FDI. Similarly, Bermejo Carbonell and Werner (2018) for India and Golitsis et al. (2018) for Albania highlighted that only a small role is played by inbound investment from the MNCs of advanced nations in the growth of the economy, despite the vast literature suggesting otherwise. It is important to note that the positive impact of FDI is not a fact; it may vary widely depending on factors like the political and macroeconomic stability of the host country, as well as its institutional capability, infrastructure, and educational system (Rakhmatullayeva et al. 2020).

Evidently, a considerable amount of the literature has empirically examined the association between foreign direct investment (FDI), economic growth, and trade openness individually for advanced nations. However, only a few researchers have studied the causality between these variables for the BRICS economies, given their increasing role in the world’s economic development in the coming years. The paper analyzes explicitly how the relationship between FDI, economic growth, and trade openness has evolved in the backdrop of changes in global economic conditions, particularly regarding the 1997 Asian financial crisis, the 2008 US sub-prime crisis, and the era of quantitative easing by the Federal Reserve Bank of USA. This is certainly a contribution to the existing literature in terms of a larger data set.

Keeping this in mind, the current research study investigates the interrelationship among the given three variables, namely FDI, economic growth, and trade openness, using the Bayesian VAR framework and the Pedroni residual cointegration test.

Therefore, based on an extensive review of the literature, the present study formulated the following hypotheses:

H1. 

There is no co-movement between FDI, economic growth, and trade openness in BRICS nations.

H2. 

There is no long-term dynamic relationship between FDI, economic growth, and trade openness in BRICS nations.

3. Data and Methodology

The current study aims to analyze the dynamic relationship between FDI, economic growth, and trade openness in BRICS economies, namely Brazil, Russia, China, India, and South Africa. For this purpose, annual data were used from 1990–1991 to 2020–2021, i.e., three decades from various secondary sources such as World Bank indicators, OECD statistics, and Penn Table (Version 10.0). Panel data estimation was employed in the study to capture the dynamic behavior of the parameters and provide a more efficient estimation and more information about the parameters. Based on the available literature in our study, a set of potential variables were identified as the determinants of both foreign direct investment and economic growth in the BRICS countries.

Table 3. Description of the variables.

Variables	Symbol	Measurement	Source
FDI inflows	FDI	Total FDI inflows in USD millions	World Bank indicators (WDI)
Gross domestic product (GDP)	GDP	Gross domestic product at constant 2015 USD prices	OECD statistics
Trade openness	TO	Ratio between the sum of exports and imports and gross domestic product (GDP).	World Bank indicators (WDI)
Exchange rate	EXR	Exchange rate in USD	OECD statistics
Inflation rate	IR	Consumer price index (CPI)	World Bank indicators (WDI)
Human capital	HC	Index of human capital per person, based on years of schooling and returns to education	Penn Table (Version 10.0)

below discusses the details about the selected variables, measurements, and data sources. The annual data for all the variables were converted into their natural log form to obtain a more normalized data set. The data obtained are an estimated balanced panel data.

For the given panel dataset, this study employed the Bayesian vector autoregression (VAR) model for the estimation. The BVAR is a statistical approach that extends the standard VAR model by incorporating Bayesian inference principles. It employs prior beliefs about the relationships between variables, allowing for more effective parameter estimation and model selection, especially in cases with limited data. The Bayesian VAR model holds advantages over the standard VAR model, particularly in terms of robustness and avoiding overparameterization. The issue of overparameterization in the traditional VAR model results in parameter instability and imprecise estimates. Also, cutting back on variables in the VAR model to reduce the parameters may lead to an omitted variable bias, adversely impacting the structural analysis and forecasting. For this reason, Bayesian VAR modeling was used for the statistical analysis in this paper, as this technique has the ability to handle overparameterization issues. In standard VAR, determining the lag order becomes challenging as the number of variables grows, often leading to excessive parameters and unreliable estimates. Bayesian VAR addresses this by incorporating prior information, enabling more efficient parameter estimation and model selection. Moreover, Bayesian VAR enhances robustness by explicitly modeling uncertainty through posterior distributions. This robustness is valuable in small-sample situations or when data exhibit non-standard characteristics. Thus, Bayesian VAR mitigates overparameterization challenges and provides a robust framework for more reliable and accurate economic and financial analyses. Also, instead of using the factor method and high-frequency data, our paper adopted the Bayesian approach to address the issue of loss of precision, which is certainly a novelty aspect of this research paper.

The Bayesian vector autoregression (VAR) model can be expressed in matrix form as follows. Let y_t be a p × 1 vector of the observed variables at time t, where p is the number of variables in the system. The VAR model with lag order k can be written as:

y_t = A₀ + A₁y_t−1 + A₂y_t−2 + … + A_ky_t−k + ε_t,

where,

• A₀ is a p × 1 intercept vector,
• A₁, A₂, …, A_k are p × p coefficient matrices for lags 1 through k,
• y_t−1, y_t−2, … y_t−k are lagged values of the vector y_t,
• ε_t is a p × 1 vector of error terms at time t.

In the Bayesian VAR, priors are assigned to the coefficients A₀, A₁, …, A_k and the covariance matrix of ε_t, incorporating prior beliefs or information about the parameters. The posterior distribution of these parameters is then estimated using Bayesian methods, such as Markov Chain Monte Carlo (MCMC) techniques, allowing for a probabilistic assessment of parameter uncertainty. This approach offers greater flexibility and robustness than traditional VAR models by accounting for uncertainty and incorporating prior knowledge into the modeling process, leading to more stable and reliable parameter estimates. Thus, obtaining stable and reliable parameter estimates using the Bayesian VAR is another novelty of this paper.

The values of four parameters (often termed hyperparameters) must be chosen before a Bayesian VAR can be estimated using the Minnesota prior: μ1, λ1, λ2, and λ3. The persistence of the variables is represented by μ1. This value is zero when the variables are stable and near to one when the variables have a unit root. If the variables are considered stationary, their number is zero; otherwise, it is often close to one. All variables are considered to be persistent at the same rate. The degree of certainty in the given priors is denoted by λ1. The certainty ensures that the prior accurately identifies the nature of the data. While a value closer to zero indicates certainty, values of ten or higher indicate uncertainty about the nature of the data. The magnitude of the cross-variable effect is given by λ2. Simply put, it establishes the strength of the relationship between the present value of variable j and the previous values of variable i in the model. It accepts values ranging from 0 to 1. The strength of the cross-variable impact increases as the value approaches 1. The decay rate of a variable’s own lags is denoted as λ3. As the value of this parameter approaches 1, the impact of a variable’s previous values on its current state diminishes. The closer this parameter is to 1, the less it is impacted by its previous values.

4. Empirical Results

This section deals with all the research techniques and tools employed in this study and the results obtained. This paper used panel data analysis techniques to estimate the dynamic behavior of the determinants of FDI in the BRICS countries. The research began with a series of preliminary analyses, including descriptive statistics, a correlation analysis, and a panel unit root test, before carrying out the Bayesian VAR analysis. The descriptive statistics are displayed in

Table A1. Descriptive statistics.

	lnFDI	lnGDP	lnEXR	lnHC	lnIF	lnTO
Mean	23.20068	27.77184	2.150484	0.863235	4.023594	3.478765
Median	23.64495	27.82848	2.112515	0.856165	4.396755	3.589802
Maximum	26.39634	30.31422	4.305410	1.233845	5.230374	4.084114
Minimum	15.02686	25.91265	−8.645027	0.411515	−5.348304	2.050899
Std. Dev.	1.984469	1.027467	1.707677	0.210143	1.609364	0.484192
Count	150	150	150	150	150	150
Skewness	−0.883579	0.324771	−2.580344	0.038413	−3.999166	−0.780065
Kurtosis	4.466093	3.071128	15.92975	2.108752	20.80459	2.785701
Jarque-Bera	32.95173	28.68527	1211.320	5.001403	238.1106	15.49956

Note. The above table shows the descriptive statistics of the variables used in this study. All variables are normally distributed, supported by significant p-values, rejecting the normality hypothesis of Jarque–Bera test.

in Appendix A.

Table 4. Correlation Matrix.

	lnFDI	lnGDP	lnEXR	lnHC	lnIF	lnTO	VIF
lnFDI	1.000000
lnGDP	0.825749	1.000000					1.244267
lnEXR	0.127466	0.031140	1.000000				1.812751
lnHC	0.335853	0.190230	0.156609	1.000000			1.446364
lnIF	0.423007	0.176927	0.630212	0.167112	1.000000		1.788610
lnTO	0.035200	−0.179697	0.430628	0.471050	0.357072	1.000000	1.800326

Note: VIF—variance-inflating factor (Source: author’s own calculation).

below presents the results of the correlation matrix for the variables considered.

The results of the correlation matrix in Table 4 show that the independent variables had a low correlation amongst them, apart from FDI and GDP. However, the variance inflation factor (VIF) for GDP was 1.244267. Thus, in conclusion, the multicollinearity between the given independent variables was very low since the VIF values were within the (less than 10) range.

4.1. Panel Unit Root Test

The panel unit root test results are shown in

Table 5. Panel unit root test results.

Variable	T-Statistic	p Value	Decision
lnFDI	−2.93450	0.0017	I (0)
lnGDP	−2.45760	0.0070	I (0)
lnTO	−5.20500	0.0000	I (0)
lnEXR	−3.01996	0.0013	I (0)
lnIF	−1.62444	0.0521	I (0)
lnHC	−3.79966	0.0001	I (0)

Source: author’s own calculation.

below. To examine the stationarity and establish the integration level of the chosen variables, this study used the Levin, Liu, and Chu (LLC) test (Levin et al. 2002). The null hypothesis of this test assumes the presence of a common unit root in the panel, whereas the alternative hypothesis indicates the stationarity of the data. According to the LLC test, cross-sectional units are assumed to be unrelated to one another. The results indicate that all the variables, namely, lnFDI, lnGDP, lnTO, lnEXR, lnIF, and lnHC, were I(0) variables, i.e., stationary in their levels as per the Levin, Liu, and Chu test statistic.

4.2. Pedroni Residual Cointegration Test

The cointegration test is most commonly used to examine the spurious regression problem, which arises only when the underlying data are non-stationary. As a result, if the unit root tests show that each variable is stable, the question of whether or not they are in long-run equilibrium emerges. To address this question, the panel cointegration test by Pedroni (2001) was used in the analysis. Pedroni (1999, 2004) suggests two cointegration tests. The four statistics used in the panel tests, panel v, panel rho, panel PP, and panel ADF, are based on the within-dimension approach. These statistics combine the autoregressive coefficients across nations for the unit root tests on the estimated residuals. These statistics account for country- and time-specific variations. The group tests are based on the between-dimension approach, which includes three statistics, group rho, group PP, and group ADF statistics. These statistics are based on averages of the individual autoregressive coefficients associated with unit root tests of the residual of each country in the panel data set.

Table 6. Pedroni residual cointegration test results.

Null Hypothesis. No Cointegration Newey–West Automatic Bandwidth Selection and Bartlett Kernel
Alternative hypothesis: common AR coefs. (within-dimension)
	Statistic	Prob.	Weighted Statistic	Prob.
Panel v-statistic	0.594128	0.2762	0.122866	0.4511
Panel rho-statistic	−0.603543	0.2731	0.140408	0.5558
Panel PP-statistic	−4.641373	0.0000	−2.199960	0.0139
Panel ADF-statistic	−1.796088	0.0362	−0.571161	0.2839
Alternative hypothesis: individual AR coefs. (between-dimension)
	Statistic	Prob.
Group rho-statistic	0.886647	0.8124
Group PP-statistic	−3.141965	0.0008
Group ADF-statistic	−2.530919	0.0057

Source: author’s own calculation.

presents the results of the Pedroni residual cointegration test below. The results reveal that the null hypothesis, which states that there is no long-term dynamic relationship between FDI, economic growth, and trade openness in the BRICS nations, cannot be rejected at the 5% significance level. The findings indicate no correlation between foreign direct investment (FDI), economic growth proxied by GDP, and trade openness in the BRICS nations over the long-term period. Similar results have been reported previously by Carkovic and Levine (2005), Sarkar (2007), Tekin (2012), and more recently by Golitsis et al. (2018). This illustrates that long-term fluctuations in GDP are unrelated to long-term fluctuations in FDI in BRICS.

4.3. VAR Lag Order Selection Criteria

The determination of the appropriate lag length in a VAR model is crucial. This is because of the potential for errors when using a low number of lags and the risk of over-parametrization when employing an excessive number of lags. This study used five lag length criteria to select the optimal lag length for the estimation of the Bayesian VAR model. The results for VAR lag length selection criteria given in

Table A2. VAR lag order selection criteria.

Lag	LogL	LR	FPE	AIC	SC	HQ
0	−664.4858	NA	0.000830	9.933123	10.06225	9.985595
1	949.9620	3061.471	5.80 × 10−14	−13.45129	−12.54743	−13.08398
2	1077.678	230.8342 *	1.50 × 10−14 *	−14.81004 *	−13.13144 *	−14.12790 *
3	1104.565	46.20596	1.72 × 10−14	−14.67503	−12.22169	−13.67806

* Indicates lag order selected by the criterion. LR: sequential modified LR test statistic (each test at 5% level). FPE: Final prediction error. AIC: Akaike information criterion. SC: Schwarz information criterion. HQ: Hannan–Quinn information criterion. Source: author’s own calculation.

in Appendix A indicate that lag order 2 was the optimal lag for the model, implying that 2 lags were adequate to capture the dynamics and relationship among the given variables.

4.4. Bayesian VAR Model

The Bayesian VAR estimates are provided in

Table 7. Bayesian VAR results.

	lnFDI	lnGDP
lnFDI(−1)	0.019353 * (0.00985)	0.002315 * (0.00045)
lnFDI(−2)	0.003867 * (0.00498)	0.000573 * (0.00023)
lnGDP(−1)	1.070231 (0.10695)	0.771741 * (0.00492)
lnGDP(−2)	0.236412 ** (0.09784)	0.185567 * (0.00451)
lnTO(−1)	0.015067 ** (0.07431)	−0.013166 * (0.00339)
lnTO(−2)	0.003796 * (0.04063)	−0.003709 * (0.00186)
lnEXR(−1)	−0.009105 * (0.02309)	0.001630 * (0.00105)
lnEXR(−2)	−0.000179 * (0.01275)	0.000426 * (0.00058)
lnIF(−1)	0.126049 * (0.03149)	0.006239 * (0.00144)
lnIF(−2)	0.041447 * (0.01733)	0.002006 * (0.00079)
lnHC(−1)	0.743858 (0.53848)	−0.044155 * (0.02459)
lnHC(−2)	0.124693 (0.48614)	0.000551 * (0.02220)
C	−14.87816 (1.64782)	1.225716 (0.07529)
R-squared	0.798352	0.995640
Adj. R-squared	0.779299	0.995228
F-statistic	41.90090	2416.737

* p < 0.05, ** p < 0.10 denote the level of significance, t-statistics in the parentheses. (Source: author’s own calculation).

below. Due to space constraints, only the outcomes of FDI and GDP are displayed here. For the analysis, the values were as follows: μ1 = 0, λ1 = 0.01, λ2 = 0.99, λ3 = 1, and λ4 = infinite. The prior used for the estimation of BVAR was the Minnesota–Litterman prior with a univariate AR initial variance–covariance error matrix. The given results show that the coefficients of both the first and second lags of foreign direct investment (FDI) were significant and positively correlated with their current level of investment. The estimated coefficients associated with the first and second lags of GDP were equal to 1.070 and 0.236. However, only the second lagged value of this variable was significant at the 10% level, whereas the first lagged value was deemed to be insignificant. This suggests that the current FDI level in the BRICS countries is influenced by the GDP value from two years ago rather than fluctuations from the previous year. Furthermore, both the first and second lags of TO positively determined the current level of foreign direct investment (FDI) at the 10% level of significance, which indicates that an increase of 1% in trade openness (TO) will induce an increase in the level of FDI of 1.50% for the first lag (t–1) and 0.03% for the second lag (t–2). Similarly, IF exhibited a significant and positive relationship with the current level of FDI at a 5% significance level for both lags.

The relationship between the first and second lags of EXR with the current level of FDI was significant at the 5% level, but the relationship was negative between the two variables. With a one percent increase in the exchange rate (EXR), the foreign direct investment (FDI) will decrease by 2.3% for the first lag (t − 1) and 1.2% for the second lag (t − 2). Further, the human capital (HC) variable had a negligible impact on FDI in the case of the BRICS countries. In conclusion, the lagged values of FDI, TO, and IF positively impact the current level of FDI, whereas the lagged values of EXR have a significant but inverse relationship with the current level of FDI. The previous empirical studies conducted by Kiyota and Urata (2004) and Kechagia and Metaxas (2022) also reported a negative association between FDI inflows and exchange rate. The lagged values of HC do not affect the incoming FDI in BRICS countries (Mohanty and Sethi 2019). In the case of the current level of GDP, all variables for both lags, including FDI, GDP, TO, EXR, IF, and HC, were significant at the 5% significance level. This means that the current GDP level in BRICS countries is influenced by the first and second lags of GDP itself, as well as FDI, trade openness (TO), exchange rate (EXR), investment (IF), and human capital (HC). However, the relationship between trade openness and economic growth reported a negative association. Similar results have been reported in previous research conducted by Hye and Lau (2015), Huchet-Bourdon et al. (2018) and Amna Intisar et al. (2020). Thus, the null hypothesis stating that there is no co-movement between FDI, economic growth, and trade openness in BRICS nations is rejected at a 5% significance level.

4.5. Impulse Response Function

The impulse response function helps to understand how a variable responds to a sudden, one-time change (impulse) in another variable while keeping all other factors constant. The impulse response function of FDI to GDP in Figure 3 below indicates that when one standard deviation (S.D.) shock was given to GDP in the first period, the FDI immediately shots up and thereafter became stable from the second period.

In the case of trade openness (TO), FDI sharply increased in the first period before starting to decline from the second period onwards. A similar response was observed in the case of FDI when one standard deviation (S.D.) shock was given to the inflation rate (IF).

Next, the impulse response function of FDI to exchange rate (EXR) exhibited a sudden fall in FDI in the first period when a shock equal to one standard deviation (S.D.) was given to EXR. Following the brief decline, FDI rose again in the second period until the fourth period and showed a gradual decrease in the succeeding periods. Lastly, the impulse response function of FDI to human capital (HC) showed a sharp rise in the first period, followed by a steady decrease over time. Further, the impulse response function of GDP to FDI in Figure 4 below reveals that when one standard deviation (S.D.) shock was applied to foreign direct investment (FDI) in the initial period, it led to an instantaneous increase in gross domestic product (GDP), followed by a period of stability starting from the next period.

Analogous patterns can be observed in the exchange rate (EXR) and the inflation rate (IR). Conversely, when trade openness (TO) experienced a shock, GDP experienced a sharp decline in the first period but gradually began to recover from the third period onwards. Additionally, the impulse response of GDP to human capital (HC) showed a consistent decline across all periods.

4.6. Variance Decomposition

Variance decomposition provides insights into the dynamic interactions among the variables in the model. It helps to identify the key drivers of variability. It is a valuable tool for understanding the relative contributions and significance of different factors in explaining the fluctuations in a panel data series. It simply indicates how important a shock is in determining the variations in the variables.

The stacked graphs in Figure 5 depict the outcomes of the variance decomposition of FDI. The results indicate that in the first year, the FDI explains a 100% variance in the forecast error variance decomposition of FDI.

In the second year, the forecast error variance explained by FDI, GDP, and IF is 98.77%, 0.74%, and 0.45%, respectively, as shown in

Table A3. Variance decomposition results of FDI.

Variance Decompositon of lnFDI
Period	S.E.	lnFDI	lnGDP	lnTO	lnEXR	lnIF	lnHC
1	0.793470	100.0000	0.000000	0.000000	0.000000	0.000000	0.000000
2	0.799882	98.77906	0.743390	0.006221	0.001575	0.456189	0.013563
3	0.804408	97.80568	1.545449	0.008184	0.005019	0.611036	0.024636
4	0.807942	97.00761	2.304312	0.008112	0.008752	0.637328	0.033882
5	0.811151	96.28173	3.012193	0.008848	0.011010	0.645112	0.041106
6	0.814123	95.61341	3.669741	0.010676	0.012248	0.647352	0.046568
7	0.816901	94.99379	4.281606	0.013095	0.013031	0.647935	0.050541
8	0.819507	94.41754	4.851927	0.015691	0.013622	0.647910	0.053312
9	0.821955	93.88054	5.384294	0.018225	0.014142	0.647661	0.055142
10	0.824258	93.37940	5.881770	0.020588	0.014646	0.647336	0.056258

Source: author’s own calculation.

in Appendix A. In the short run, FDI is explained by its own lagged values. The role of other variables in the short run is limited. Likewise, 93% of FDI alone explains itself in the long run, suggesting that FDI plays a substantial role, followed by GDP, in influencing fluctuations more than the other variables. Further, the variance decomposition results of GDP, as shown in

Table A4. Variance decomposition results of GDP.

Variance Decompositon of lnGDP
Period	S.E.	lnFDI	lnGDP	lnTO	lnEXR	lnIF	lnHC
1	0.068016	1.768626	98.23137	0.000000	0.000000	0.000000	0.000000
2	0.086262	2.334837	97.42796	0.097602	0.036496	0.098998	0.004109
3	0.101864	2.648237	96.91751	0.162025	0.067883	0.193919	0.010425
4	0.114465	2.819908	96.60847	0.195504	0.092967	0.262586	0.020560
5	0.125256	2.924029	96.40376	0.213500	0.113225	0.311832	0.033656
6	0.134654	2.989469	96.26036	0.223191	0.129985	0.347528	0.049466
7	0.142969	3.031721	96.15380	0.228369	0.144233	0.374262	0.067616
8	0.150401	3.059156	96.07062	0.230917	0.156636	0.394866	0.087805
9	0.157101	3.076718	96.00286	0.231853	0.167653	0.411166	0.109754
10	0.163177	3.087469	95.94561	0.231758	0.177600	0.424347	0.133211

Source: author’s own calculation.

in Appendix A, indicate that in the first year, 98% of the variances in the forecast error are explained by GDP itself and only 1.76% by FDI. The remaining variables have no explaining power in both short-term and long-term periods.

5. Conclusions and Policy Implications

This research paper aims to understand the intricate relationship between foreign direct investment (FDI), economic growth, and trade openness within the BRICS nations, namely, Brazil, Russia, India, China, and South Africa, across three decades. The research aims to fill a significant gap in the literature by focusing on this crucial group of emerging economies, given their economic importance in the world economy in the coming decades. The panel data were collected from 1991 to 2020 from various sources, including World Bank indicators, OECD statistics, and Penn Table (Version 10.0). This extended time frame allowed for a thorough examination of the trends and variations in the BRICS countries. This paper used the Bayesian VAR framework to examine the dynamic interactions between FDI, economic growth, and trade openness variables. The Pedroni residual cointegration test was employed to check the existence of a long-run relationship between the given variables in the BRICS nations. The research findings provided evidence that foreign direct investment (FDI) exhibits a substantial interaction with economic growth in the short term within the BRICS countries. This research contributes to methodological innovation by introducing the Bayesian VAR framework, offering a deeper understanding of dynamic interactions among these key variables. The BVAR model holds various advantages over the traditional VAR model, specifically in terms of robustness and avoiding overparameterization of the model. The inclusion of this framework produces stable and reliable estimates, marking a novelty feature of this paper. The findings of the Bayesian VAR model highlight the significant role of FDI as a catalyst for immediate economic gains in these emerging economies. However, an intriguing dimension of this investigation emerges when considering the long-term relationship. The empirical evidence suggests that while FDI showcases a substantial link with economic growth in the short run, the relationship remains weak in the long run. Trade openness exhibits a positive correlation with foreign direct investment (FDI) but demonstrates a negative association with gross domestic product (GDP) in the short run. In practical terms, this suggests that while increased openness to international trade attracts more foreign direct investment (FDI), it may also be linked to a decrease in overall economic output or growth measured by GDP in BRICS countries. The intricate relationship between these variables emphasizes the need for a thorough assessment of the impact of trade policies on both FDI and GDP. This research study included competitiveness in the form of trade openness (TO). However, future research studies can explicitly include global competitiveness measures such as the global competitive index (GCI) or the easing of doing business index as a variable to study capital flows, both inflow and outflow.

Following the findings, several policy implications emerge from this study to exploit the full potential of short-term FDI impacts on economic growth in the BRICS nations. Governments should encourage and promote FDI through targeted investment promotion organizations and take sector-specific initiatives that highlight the immediate benefits of inward investment. Policymakers should ensure that infrastructure development aligns with sectors likely to receive short-term FDI inflows. This includes investments in transportation, energy, and communication systems. Moreover, flexible labor market policies can expedite job creation in response to short-term FDI inflows. Also, policies that facilitate skill development and workforce readiness are essential. Therefore, governments in these emerging economies should establish mechanisms to monitor the short-term impacts of FDI and adjust policies accordingly to maximize economic gains. However, the BRICS nations exhibit diverse economic structures and governance frameworks. Thus, the policy implications should be tailored to the specific circumstances of each country. Moreover, short-term FDI impacts can be influenced by global economic conditions, geopolitical events, and international market fluctuations, which are beyond the control of individual nations.

In summary, while this study recognizes the considerable short-term positive effects of FDI on economic growth in the BRICS countries, it also stresses the importance of strategic policy measures to optimize these impacts. Therefore, policymakers should adopt a dynamic and adaptable approach to leverage the immediate benefits of FDI while considering the broader context of sustainable economic development.