Determinants of Foreign Direct Investment Inflows in BRICS Nations: A Panel Data Analysis

Abstract

The aim of this study is to determine the most significant factors that lead to continuous rise in the foreign direct investment (FDI) inflows in the five most emerging economies of the world, that is, Brazil, Russia, India, China, and South Africa (BRICS). The present study employs panel data regression analysis to find out the most significant determinants affecting the FDI inflows in BRICS countries as a whole. Under this technique, all the three regression models, that is, common constant (ordinary least square [OLS]), fixed effects, and random effects are tested to explore the determinants of FDI in BRICS countries over a period of 31 years, that is, 1983–2013 (except for Russia whose data is available from 1995–2013 for the selected variables in the study). The empirical results of the modified random effects model reveal that industrial production index (IPI), inflation rates, unemployment rates, trade openness and real effective exchange rate (REER) are significant at 1 percent significance level whereas labor cost is statistically significant at 10 percent significance level which means that these are the most significant determinants in attracting FDI inflows in BRICS countries. This study is significant because there are various researchers who have contributed to the literature of determinants of FDI but there is hardly any study which documents the factors attracting FDI inflows in BRICS countries covering a long period of more than three decades.

Keywords

Foreign direct investment panel data analysis determinants of FDI BRICS emerging economies fixed effects model random effects model financial econometrics

Introduction

One of the noticeable tendencies in the world economy over the past three decades has been its increasing global economic integration. The world economy has seen an increasing internationalization with rising volumes of international trade and increasing levels of foreign investment flows, particularly that of the direct investment. However, the impact of globalization has been uneven across sectors, regions and countries. Some countries and regions have been able to get more integrated with the world economy with growing magnitudes of inward FDI flows and other types of cross-border transactions, while others have been left relatively untouched. The BRICS countries are a notable example of such world-wide economic shift that is evolving today.

The era of globalization has brought about drastic changes in the outlook of the leading manufacturers and producers around the world to explore investment opportunities in these fastest developing countries of the world, that is, BRICS countries. Because of the competitive advantage that these foreign investors get by investing in these BRICS countries their domestic counterparts are facing a huge economic exposure. Thus, this study will help such companies to understand and evaluate the factors that are motivating their local counterparts to become multinational and how they can also gain by investing in these emerging markets.

Second, many researchers and policy makers have opined that FDI is an important source of investment, capital formation, and growth for emerging markets. They assert that this becomes a reason for rising competition for such emerging market economies to attract more FDI flows. Thus, this study will enable the policy makers across the world to understand the major factors acting behind the rising FDI flows in these five countries and also motivate other least developed countries (LDCs) to move up from the category of LDC to emerging market economies like BRICS.

Third, this study has policy implications for the policy makers of other emerging market economies because by comparing the performance of their country with BRICS they will be able to learn about the areas in which they are lagging behind. Furthermore, by benchmarking emerging market economies against one another over the last three decades, the nations that have consistently performed well and the factors that contribute to their success can be identified.

This study is also significant because there are various researchers who have contributed to the literature of determinants of FDI but there is hardly any study which documents the factors attracting FDI inflows in BRICS countries. The objective of this study is thus to determine the most important economic determinants attracting FDI inflows in the context of BRICS countries as a whole for a long time period of 31 years.

An Overview of FDI Inflows in BRICS Countries

Over the last two decades, BRICS economies have become the most emerging economies of the world because of a continuously increasing share in world GDP and world trade, trade openness, foreign exchange reserves, and FDI inflows and outflows. Jim O’ Neill (2001) at Goldman Sachs coined the idea of the BRICS in his notable paper, “Building better global economic BRICS” when major structural changes were already taking place in the BRIC countries. Brazil had put in place a radical economic stabilization plan to turn around hyperinflation and boost privatization in the late 1980s, while India had introduced extensive economic reforms in the early 1990s. On the other hand, China had emerged unharmed from the Asian economic crisis of the late 1990s and Russia had started putting in place a strategy to rebuild its lost economic status.

Persistent economic activities coupled with a growth-oriented strategy in BRICS countries, since the 1990s, have resulted in significant infrastructural and other favorable changes. These changes have transformed BRICS countries into attractive destinations for FDI. The BRICS members are all developing or newly industrialized countries but they are distinguished by their large, fast-growing economies and significant influence on regional and global affairs; all five are G20 members. As of 2013, the five BRICS nations represent almost 3 billion people with a combined nominal GDP level of US$ 16.039 trillion and US$ 4 trillion approximately as combined foreign reserves. Appendix Table A1 presents the total FDI net inflows into BRICS by the host countries. The table reveals that the year 2013 was a strong year for BRICS with FDI net inflows of US$ 535.62 billion. In year 2013, China was the largest recipient of FDI inflows accounting for 65 percent of the total FDI flows in BRICS.

With this background, the study attempts to determine the various factors influencing the increased FDI inflows in these countries over a long period of 31 years (except for Russia for which the data was available from 1995). Also, the study throws a light on how this would channelize the efforts made by these economies in the right direction and enhance their efficiencies to become one of the supreme powers of the world.

The study is also motivated by the fact that BRICS appear to have prosperity of economic and social development in the forthcoming decades and it is expected that the economic growth of these countries will be tremendous in such a manner that they will throw competition and challenges toward the developed countries as well (Goldman Sachs Annual Report, 2013). Therefore, this study is important not just for the policy makers of these five countries but also for the other least developed countries (LDCs) and emerging market economies (EMEs) following the same path like the BRICS.

The rest of this article is structured as follows: The next section provides a “Literature Review” on the potential determinants of FDI. The section after that presents the “Data, Methodological Framework, and Modeling.” This is followed by “Empirical Results, and Discussions” section and “Findings” section. The next section summarizes the findings and results along with the policy implications of the study, while the final section highlights the limitations and future scope of this study.

Literature Review

FDI is assumed to facilitate economic growth in emerging markets by providing innovative technological know-how, capital, and access to diverse markets for the production of goods and services. However, attracting FDI is a major challenge for host countries as they need to identify the major push and pull factors that attract FDI to their countries.

All the below mentioned studies, point the role of economic factors in attracting FDI flows in emerging markets but there are hardly any studies which have examined the determinants of FDI in emerging markets as a group. Therefore, we extend previous studies by analyzing the economic determinants affecting the FDI inflows in BRICS countries as a whole over a very long period of three decades.

There are several studies which have analyzed the determinants of FDI for individual countries or groups of countries that are part of developing or developed markets. A brief summary of such studies is presented in this section. Marcelo Braga Nonnemberg and Mario Jorge Cardoso de Mendonca (2004) concluded that FDI is correlated to the level of schooling, economy’s degree of openness, inflation, risk, and average rate of economic growth in developing countries. Khondoker Abdul Mottaleb and Kaliappa Kalirajan (2010) also demonstrated that developing countries with larger GDPs, higher GDP growth rates, higher proportion of international trade, and a more business-friendly environment are more successful in attracting FDI.

Various studies pertaining to specific developing countries include studies by Boopen Seetanah and Sawkut Rojid (2011) who emphasized on trade openness, wages, and the quality of labor as the most instrumental factors for attracting FDI in Mauritius. Andreia Alexandra Faria Severiano (2011) showed the importance of GDP per capita, degree of openness to trade, exchange rate, minimum wage, corporate tax rate, and labor market flexibility while analyzing the determining factors for FDI in primary, secondary, and tertiary sectors of Portugal. Another study which confirmed the results of the previous studies was by Gilmore et al. (2003) who concluded that FDI is a preference to other forms of foreign market entry, size and growth of the host market, government emphasis on FDI and financial incentives, economic policy, cultural closeness, costs of transport, materials and labor, resources, technology, political stability and infrastructure are important determinants of FDI for Northern Ireland and Bahrain. Another study by Ngoc Anh Nguyen and Thang Nguyen (2007) in the Vietnam context gave importance to market, labor, and infrastructure in attracting FDI. Elizabeth Asiedu (2002) suggested that good infrastructure, liberalized trade regimes, and better government policies can help Africa to get more FDI flows. The main result is that natural resources and large markets promote FDI (Elizabeth Asiedu, 2006). Openness to trade, the size of the domestic market, and stock of the human capital played a positive role, while political instability and labor cost a negative role in attracting FDI in the African markets as explored by Sawkut et al. (2009). Yuko Kinoshita and Nauro F. Campos (2003) explored institutions, agglomeration, and trade openness as main determinants for FDI inflows in transition economies.

A study in China by C. H. Ho Owen (2004) indicated that market size, wage rate, degree of economic reform, and innovation activities are important determinants of sectoral FDI in China. Chien-Hsun Chen (1996), Lv Na and W. S. Lightfoot (2006) determined that market expansion potential, labor cost, allocative efficiency, transportation linkages, and technological filtering are able to attract more FDI in the mainland China. Junjie Hong (2008) also emphasized the labor cost as an essential determinant for attracting FDI flows in China. DeAngelo et al. (2010) concluded the importance of the consumer market and strength of consumer sales as the most important factor in explaining capital movements into Brazil. Bergsman et al. (1999) suggested that Russia should switch to a more modern policy approach to FDI by eliminating the relatively extensive non-tariff protection given to the domestic market, phasing out existing tax preferences for foreign investors, and reducing significantly restrictions on FDI to a limited number of activities. Priya Gupta and Archana Singh (2014) also concluded in a very recent study on BRIC Nations that the most important factors for attracting FDI inflows are inflation rate, international liquidity, debt service as a percentage of export of goods and services of the country, current account as percentage of GDP, current account as percentage of export of goods and services, budget balance as a percentage of GDP, and percentage unemployment in the country.

There are various studies pertaining to determinants influencing the FDI inflows in India specifically or as a part of some group like Vijayakumar et al. (2010) concluded that market size, labor cost, infrastructure, currency value, and gross capital formation are the potential determinants of FDI inflows of BRICS countries. Another study in the context of BRICS countries was conducted by Ranjan and Agrawal (2011) who found that the most important determinants of FDI inflows in BRICS are market size, trade openness, labor cost, infrastructure facilities, and macroeconomic stability and growth prospects. Market size and trade openness were also highlighted by Jadhav (2012) in his paper on BRICS countries. Muhammad Azam and Ling Lukman (2010) also revealed that market size, external debt, domestic investment, trade openness, and physical infrastructure are the important economic determinants of FDI in Pakistan, India, and Indonesia. Balasundram Maniam and Amitava Chatterjee (1998) conferred the importance of exchange rate for attracting US FDI in a developing country like India. Monica Singhania and Akshay Gupta (2011) revealed that GDP, inflation rate, scientific research, and FDI Policy changes have had a significant impact on FDI inflows into India. The financial strength of the state, development level of the state, size of the market, and level of infrastructure were some other determining factors for FDI in India as studied by Neerja Dhingra and H. S. Sidhu (2011).

Thus, it can be seen from the above literature that there may be various studies regarding the determinants of FDI inflows in the context of both developed and developing economies but there is hardly any study which explores these determinants of FDI inflows in the context of BRICS countries as a whole. From the review of the literature regarding determinants of FDI it has been observed that researchers agreed about the impact of many variables on FDI, but there is lack of uniformity in the opinion regarding the influence of some variables like inflation, exchange rate, trade openness, GDP, labor cost, and international liquidity on FDI inflows. This necessitates reinvestigation of factors influencing FDI in case of BRICS nations.

Potential Variables Used in the Study Determining the FDI Inflows

Based on the discussed literature review, our study estimates a set of potential determinant variables that influence the FDI flows. The detailed definitions, measurements, and expected relationship of the selected independent variables of this study are reported in Table 1.

Table 1.

Explanatory Variables Definition, Measurement and Their Expected Relationship with Dependent Variable

Variable	Definition	Measurement	Expected Relationship with Dependent Variable
Y	Foreign direct investment	Log of net inflows of FDI in current US$ bn.	Dependent variable
X1	Market size and growth prospects	GDP annual percentage growth rate	+
X2	Industrial production index	Level of industrial production index (IPI)	+
X3	Inflation rates	Log of annual percentage consumer price index (CPI)	+/-
X4	Unemployment rates	Log of percentage of total labor force	-
X5	Trade openness	Log of sum of export plus import of goods and services in current US Dollars as a percentage of gross domestic product in current US Dollars	+
X6	Exchange rate	Log of real effective exchange rate (REER) index	+/-
X7	Gross capital formation	Annual percentage growth rate	+/-
X8	International liquidity	Log of import cover ratio (measured as the proportion of foreign reserves in current US Dollars and monthly imports of goods and services in current US Dollars)	+
X9	Labor cost	Net workers’ remittances and compensation in US Dollars	-

Source: Author’s own.

Data, Methodological Framework, and Modeling

The dependent variable in this study is the FDI inflow in US billion dollars and the independent variables that are expected to determine FDI flows are carefully chosen, based on the previous literature and availability of dataset for the selected period. The dataset consists of annual dataset from 1983–2013 for the four emerging economies, namely, Brazil, India, China, and South Africa and for Russia the data is available from 1995–2013. The required dataset for the BRICS countries was obtained from various sources like the World Development Indicators, published by World Bank, World Economic Outlook published by International Monetary Fund (IMF), Oxford Economics Annual Database and Bruegel Database.

In order to evaluate the potential determinants of FDI inflows for the BRICS countries as a group, panel data analysis (Balestra, 1992) has been employed. A panel is a cross-section or group of people who are surveyed periodically over a given time span. A panel dataset offers several econometric benefits over traditional pure cross-section or pure time series datasets. Panel data analysis is being used extensively in economics and finance research to study cross-country economic issues (Maddala, 1999; Webb & Hall, 2009). The most obvious advantage is that the number of observations is typically much larger in panel data, which will produce more reliable parameter estimates and, thus enable us to test the robustness of our linear regression results. Panel data also alleviates the problem of multicollinearity because when the explanatory variables vary in two dimensions (cross-section and time series), they are less likely to be highly correlated.

Along with the common constant model (OLS regression), both fixed effects (FE) model and random effects (RE) models have also been tested to explore the key determinants of FDI inflow into BRICS countries due to the fact that the former takes into consideration the country-specific effect and the latter considers the time effect.

(a) The Common Constant Model (also called as pooled OLS method): The null hypothesis is that all the constants are the same (homogeneous), and thus, the common constant method is applicable:

H_{0} = α_{1} = α_{2} = α_{N}

(1)

However, this case is quite restrictive and the case of more interest involves the inclusion of fixed and random effects in the method of estimation.

(b) Fixed Effects (FE) Model: This model treats the constant as group specific, that is, there are different constants for each groups.

The FE model is specified as under:

y_{i t} = α_{i} + \sum_{k = 1}^{k} X_{i t k} β_{k} + ε_{i t}

(2)

where

i = 1,2,…..,N

t = 1,2,…..,T

where $y_{i t}$ represents the value of the dependent variable, that is, FDI inflows in cross-section i (five countries in our case); T is the length of time series, that is, 1983–2013 for Brazil, India, China, and South Africa and 1995–2013 for Russia; k is the number of independent variables explaining the dependent variable. The term $α_{i}$ denotes unobserved country-specific effects that are assumed to be fixed over time and different across country i. $X_{i t}$ and $β$ represent the vectors of explanatory variables and their parameters respectively. The subscript i indicates individual countries, while t shows different time periods. $ε_{i t}$ represents the vector of the error component which is assumed to be independently distributed across i and over t with mean zero and variance σ2.

(c) Random Effects (RE) Model: This model of estimation handles the constants for each section as random parameters rather than fixed or constant. Hence, the variability of the constant for each section comes.

In the RE case, the model is defined as

y_{i t} = μ + \sum_{k = 1}^{k} X_{i t k} β_{k} + v_{i t}

(3)

where

i = 1, 2,…..,N

t = 1, 2,…..,T

$v_{i t} = α_{i} + ε_{i t}, t = 1, . \dots,$ T are the composite errors

For each t, $v_{i t}$ is the sum of the unobserved effect and an idiosyncratic error (Wooldridge, 2010). $α_{i}$ are assumed to be independently distributed across i, with mean zero and variance $σ_{α}^{2}$ and uncorrelated with $X_{i t} .$ The error term $ε_{i t}$ is assumed to be independently distributed across i and over t with mean zero and variance σ2.

Finally, it can be seen that in the panel data analysis, the fixed effects model assumes that each country differs in its intercept term whereas the random effects model assumes that each country differs in its error term.

Hausman Specification Test

Generally, when the panel is balanced it is expected that the fixed effects model would work well and when the panel is unbalanced, the random effects model would be a better model. However, the Hausman specification test (1978) guides us to choose the appropriate panel data model. Therefore, in such case following hypotheses are tested:

Null Hypothesis: H₀: Cov (α _i , X _it ) = 0, that is, random effect model is suitable, if null hypothesis is accepted.

Alternate Hypothesis: H_a: Cov (α _i , X _it ) ≠ 0, that is, fixed effect model is suitable, if alternate hypothesis is accepted.

If p < 0.05 → FE is suitable

If p > 0.05 → RE is suitable

where

p is the probability value of the test statistic

If the value of the statistic is large, then the difference between the estimates is significant, so the null hypothesis can be rejected and a fixed effects model can be used. On the contrary, a small value of the Hausman statistic implies that the random effect is a more appropriate estimator.

The general specification of the parameters (dependent and independent variables) of the model in this study is as follows:

Y_{i t} = α + β_{1} X 1_{i t} + β_{2} X 2_{i t} + β_{3} X 3_{i t} + β_{4} X 4_{i t} + β_{5} X 5_{i t} + β_{6} X 6_{i t} + β_{7} X 7_{i t} + β_{8} X 8_{i t} + β_{9} X 9_{i t} + μ_{it}

(5)

In the above specification, $Y_{i t}$ is the log of net inflows of foreign direct investment (FDI) in current US Dollars for country i at time t. This represents the dependent variable of the study.

The right hand side of the specification model includes all the independent variables which are defined as follows:

X1_it is the gross domestic product (GDP) (annual percentage growth rate) used as a proxy of market size and growth prospects for country i at time t.

X2_it is the industrial production index (IPI) for country i at time t.

X3_it is the log of inflation rates (annual percentage consumer price index) for country i at time t.

X4_it is the log of unemployment rates (measured as percentage of total labor force) for country i at time t.

X5_it is the log of trade openness (measured as the sum of export plus import of goods and services in current US Dollars as a percentage of gross domestic product in current US Dollars) for country i at time t.

X6_it is the log of real effective exchange rate used as a proxy of exchange rate movements for country i at time t.

X7_it is the gross capital formation (annual percentage growth rate) for country i at time t.

X8_it is the log of import cover ratio (measured as the proportion of foreign reserves in current US Dollars and monthly imports of goods and services in current US Dollars) used as a proxy of international liquidity for country i at time t.

X9_it is the labor cost (measured as the net workers’ remittances and compensation in US Dollars) for country i at time t.

μ_it is the stochastic disturbance term.

The following hypotheses have been formulated with the help of carefully chosen independent variables (based on the availability of data for the selected period of study):

H1: Gross domestic product (GDP) does not have a significant impact on FDI inflows of country i at time t.

H2: Industrial production index (IPI) do not have a significant impact on FDI inflows of country i at time t.

H3: Inflation rates do not have a significant impact on FDI inflows of country i at time t.

H4: Unemployment rates do not have a significant impact on FDI inflows of country i at time t.

H5: Trade openness does not have a significant impact on FDI inflows of country i at time t.

H6: Exchange rates do not have a significant impact on FDI inflows of country i at time t.

H7: Gross capital formation does not have a significant impact on FDI inflows of country i at time t.

H8: International liquidity does not have a significant impact on FDI inflows of country i at time t.

H9: Labor cost does not have a significant impact on FDI inflows of country i at time t.

Empirical Results and Discussions

While applying the panel data models, various statistical tests need to be applied to check which model out of the three estimated models (explained in the previous section) is the model of best fit. Thus, first, the study checks whether FE model is better than the OLS model. For this purpose, the standard F-test can be used to analyze whether fixed effects (i.e., different constants for each group) should indeed be included in the model or not.

If F-statistics is bigger than the F-critical then the null hypothesis can be rejected (which assumes that all the constants are homogeneous) and therefore, the FE model should be used as a model of estimation.

Second, the comparison is made between the FE model and RE model in which Hausman specification test can be useful. If the value of the statistic is large and the difference between the estimates is significant, the null hypothesis can be rejected (which assumes the Hausman statistic is asymptotically distributed as chi-square with k degrees of freedom) and the fixed effects model can be used. On the contrary, a small and non-significant value of the Hausman statistic implies that the random effect model is a more appropriate estimator.

Lastly, the Breusch–Pagan Lagrange multiplier (LM) test (1980) is also computed to test whether the RE model is preferable over the common constant model (OLS regression). In other words, it can be proved with the help of this test whether there is an evidence of significant differences across countries or not. Otherwise, a simple OLS regression model can be run.

As a part of preliminary investigation, the descriptive statistics and correlation matrix of the selected variables in the study is given in Tables 2 and 3 respectively.

Table 2.

Descriptive Statistics of Dependent and Independent Variables in the Study

Variable	No. of Observations	Mean	Standard Deviation	Minimum Value	Maximum Value
Y	143	1.680824	2.430166	–5.898806	6.301514
X1	143	5.097372	4.224576	–7.820885	15.17602
X2	143	4.234483	0.6029982	2.093307	5.435548
X3	143	2.473732	1.362577	0.3259758	7.988792
X4	143	2.071456	0.7112716	0.5877867	3.325036
X5	143	–1.0767	0.5065465	–2.119541	–0.3165565
X6	143	4.594341	0.2529566	3.910074	5.290761
X7	143	6.683131	13.34898	–45.2	75.2
X8	143	1.836274	0.7456074	–0.1712762	3.275048
X9	143	0.4592223	2.83933	–11.87157	13.22716

Source: Authors, own.

Note: Author’s own compilation based on the computations done on STATA (Version 12.0).

Table 2 suggests that all the selected variables in the study have equal number of 143 observations. This means that the panel is balanced with no missing observations. The results in Table 2 also displays that gross capital formation (X7) has the highest mean value and standard deviation of 6.683131 and 13.34898 respectively in the data distribution.

Some more important information can be gathered from Table 3 which reveals that most of the important determinants of FDI inflows in the model have turned out to be in the expected direction as defined under Table 1 earlier. These signs of determinants have been expected as per the available theories in literature that is reviewed under the “Literature Review” section earlier. Only X3 (inflation rates), X4 (unemployment rates), and X6 (exchange rates) are showing a negative sign meaning, thereby, an inverse relationship with the dependent variable, that is, FDI inflows are adversely affected by the upward movement in these three variables. Other than these three variables, X1 (GDP growth rates), X2 (industrial production index), X5 (trade openness), X7 (gross capital formation), and X8 (international liquidity) are positively correlated with the FDI inflows as expected. Further, the results in Table 3 display that none of the independent variables are found to be correlated with each other, hence, solving the multicollinearity problem in the model.

Table 3.

Correlation Matrix of the Selected Variables in the Study

Variable	Y	X1	X2	X3	X4	X5	X6	X7	X8	X9
Y	1.0000
X1	0.2693	1.0000
X2	0.3997	–0.1920	1.0000
X3	–0.2014	–0.2722	–0.0921	1.0000
X4	–0.4073	–0.5292	0.3685	–0.1967	1.0000
X5	0.3605	0.0607	0.5339	–0.3018	0.2356	1.0000
X6	–0.0925	0.0876	–0.1411	–0.4058	0.0307	–0.0653	1.0000
X7	0.1087	0.6477	–0.0474	–0.1575	–0.2137	0.0246	0.0341	1.0000
X8	0.5407	0.4351	0.1071	–0.1440	–0.5916	–0.0640	–0.0197	0.2472	1.0000
X9	0.1012	0.1531	0.1809	–0.0500	–0.0623	0.0570	–0.0100	–0.0588	0.2841	1.0000

Source: Authors’ own.

Note: Author’s own compilation based on the computations done on STATA (Version 12.0).

In order to capture the distribution of FDI across the BRICS countries over a period of 30 years (except for Russia), the estimates of Equation (3) were generated with the following panel data linear regression models: (1) common constant (OLS regression) model (2) FE model, and (3) RE model.

The estimation results of all the three models, that is, common constant (OLS), FE, and RE, explained earlier, are presented in Table 4.

Table 4.

Determinants of FDI Inflows as per the Three Models of Panel Data Analysis

Dependent Variable: FDI Inflows
Independent Variables	Common Constant Model (OLS)	Fixed Effects Model (FE)	Random Effects Model (RE)
Intercept	4.536113(1.38)	–0.4396563(–0.11)	4.536113[1.38]
X1	–0.0617494(–1.13)	0.0040654(0.07)	–0.0617494[–1.13]
X2	1.989922(6.14)*	1.800015(3.74)*	1.989922[6.14]*
X3	–0.5326794(–4.05)*	–0.5444495(–3.91)*	–0.5326794[–4.05]*
X4	–2.36331(–7.24)*	–1.553407(–2.95)*	–2.36331[–7.24]*
X5	0.862191(2.54)**	1.2013(1.41)	0.862191[2.54]**
X6	–0.9551579(–1.66)***	0.0367769(0.05)	–0.9551579[–1.66]***
X7	–0.0040989(–0.31)	–0.0076622(–0.57)	–0.0040989[–0.31]
X8	0.332553(1.24)	0.1190359(0.38)	0.332553[1.24]
X9	–0.0440807(–0.91)	–0.0013914(–0.03)	–0.0440807[–0.91]
Mean VIF	2.17
F-Test	-	18.13*	-
Wald c2	-	-	228.22*
R² within	-	0.5585	0.5382
R² between	-	0.5480	0.9334
R² overall	0.6318	0.5499	0.6318
Hausman Test (p-value)	-	16.86(0.0510)
Breusch–Pagan Lagrange multiplier (LM) Test: Var(u) = 0	-	-	c² = 0.00Prob > c² = 1.0000

Source: Authors’ own.

Note: Parentheses () and [] show the t-value and z-statistics respectively. *, **, and *** denote the significance at the 1 percent, 5 percent, and 10 percent levels respectively. This table is author’s own compilation based on the computations done on STATA (Version 12.0).

From Table 4, it can be seen that the first column (common constant model) shows the estimation results of regression Equation (5). As pointed out earlier, the problem with OLS methodology is that it implies that there are no differences between the estimated cross-sections (BRICS in our case) and it is useful under the hypothesis that the dataset is a priori homogeneous. Therefore, this case is quite restrictive but it allows us to check for the existence of multicollinearity in the model by way of a variance inflation factor (VIF). The literature points out that there is indication of multicollinearity if the mean VIF is greater than 5 (Judge et al., 1982). In Table 4, the mean VIF from the OLS method is found to be 2.17 which implies that there is no indication of multicollinearity problem in the model considered. Besides, the F-test (OLS versus FE) reveals that the null hypothesis (OLS model) is rejected. So, FE model is preferred to common constant (OLS) model.

In Table 4, the estimation results of the FE model and RE model are also reported in column three and four respectively. Now, the next concern is the choice between FE and RE models. To select an appropriate model for the empirical analysis, the Hausman specification test was conducted. The test gave a chi-square value of 16.86 which is not significant at 5 percent significance level, suggesting that the generalized least squares (GLS) estimators of the RE model are the preferred ones.

Lastly in Table 4, when the RE Model was compared with the common constant model with the help of the Breusch–Pagan Lagrange multiplier test, it was observed that the Prob > χ2 is more than 0.05. Therefore, we failed to reject the null (that variance across the five countries is zero) and concluded that random effects model is not appropriate. In other words, there is no evidence of significant difference across countries and thus we can run a simple OLS regression. But, we are still in the favor of random effects model instead of common constant model because of two reasons, one that the results of the latter were observed to be similar with the former and two, because the Hausman specification test also favored the RE model in comparison with FE model. Moreover, the random effects model has more estimation advantages than the common constant model, since the data classification seems to be a priori homogeneous. The estimation ensures homogeneity by choosing the sample countries which are assumed to be the most emerging economies of the world in terms of their growth and market potentials. Therefore, it was decided to drop the common constant model and not discuss its results further in this study.

However, once the above results are arrived at, one needs to analyze the residual diagnostics like the test of heteroskedasticity, the test of cross-sectional dependence/contemporaneous correlation, and also the test of serial correlation of both the models to derive the model of best fit to explain the dependent variable (FDI inflows in our case). If in case any of these assumptions pertaining to residual diagnostics are violated, then corrections are to be made in the aforesaid models by incorporating these assumptions and a robust model can be arrived at with the help of “cluster command” in case of FE and RE models in STATA (the statistical software). This will be done to make sure about the robustness of parameter coefficient in explaining the factors that determine the FDI inflows to the BRICS countries.

Modified Wald test for group-wise heteroskedasticity for both the FE and RE models gives a chi-square value of 1435.29 that is statistically significant at 5 percent significance level. Thus, we can conclude that there is substantial amount of heteroskedasticity in both the models. Another test for evaluating the cross-sectional dependence/contemporaneous correlation among the residuals is the Breusch–Pagan Lagrange multiplier test of independence which showed that for the FE and RE models, the chi-square distribution is 16.635 which is not significant at 95 percent confidence interval and thus we can not reject the null hypothesis, that is, there is no cross-sectional dependence or residuals across countries are not correlated. Lastly, the test for serial correlation is conducted where the F-statistic comes out to 11.393, which is statistically significant at 0.05 level, implying that the residuals have the problem of autocorrelation in both FE and RE models.

From the above results, it is evident that both the models are suffering from the problem of heteroskedasticity and are also violating the assumption of no autocorrelation among the residuals. Therefore, for FE and RE models, the cluster option in STATA can be applied along with the regression command to produce robust standard error estimates for such linear panel models. This would solve the problem of both heteroskedasticity and autocorrelation (Daniel Hoechle, 2007). The results as per both modified models are presented in Table 5.

Table 5.

Determinants of FDI Inflows as per both the Modified Models of Panel Data Analysis (with Cluster Option)

Dependent Variable: FDI Inflows
Independent Variables	Fixed Effects Model–Cluster Option (FE)	Random Effects Model–Cluster Option (RE)
Intercept	–0.4396563(–0.10)	4.536113[2.35]**
X1	0.0040654(0.09)	–0.0617494[–1.37]
X2	1.800015(3.54)**	1.989922[11.13]*
X3	–0.5444495(–5.42)*	–0.5326794[–5.92]*
X4	–1.553407(–2.50)***	–2.36331[–16.57]*
X5	1.2013(1.15)	0.862191[2.68]*
X6	0.0367769(0.03)	-0.9551579[–4.43]*
X7	–0.0076622(–1.32)	–0.0040989[–1.18]
X8	0.1190359(0.66)	0.332553[1.33]
X9	–0.0013914(–0.10)	–0.0440807[–1.91]***
R2 within	0.5585	0.5382
R2 between	0.5480	0.9334
R2 overall	0.5499	0.6318

Source: Authors’ own.

Notes: Parentheses () and [] show the t-value and z-statistics, respectively. *, **, and *** denote the significance at the 1 percent, 5 percent, and 10 percent levels respectively. This table is author’s own compilation based on the computations done on STATA (Version 12.0).

In the above table, we have estimated the fixed effects (FE) model and random effects (RE) model with the cluster option for the selected study period. The robustness of parameter coefficients is used to explain the relationship between FDI inflows and the selected independent variables. The fixed effects model is rejected in the analysis based on the Hausman specification test (1978). Therefore, we shall discuss only the results of random effects model (with cluster option) as this is the model of best fit.

Findings

The empirical results of the random effects model (with cluster option) presented in Table 5 reveals that IPI is significant and positively affecting FDI inflows as expected. Both inflation rates and unemployment rates are also significant and show a negative sign symbolizing an adverse effect on FDI inflows (results confirm with Gupta & Singh, 2014; Singhania & Gupta, 2011; Nonnemberg & Mendonca, 2004). It implies that as the inflation and unemployment rates increase in these emerging economies the resultant effect on FDI inflows is negative as the MNCs might not like to enter in such a market where their cost of production would increase. This might cut down their margins and thus acts as a demotivating factor for them while analyzing the alternative of making long-term investment into BRICS countries. A positive coefficient of trade openness, as expected, shows that all the BRICS countries are following a liberal trade regime and are thus able to successfully attract increasing FDI inflows. On the other hand, the results reveal a negative sign of coefficient of REER implying that there is an inverse relationship between FDI inflows and exchange rate movements in the host country. In other words, this implies that a depreciation of the host currency would increase FDI into the host country and an appreciation of the host currency would decrease FDI. All these five determinants are statistically significant at 1 percent significance level. The relationship between labor cost and FDI inflows is also negative as expected in the literature of other developing countries and is statistically significant at 10 percent significance level. This finding is also supported by the findings of M. V. Posner (1961) and Seev Hirsch (1976). It was concluded from these studies that the direction of the trade is decided by the comparative costs existing in these different countries which the MNEs take as absolute costs for taking decisions on investment flows.

Thus, we can conclude that these six determinants are the most crucial factors in attracting FDI inflows in BRICS countries.

Besides, the results also show that other variables such as market size (proxy used gross domestic product (GDP) growth rates) and gross capital formation have a very low and insignificant coefficient value and an opposite sign from the expected direction implying that they are not very relevant for the BRICS countries for attracting more FDI inflows. Libor Krkoska (2001) also concluded that the relationship between FDI and capital formation is not simple. Usually, a better investment climate leads to higher FDI inflows but in certain cases of privatization it might not lead to any increase at all or may even lead to a reduction in future FDI inflows. The sign of the coefficient of international liquidity (proxy used import cover ratio) is although found to be positive, as expected, but is statistically insignificant implying that even this does not play any significant role in attracting FDI inflows in BRICS countries. Finally, it can be asserted from the findings that most of the determinants are behaving in the same direction as expected in other developing countries of the world and are extremely useful (99 percent confidence level for all except labor cost) in attracting FDI inflows in BRICS.

The findings are well supported by the theories that exist in the international business environment which determines the movement of FDI flows to a specific country.

Conclusion and Policy Implications

The findings indicate that the governments of the BRICS countries should put incremental efforts to enhance production levels because a growing IPI indicates that the companies in the industry are performing well which attracts more FDI in such sectors. The BRICS countries should also adopt more open and liberal trade philosophy, provide tax incentives, etc. that will provide more confidence to the foreign investors to invest their capital in these countries. The government of these most emerging economies of the world should also keep a check on unemployment rates, inflation rates, exchange rate movements, and labor cost as these might act as a deterrent toward FDI inflows.

The results of this study also suggest that each of these five countries should put in individual efforts also to have more integration among each other. For instance, in case of Russia it can be seen that the country has undergone several reforms after the collapse of Soviet Union but still these reforms, as they were expected to, were not fully able to decrease concentration of economic power. Other than this, the country is still extremely weak in terms of protection of intellectual property rights and has an interfering atmosphere created by the government. Although Russia is amongst the leaders of the world for producing oil and natural gas but its manufacturing sector is still not competitive in the world. It is because of its reliance on commodity exports which are vulnerable to economic cycles that are prone to volatility in global prices. Therefore, the Russian government should put more emphasis on economic reforms and loosening of strict foreign sanction parameters.

The case of South Africa is quite similar to Russia, as although South Africa is one of the emerging markets in the world because of its vast supply of natural resources but it suffers from some major economic challenges leading to a retarded economic growth that in turn detains the foreign investors to keep their stakes high. Inadequate infrastructure facilities (like unstable electricity supplies) coupled with three major economic problems of unemployment, poverty, and income inequality (which are among the highest in the world) are the biggest challenge for the South African government. There are other lacunas also that hamper the growth rate of this economy like shortage of skilled labor, global incompetence, and frequent stoppage of work due to strikes and lockouts. Policy measures to resurrect the system are required at the priority level for the present government.

It is also suggested that the countries like Brazil, India, and China also need to sustain their positions in the competitive world. Like in case of Brazil, since GDP growth has slowed post 2011, due to several factors like heavy reliance on exports of raw commodities, low productivity and high costs of operations, high inflation rate, and low levels of investment, the government of Brazil should bring reforms to resurrect the system. Policy measures like promoting the domestic manufacturers by providing them tax reliefs and financial packages, educating them about new avenues in the manufacturing sector, imposing local content, and technology transfer requirements on foreign businesses, etc. might prove to be useful for bringing more FDI inflows in future if policy makers start considering them on a priority basis.

For China, friendly business climate, structural changes, better infrastructure facilities to promote exports, strategic policy initiatives of providing economic freedom, and flexible laws can be identified as driving forces for attracting FDI. While India has risen due to its human capital, size of the market, rate of the growth of the market, and political stability. In case of India, to enhance more FDI inflows, the policy makers need to ensure more economic and political stability, better infrastructure facility, a peaceful environment having a proper law and order mechanism, reduce the external liabilities, and also appropriate monetary and fiscal policy changes should be brought in for global integration.

Limitations and Scope of the Study

A similar or more comprehensive study can be extended to other upcoming groups like PIN (Pakistan, Indonesia, and Nigeria), MINT (Mexico, Indonesia, Nigeria, and Thailand), etc. which may follow the path of BRICS economies in formulating their strategies for attracting better FDI inflows. The most important limitation of this study is that the variables used in the study do not include the noneconomic/institutional variables. Therefore, further study with both economic and institutional variables is, henceforth, proposed. Another limitation is that this study looks at determinants of FDI for BRICS as a whole. Despite the fact that these five countries are classified as emerging markets there are differences among them on institutional and economic frontiers. Therefore, it will be interesting to analyze the determinants of FDI for each of these countries individually.

Footnotes

Appendix

Table A1. FDI Net Inflows (US Billion $) in Brazil, India, China, and South Africa (1983–2013); Russia (1995–2013)

Year	Brazil	Russia	India	China	South Africa
2013	80.843	70.65372	28.15303	347.8487	8.118154
2012	76.11066	50.58755	23.99569	295.6256	4.626029
2011	71.53866	55.08363	36.49865	331.5917	4.139289
2010	53.34463	43.16777	27.39689	272.9866	3.693272
2009	31.48093	36.5831	35.58137	167.0708	7.62449
2008	50.7164	74.78291	43.40628	186.7976	9.885001
2007	44.57949	55.87368	25.22774	169.3898	6.586792
2006	19.37809	37.59476	20.02912	133.2725	0.623292
2005	15.45998	15.50806	7.269407	111.2102	6.522098
2004	18.16569	15.44437	5.771297	62.10804	0.701422
2003	10.14352	7.95812	4.322748	49.45685	0.783136
2002	16.5902	3.461132	5.62604	49.30798	1.479805
2001	22.45735	2.748286	5.471947	44.241	7.270345
2000	32.77924	2.71423	3.584217	38.3993	0.968831
1999	28.576	3.30943	2.168591	38.753	1.503332
1998	31.913	2.76126	2.634652	43.751	0.550339
1997	19.65	4.864643	3.57733	44.237	3.810544
1996	11.2	2.579	2.426057	40.18	0.816389
1995	4.859	2.065	2.143628	35.8492	1.248425
1994	3.072	-	0.973271	33.787	0.37441
1993	1.292	-	0.55037	27.515	0.011291
1992	2.061	-	0.276512	11.156	0.003358
1991	1.103	-	0.073538	4.366	0.254134
1990	0.989	-	0.23669	3.487	–0.07572
1989	1.131	-	0.2521	3.393	–0.20121
1988	2.804	-	0.09125	3.194	0.158437
1987	1.169	-	0.21232	2.314	–0.19167
1986	0.345	-	0.11773	1.875	–0.05049
1985	1.441	-	0.10609	1.659	–0.45264
1984	1.594	-	0.01924	1.258	0.4195
1983	1.609	-	0.00564	0.636	0.0709

Source: Authors’ own.

Note: Authors’ own compilation based on data extracted from World Development Indicators 2014, World Bank.

Author’s Biography

Priya Gupta, is working as Faculty of Finance at Apeejay School of Management, New Delhi, India. She is also pursuing her PhD from Delhi School of Management, Delhi Technological University (formerly Delhi College of Engineering). She has an experience of over 8 years in academics. Her area of research is foreign direct investment in emerging economies. She is an MCom (Delhi University), MBA (Gold Medalist), NET (Commerce), and CA (Inter.) qualified. She has several research papers published in both national and international journals of repute in addition to presenting papers at international and national conferences at IIMs, IIT (Delhi), Jamia Millia Islamia University, Delhi University, to name a few.

Archana Singh, is working as Faculty of Finance at Delhi School of Management, Delhi Technological University (formerly Delhi College of Engineering). She has 3 years of industry experience and more than 15 years of experience in academics and research in the specific areas of financial accounting, cost accounting and security analysis, and portfolio management. She has done her PhD in management from University School of Management Studies, Guru Gobind Singh Indraprastha University, Delhi. Her broad area of research is valuation. She has done MBA in finance from L. N. Mishra Institute of Economic Development and Social Changes (an autonomous body under Magadh University), Patna. During her academic career, she has to her credit the publication and presentation of various research papers at various national and international journals and conferences.

References

Asiedu

(2002). On the determinants of Foreign Direct Investment to Developing Countries: Is Africa different? World Development , 30(1), 107–119.

Asiedu

(2006). Foreign direct investment in Africa: The role of natural resources, market size, government policy, institutions and political instability. The World Economy , 29(1), 63–77.

Azam

Lukman

(2010). Determinants of foreign direct investment in India, Indonesia and Pakistan: A quantitative approach. Journal of Managerial Sciences , 4(1), 31–44.

Balestra

(1992). Introduction to Linear Models for Panel Data. In Mátyás, L. and Sevestre, P. (Eds), The Econometrics of Panel Data (21–29). Netherlands: Springer.

Bergsman

Broadman

H. G.

Drebentsov

(1999). Improving Russia’s Foreign Direct Investment Policy Regime. In Broadman

H. G.

(Ed.), Russian Trade Policy Reform for WTO Accession . Washington, DC: The World Bank.

Breusch

T. S.

Pagan

A. R.

(1980). The Lagrange Multiplier Test and Its Applications to Model Specification in Econometrics. The Review of Economic Studies , 47(1), 239–253.

Bruegel REER (Real Effective Exchange Rate) Annual Database. (n.d.). Retrieved May 30, 2015, from www.bruegel.org/datasets/real-effective-exchange-rates-for-178-countries-a-new-database/

Chen

C. H.

(1996). Regional determinants of foreign direct investment in mainland China. Journal of Economic Studies , 23(2), 18–30.

DeAngelo

Eunni

R. V.

Fouto

N. M. M. D.

(2010). Determinants of FDI in Emerging Markets: Evidence from Brazil. International Journal of Commerce and Management , 20(3), 203–216.

10.

Dhingra

Sidhu

H. S.

(2011). Determinants of foreign direct investment inflows to India. European Journal of Social Sciences , 25(1), 21–31.

11.

Gilmore

O’ Donnell

Carson

Cummins

(2003). Factors influencing Foreign Direct Investment and International Joint Ventures: A comparative study of Northern Ireland and Bahrain. International Marketing Review , 20(2), 195–215.

12.

Goldman Sachs Annual Report (2013). Retrieved June 26, 2015, from http://www.goldmansachs.com/s/2013annualreport/assets/downloads/GS_AR13_AllPages.pdf

13.

Gupta

Singh

(2014). Macro economic determinants of foreign direct investment in BRIC nations. South Asian Journal of Marketing & Management Research , 4(2), 1–12.

14.

Hausman

J. A.

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

15.

Hirsch

(1976). An international trade and investment theory of the firm. Oxford Economic Papers , 28(2), 258–270.

16.

Hoechle

(2007). Robust standard errors for panel regressions with cross-sectional dependence. Stata Journal , 7(3), 281–312.

17.

Hong

(2008). WTO accession and foreign direct investment in China. Journal of Chinese Economic and Foreign Trade Studies , 1(2), 136–147.

18.

International Monetary Fund (2014). World Economic Outlook 2014 Database. Retrieved May 30, 2015, from https://www.imf.org/external/pubs/ft/weo/2015/01/weodata/index.aspx.

19.

Jadhav

(2012). Determinants of foreign direct investment in BRICS economies: Analysis of economic, institutional and political factor. Procedia-Social and Behavioral Sciences , 37, 5–14.

20.

Judge

G. G.

Hill

R. C.

Griffiths

W. E.

Lütkepohl

Lee

T. C.

(1982). Introduction to the theory and practice of econometrics. New York: John Wiley & Sons.

21.

Kinoshita

Campos

N. F.

(2003). Why does FDI go where it goes? New evidence from the transition economies (William Davidson Institute Working Paper No. 573). Retrieved June 26, 2015, from https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.414540

22.

Krkoska

(2001). Foreign Direct Investment Financing of Capital Formation in Central and Eastern Europe (Working Paper 67). London: European Bank for Reconstruction and Development.

23.

Maddala

(1999). On the use of panel data methods with cross-country data. Annales d’Economie et de Statistique , 55/56(July–December), 429–448.

24.

Maniam

Chatterjee

(1998). The determinants of US foreign investment in India: Implications and policy issues. Managerial Finance , 24(7), 53–62.

25.

Mottaleb

K. A.

Kalirajan

(2010). Determinants of foreign direct investment in developing countries: A comparative analysis. Margin: The Journal of Applied Economic Research , 4(4), 369–404.

26.

Lightfoot

W. S.

(2006). Determinants of foreign direct investment at the regional level in China. Journal of Technology Management in China , 1(3), 262–278.

27.

Nguyen

A. N.

Nguyen

(2007). Foreign direct investment in Vietnam: An overview and analysis; determinants of spatial distribution across provinces. Development and Policies Research Center, MPRA Paper No. 1921. Retrieved June 26, 2015, from https://mpra.ub.uni-muenchen.de/1921/1/MPRA_paper_1921.pdf

28.

Nonnemberg

M. B.

Mendonca

M. J. C.

(2004). The determinants of foreign direct investment in developing countries. SSRN . Retrieved June 26, 2015, from https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.525462

29.

O’ Neill

(2001). Building better global economic BRICS. Goldman Sachs. Retrieved June 26, 2015, from http://www.goldmansachs.com/our-thinking/archive/archive-pdfs/brics-book/brics-full-book.pdf

30.

Owen

H. C. Ho

. (2004). Determinants of Foreign Direct Investment in China: A Sectoral Analysis. Economic Discussion Paper No. 18. The University of Western Australia, Business School, Australia. Retrieved June 26, 2015, from http://www.business.uwa.edu.au/school/economics/?a=102556

31.

Oxford Economics Annual Database. (n.d.). Retrieved June 1, 2015, from https://www.oxfordeconomics.com/forecasts-and-models/countries/data-and-forecasts/global-economic-databank/overview

32.

Posner

M. V.

(1961). International trade and technical change. Oxford Economic Papers , 13(3), 323–341.

33.

Ranjan

Agrawal

(2011). FDI inflow determinants in BRIC countries: A panel data analysis. International Business Research , 4(4), 255–263.

34.

Sawkut

Seetanah

Ramessur-Seenarain

Sannassee

(2009). Determinants of FDI: Lessons from the African economies. Journal of Applied Business and Economics , 9(1), 1–16.

35.

Seetanah

Rojid

(2011). The determinants of FDI in Mauritius: A dynamic time series investigation. African Journal of Economic and Management Studies , 2(1), 24–41.

36.

Severiano

A. A. F.

(2011). The determinants of FDI in Portugal: A sectoral approach (Dissertation). Lisbon, Portugal: Universidade Católica Portuguesa. Retrieved July 27, 2015, from http://www.clsbe.lisboa.ucp.pt/resources/Documents/MSC/Teses%20dos%20Alunos/Andreia%20Severiano.pdf

37.

Singhania

Gupta

(2011). Determinants of foreign direct investment in India. Journal of International Trade Law and Policy , 10(1), 64–82.

38.

Vijayakumar

Sridharan

Rao

K. C. S.

(2010). Determinants of FDI in BRICS countries: A panel analysis. International Journal of Business Science and Applied Management , 5(3), 1–13.

39.

Webb

Hall

(2009). Application of a dynamic panel data estimator to cross-country coffee demand: A tale of two eras. Journal of Economic Development , 34(1), 1–17.

40.

Wooldridge

J. M.

(2010). Econometric analysis of cross section and panel data (pp. 257–265). Cambridge, Massachusetts; London, England: MIT Press. Retrieved June 26, 2015, from http://www.ipc-undp.org/evaluation/apoio/Wooldridge%20%20Crosssection%20and%20Panel%20Data.pdf

41.

World Bank. (2014). World Development Indicators 2014 Database . Retrieved May 30, 2015, from data.worldbank.org/data-catalog/world-development-indicators