Abstract
Developed financial system is one of the important determinants of Foreign Direct Investment (FDI). The article analyzes the impact of financial system development on FDI with respect to BRIC countries for the period 1991 to 2010. Using the panel data analyses, fixed and random effect, our results conclude that FDI inflows to BRIC countries are influenced by banking sector and stock market variables, used as a proxy for financial development. FDI is positively influenced by size of banking sector and stock market capitalization. However, more domestic credit by banking sector negatively influences FDI inflows to these countries over the period of study. The study contributes to the existing literature as it analyzes the influence of financial sector development on China’s FDI too, not included in earlier cross-countries studies.
Introduction
The process of globalization has made the world a single global village. This process is now irreversible, caused primarily by trade and investment across economies resulting in strong worldwide market for goods, services and capital in particular. Foreign Direct Investment is one of the important outcomes of globalization. The global stock of Foreign Direct Investment (FDI) increased from US$ 2.081 trillion in 1991 to USD 17.743 trillion in 2009 (UNCTAD, 2010). FDI is increasingly considered as an important source of economic growth and development for emerging economies. The benefit of FDI can be in the form of knowledge & technology spillovers, employment generation and enterprise development (NACER, 2009; OECD, 2002; Magnus et al., 2008). Countries all over the world, especially developing countries, strive to attract FDI to their countries. This is evident from the increasing deregulation and liberalization of international investment regimes by the respective countries all over. However, the pace of reforms varies from one country to another.
Competitiveness of a country to attract FDI depends on various factors such as policy and economic variables, its infrastructure and its business facilitation measures. Apart from these factors, the financial system of the country also plays a significant role in attracting FDI and contributing to the economic growth of the host country. Financial intermediaries like banks and stock markets help in mobilization of savings and investment, creating new business opportunities, facilitating trading in securities, optimum allocation of capital to projects giving high yields and so on. In other words, a well-developed financial system acts as a catalyst for materializing the benefits from FDI and to increase its inflow into the host country.
The financial system of a country includes its banking sector, other financial institutions and the stock market. An efficient banking sector can contribute to FDI inflows by providing financial support in terms of lower costs and quicker transactions, loans availability and better foreign exchange services. The efficient stock market provides for better linkages between domestic and foreign investors. In this regard, it is imperative to understand the link between the FDI and efficient financial system and to identify those variables that contribute to FDI inflows.
Statistics reveal that FDI is showing an increasing trend in developing economies. The percentage of FDI stock in developing countries is 28 per cent of the world FDI stock in 2009.
The stock of FDI to developing countries increased from USD 0.524 trillion to USD 4.894 trillion from 1990 to 2009 as shown in the Figure 1.
There exists voluminous literature on FDI and its determinants. Bevan, Alen A et al. (2000) reported that country risk, unit labour costs and host market size are the major determinants in Central and Eastern Europe. OECD (2002) maintains that national policies and international investment architecture are significant factors for attracting FDI to a larger number of developing countries. In this backdrop, there is a need to have transparent and effective policy environment with sound institutional capabilities to implement them. Major studies are focused on economic and policy variables of the host country as a determinant of FDI.
The BRICs (Brazil, Russia, India and China) are the four biggest emerging economies, as they account for two fifths of the total Gross Domestic Product (GDP) of all emerging economies. Wilson and Purushothaman (2003)argue that BRIC economies could become a much larger force in the world economy than the G6 countries (United States, Japan, Germany, France, Italy, and the United Kingdom) in less than 40 years and by 2025, they could account for more than half the size of the G6. BRIC countries are among the best international investment destinations in the world. This can be evident from the fact that FDI stock of BRICs increased from 25 billion Euros in 2004 to 115 billion Euros in 2009 (Eurostat, 2010).
Apart from this, they are fast-growing economies with the biggest labour force. In terms of population and land area, they account for approximately 40 per cent of world’s population and one third of land area of the world as reflected in Table 1.
In terms of GDP volume and growth, emerging economies overtook the developed economies by their share in World GDP in terms of Purchasing Power Parity (PPP) for the year 2005. The main contributors to this GDP growth are BRIC countries. In 2009, BRIC countries were among the top-10 economies out of the total of 178 countries, with China on second position, India on fourth, Russia on sixth and Brazil on ninth position in terms of GDP volume (World Development Indicators Database, World Bank 2010).
Interestingly, in terms of size of the economy measured by gross national income, these countries account for 17.30 per cent of the world’s total gross national income. Therefore, the focus of our study is BRIC countries. Our results conclude whether the development of financial system influence FDI in these countries and suggest some policy implications.
Before we proceed to the analysis, it is important to know what FDI is. FDI is a category of international investment in which a resident entity in one economy (the foreign direct investor) acquires a lasting interest in an enterprise resident in another economy (the direct investment enterprise). However, the foreign investor should own at least 10 per cent of on overseas direct investment enterprise to control and influence the management of the investment (IMF, 1996). FDI consists of both the initial transaction that creates investments as well as subsequent transactions between the direct investor and the direct investment enterprises in the host country aimed at maintaining, expanding or reducing investments.
Population and Land Area of BRIC and the World in 2009
The data on FDI includes three broad components, new equity flows (either through M&A or Greenfield investment), intra-company loans and reinvested earnings.
The study fills the gap in the literature by highlighting the role of financial system in attracting FDI to four-biggest emerging economies. The relationship is analyzed by applying panel-data analysis using fixed effects and random effects in Ordinary Least Square (OLS) framework. Apart from this, panel corrected error terms regression is also used to incorporate the effect of autocorrelation in the error terms within the panel.
The article is divided into four sections. Section 1 is the introduction. Section 2 contains review of literature. Section 3 explains the data and methodology used for analysis. Empirical results are shown in Section 4, followed by the concluding remarks.
Brief Literature Review
There are three important set of determinants that influence the FDI inflows, namely, presence of ownership-specific competitive (O) advantages in a transnational corporation, the presence of location advantages (L) in a host country and the presence of superior commercial benefits internally in a firm (I) (Dunning, 1993). Singh and Jun (1996) argue that political risk, business conditions and macroeconomic policies matter for FDI in developing countries.
A study by Sun Qian et al. (2002) of 30 provinces of China provides that FDI determinants move through time. Labour quality and infrastructure are important determinants of the distribution of FDI. High labour quality and good infrastructure attract foreign investors. Also, China’s political stability and openness to the foreign world is another important factor for attracting foreign capital. Globerman et al. (2002) analyzed that for developing and developed countries, governance infrastructure in the form of institutions and policies is an important determinant of FDI inflows and outflows. Moosa and Cardak (2006) in their study of 138 countries concluded that countries with high degree of openness and low country risk attract more FDI. Sahoo and Pravakar (2006) further provide that major determinants of FDI in South Asia are market size, labour force growth, infrastructure index and trade openness. Sung-Hoon Lim (2008) in his study of China concludes that Investment promotion positively affects the attraction of FDI.
Despite the obvious effect of FDI on economic growth, a good financial system also plays a significant role. Alfaro et al. (2004) argue that spill overs of FDI for the host economy might crucially depend on the extent of the development of domestic financial markets. The studies by King and Levine (1993) and Levine and Zervos (1998) concluded that a good financial system can promote economic growth.
Agarwal and Mohtadi (2004) study the role of financial market development in the financing choice of firms in 21 developing countries over the period 1980–1997. Their analysis reveals that FDI as proportion of GDP and investment as a proportion of GDP are positively correlated with both the stock market variables and the banking variables.
Nasser and Gomez (2009), using data of 15 Latin American countries, analyzed that FDI is directed to the counties that are financially developed and institutionally strong.
Financial Sector Development in BRICs-A Synopsis
Brazil
The financial reforms started in 1994 with a continuous phase of improvement of institutions, procedures and security of the system. The sector opened up to foreign and domestic private participation. The normative function of regulation and supervision is conducted by Banco do Brazil, National Monetary Council and Securities and Exchange Commission. The intermediation function is performed by commercial banks and saving banks. The financial market of Brazil is modern with state-of-the-art payment system. The recent credit market reforms in Brazil have paved the way for financial inclusion. The Central Bank of Brazil is implementing Basel II to improve the banking sectors regulation and supervision. The entry of foreign banks in Brazil has fostered competition while improving the profitability and efficiency of other banks. The Brazilian banking system consists of state-owned, foreign and private domestic banks.
Russia
Banking reforms in Russia actually started after the crises of 1998 when there was a complete rebound of banking system through restructuring of several large banks and technological advancement in the banking system. Besides the Sberbank, new five clusters of commercial banks were formed. The major weaknesses of Russian banking sector include lack of capabilities to lend large industrial companies, relatively high level of foreign investment with low protection of foreign investors (Hummel and Plakitkina, 2004). There is a need to have financial deepening in Russia to diversify in sectors other than oil sector. Apart from corporate banking, retail banking is also gaining ground, thereby fuelling demand for other banking products. The formal financial market of Russia is still untapped with approximately 42 per cent of the population not having access to financial services (BRICS Report, 2012). Though, the stock markets have grown rapidly due strong commodity prices and a domestic boom in recent years, a large proportion of Russian stocks are still traded outside Russia via depository receipt programmes (Moser, 2007).
India
The banking sector in India has undergone a revolutionary change since 1991 when Narasimham Commission-I in 1991 had provided the blue print for the first-generation reforms in the financial sector. Prior to 1990, the financial sector was highly regulated. The unprecedented financial reforms were aimed at development of stable and efficient financial sector with increased operational efficiency, diversification of products lines and services, privatization of banks and enhanced transparency and disclosure norms to ensure market discipline. As a result, the number of offices of scheduled commercial banks increased from 61,248 in 1998 to 93,080 in 2011. Also, per capita deposit of scheduled commercial banks increased from Rs 6,170 in 1998 to Rs 46,321 in 2011 (RBI, 2006; RBI, 2011). Financial inclusion is given top most priority by government now to increase penetration of affordable banking services in rural and unorganized sector (RBI, 2010). The major turning point in capital market reforms was the constitution of Securities and Exchange Board of India (SEBI) and abolition of capital issues control act resulting in tightened entry norms, improved disclosure norms, dematerialization of securities, on-line trading, reduced settlement period, reduced information asymmetries, minimizing transaction costs, and so on.
China
The financial sector development in China is impressive since the inception of financial reforms in 1978. The main thrust of financial sector development is devoted to institutional development, creating markets and liberalization of financial operations. The capital market focused on development of financial instruments. The financial liberalization in China included commercialization of state-owned banks, establishment of effective banking supervision, and reform of the accounting system used by the banks. China’s stock market expanded rapidly since the constitution of Shanghai and Shenzhen Stock exchange in early 1990s. Over the years of steady reforms, the financial system of China has become robust due to improvement in asset quality, capital adequacy and management.
Data and Methodology
This section describes the data used for empirical analysis. The data consists of yearly observations from 1991 to 2010 for the four emerging economies, namely, Brazil, Russia, India and China. The dependent variable is a log of FDI inflows in the respective country in USD taken from World Development Indicators published by World Bank (2011). It is denoted by LFDI. The variables used for measuring the financial development are described in Table 2.
List of Variables Used in the Analysis
Most of these variables are used by King and Levine (1993), Hearn et al. (2010), Nasser and Gomez (2009), Alfaro et al. (2003) and Abdel Aal Mahmoud (2010) to establish the relation between FDI and financial development variables. This study complements the existing literature by providing empirical analysis on BRIC countries only, as the existing studies contain large number of countries but do not have China in their list, which is the one of major recipients of FDI in the world. The data series are taken from 1990 onwards because of the non-availability of data prior to that period for Russia and a few variables for China. Thus, to ensure homogeneity in time period, the data from 1991 is taken for all the countries under study.
Research Framework
We have used the panel data analysis including pooled ordinary least square (POLS), fixed effect and random effect. Hsiao (1995) and Baltagi (1995) argued that panel data sets possess several major advantages as they suggest individual heterogeneity to reduce the risk of obtaining biased results, provide a large number of data points (observations) to increase the degrees of freedom and variability and to be able to study the dynamics of adjustment. Before the analysis is done, the data is checked for the presence of unit roots. In other words, stationary properties of panel data is examined using Im-Pesaran-Shin panel unit root test. In POLS, the results show the impact on all the countries included in the panel ignoring the heterogeneity across countries.
The equation for POLS is given by:
Where LFDI is the log of FDI inflows to BRIC countries which is a dependent variable,
α is the pooled intercept,
LLLY is liquid liabilities of banking sector,
LDOMCRE is the log of domestic credit by banking sector to GDP,
LBROE is log value of average return on equity of banks,
LBCYR is the log of bank cost to income ratio,
LMCAP is log value of stock market capitalization,
LSTURN is log of turnover ratio,
βi are the slopes of independent variables (IVs) and ε is the error term.
The core difference between fixed and random effect models lies in the role of dummy variables. If dummies are considered as a part of the intercept, this is a fixed-effect model. In a random effect model, the dummies act as an error term. It can be argued that financial development and FDI inflows accrue to these countries at different points of time. In order to control for unobserved heterogeneity across counties, we have used countries as dummy variable absorbing effect particular to each country and analyzing the pure impact of financial system development and efficiency variables. As a result, there is different intercept for each country but the slopes will be the same for all countries. Owing to this intra-panel variation, the random effects model has the distinct advantage of allowing for time-invariant variables to be included among the regressors.
We have used the Least Square Dummy Variable (LSDV) model to identify the fixed group effect on each country where model will include a dummy for each country. The model will drop one dummy and use it as reference group. Equivalent results can be obtained if a different country is taken as a reference group (Park, 2009). The effect of dummy variable is that there will be n-1 countries included in the model. After introducing the country as dummy variable, the eq. (1) will become:
FDIit is the Dependent variable (DV) where i = country and t = time.
Cn is the binary dummy for country. γ is the coefficient for the binary country dummy.
We control for time variable whenever unexpected variation or special events affect the dependent variable. So, the equation for fixed effects using time fixed effect will become:
Where δ is the coefficient of binary time regressor and Tt is the binary time variable. And there will be t-1 time period in the model.
Fixed effects model assumes that time invariant characters are perfectly collinear with country so it cannot cause any change in the DV as it is constant for each country. If we consider the effect of these time invariant factors, random effect model is used. In other words, when variation across countries is assumed to be random and uncorrelated, random effect will give more valid results. The equation for random effect model is:
Where μit is the between country error and εitis the within country error.
To check for heteroskedasticity, which is major problem in panel-data analysis, we employed Breusch-Pagan/Cook-Weisberg test for heteroskedasticity. High chi-square value indicates the presence of heteroskedasticity. In order to control for heteroskedasticity, we have used robust standard errors so that our error terms are free from any bias. Further, the use of robust standard errors does not change coefficients estimates and gives correct accurate p values even if the errors are not independently and identically distributed.
Also, in order to examine the issue of multi-collinearity in the IVs, we have computed variance inflation factor (VIF) values. Presence of multi-collinearity can increase standard errors and the results can be spurious. The VIF values measure the relationship of all the variables simultaneously rather than two variables shown by correlation matrix. It is calculating when each independent variable (IV) is regressed on the other IVs and if the other IVs can predict the particular IV, then it will have high VIF value. VIFs are generally considered bad if they exceed 5.
Empirical Results
The aim of the paper is to identify the financial sector variables that determine the FDI inflows to BRIC countries. The analysis reveals that BRIC have almost similar patterns of FDI inflows. Figure 2 shows that the FDI inflow to India in the beginning of 1990s was lowest among the BRIC countries. However the performance of India is consistent with Russia and there is a substantial increase after 1995 onwards. Interestingly, China continues to lead with highest FDI inflows throughout the period of study. In 2009, the inflows of FDI reduced to these countries, which can be due to global downturn.
FDI inflows to Brazil showed substantial increase after 1993 and reached its peak level in 2000 but there was a decline after 2000 onwards. The standard deviation of LFDI and LMCAP is high as compared to other variables as depicted by Table 3.
The results of I-P-S panel unit root test presented in Table 4 reflects that the variables are not stationary at level. The major advantage of I-P-S unit root test over Levin-Lin-Chiu test for unit root in panel data is that it can be used for unbalance panel data. In order to make the series stationary the first difference values are checked for stationarity. It can be inferred from Table 4 that all variables are stationary at their first difference.
Summary Statistics of the Variables for BRIC Countries
Thus, the first difference values are taken for further analysis of data. In other words, LFDI will now be ΔLFDI; LDOMCRE will be ΔLDOMCRE and so on, where Δ indicates first difference values.
Results of Im-Pesaran-Shin Panel Unit Root Test
The results of Breusch-Pagan/Cook-Weisberg test indicate the presence of heteroskedasticity:
Ho: Constant variance of error terms, i.e., homoskedasticity H1: Heteroskedasticity Variables: fitted values of ΔLFDI chi2(1) = 113.17 Prob > chi2 = 0.000
The high value of Chi indicates the hetroskedasticity. Consequently, we have used robust standard errors in order to control for hetroskedasticity in the results.
Table 5 indicates that there is no multi-collinearity in the variables used in the analysis as all the VIF values are lesser than 5.
Result of Multi-collinearity among the Variables Used in Analysis
The panel data analysis is done on the first difference values of the variables using POLS, FE and RE. POLS model gives a common constant for all the countries included in the panel. While FE takes into consideration the country specific effect in the intercept values.
The results for POLS, fixed effect and random effect are presented in Table 6.
The results depicted in Table 6 show that POLS and RE is giving similar values of coefficients but their significance is different. In POLS the significant variable are LLY, LDOMCRE and LMCAP with R-square value as 62 per cent. However, the significant variable in RE model is LLLY only with R-square value 62 per cent.
As per FE model, the significant variable affecting FDI inflows to BRIC countries include LLLY and LDOMCRE from banking sector development and LMCAP from capital market development. In other words, the size of banking sector measured by liquid liabilities, domestic credit by banking sector and market capitalization influence FDI inflows to BRIC countries, as their coefficients are statistically significant. A unit increase in LLY and LMCAP, positively effects FDI inflows and increases FDI by 2.49 and 0.14, respectively. However, the negative coefficient value of LDOMCRE indicates that it affects the FDI inflows negatively to BRIC countries during the period of study. The constant value in FE represents the intercept of Brazil, and dummy-Russia, dummy-India and dummy-Brazil are the differential intercept coefficients that represents by how much the intercept of Russia, India and China differ from intercept of Brazil.
Gujarati (2003) asserts that there is little difference in the values of parameters estimated by FE and RE if the time series data is large and cross-sectional entities are small. As a result, FE estimators are recommended for conclusion. Also, the results of BP-LM test for random effects also corroborate the FE.
The time effect is tested following equation (3) and the results of F test is not significant, which indicates that there is no need for analyzing time effect, as there is no significant difference in values or unexpected variations in the values of countries under study over time. (Results are available on request from the corresponding author.).
Results of POLS, Fixed Effect and Random Effect
Results of Prais-Winston Regression with Panel Corrected Standard Errors
In order to control for autocorrelation in the residuals inherent within the panels from one time period to another, the results of a robust panel regression model given by Prais-Winston are presented in Table 7. The analysis is based on panel corrected standard errors and it gives a different estimation of parameters and standard errors. It can be inferred from the Table that the significant variables are LMCAP, LDOMCRE, LLLY and LBROE.
LBROE becomes significant when the panels are assumed to be auto-correlated of degree one. In other words, when panel specific AR (1) is taken into consideration, the value of R-square increases to 0.6334 along with LBROE as statistically significant variable affecting FDI to BRIC countries.
Concluding Remarks
As stated earlier, the purpose of this study is to analyze the impact of financial system development on the inflows of FDI to BRIC countries. It can be construed from the analysis that the country’s financial system development plays a very vital role in the accelerating the FDI inflows. This study on BRIC economies reveals that for all the countries taken together, one capital market variable, namely, MCAP and two banking sector variables, namely, LLY and LDOMCRE, influence FDI. When FDI comes through M&A, the capital market development is more relevant from the prospective of foreign investor. Furthermore, well-functioning stock markets play a significant role in removing entry and exit barriers for foreign investors and facilitating better linkages between domestic and foreign markets.
The size of the banking sector measured by LLY is a statistically significant factor determining FDI inflows to BRIC countries. Thus, the policymakers of these countries should take necessary steps to facilitate the banking sector, as it can influence FDI inflows positively. This is essential because banks provide the services to the foreign investors regarding the investment and repatriation issues. Apart from development, the efficacy of the banking sector measured by LBROE also influences FDI as inferred from Table 7. It is evident from the analysis that foreign investors not only consider the development of financial sector but also take cognizance of the efficiency part of financial sector. Thus, both development and efficiency of the banking sector should be the prime concern for policymakers of these countries. However, it is surprising that domestic credit provided by the banking sector has a negative influence on FDI inflows in BRIC countries for the period under study. It implies that more credit by banking sector was not able to attract FDI to these economies over the study period. This can be due to the fact that more domestic credit by banking sectors makes FDI unattractive for domestic investors. As a result, domestic companies raise funds from domestic sources rather than taking the route of ADRs/GDRs and other forms of foreign capital.
It can be concluded from the study that the efficient financial system consisting of banking sector and capital market have a strong influence on FDI inflows to BRIC countries. The policy makers must strive to improve the efficiency of these financial intermediaries by following international standards in banking and integrating their domestic stock markets with best international stock markets of the world.