Effect of institutional quality on FDI inflows in South Asian and Southeast Asian countries

Abstract

Foreign Direct Investment (FDI) plays a pivotal role in the economic development of countries and institutional quality encompassing aspects such as political stability, regulatory environment, and the rule of law, plays a vital role in attracting and retaining foreign investment. This study investigates the impact of institutional quality on the level of FDI inflows in the South Asian and Southeast Asian countries over the period 2002–2019. We have constructed an institutional quality index by using Principal Component Analysis (PCA) on six governance indicators. Iterated generalized least squares (I-GLS) in the fixed effect model has been employed for the estimation of the results. The results show that the institutional quality index has a positive and significant impact on the FDI inflows in both the regions. This implies that an ideal governance system comprising of low corruption, political stability, absence of violence, voice and accountability, regulatory quality and proper judicial system helps to attract FDI inflows in the South Asian and Southeast Asian countries. It also has positive spillovers to other economic activities such as GDP growth, international trade and financial development that are vital for economic growth and development.

1. Introduction

Foreign Direct Investment (FDI) plays a pivotal role in fostering sustainable economic growth and employment in host countries. As global economic integration deepens, FDI stands out as the engine propelling the recipient nation's progress. This influx of capital not only provides much-needed financial resources to emerging countries but also facilitates the transfer of managerial expertise and technological advancements. Consequently, there is heightened competition among both developing and developed nations to attract FDI, leading governments to offer various incentives to investors. These incentives typically take the form of fiscal measures, including relaxed repatriation laws and reduced tax levels for foreign investors. Additionally, the provision of essential infrastructure emerges as a critical factor in attracting foreign investments, prompting governments to allocate significant expenditures to bolster their economies.

In recent years, political economy variables and good governance have gained recognition as crucial determinants of FDI inflows [1]. Factors such as inconsistent government policies, low institutional quality, and high business transaction costs act as deterrents to potential investments [2,3]. The quality of institutions and governance is viewed as instrumental in reducing transaction costs associated with operations, logistics, production, research and development, and risk monitoring [4]. Poor institutional quality, characterized by unregulated institutions, high corruption levels, and inadequately protected property rights, poses a threat to investments by limiting enterprises' operational capabilities, risk diversification, and dispute resolution. Investors, therefore, must evaluate transaction costs to gauge the overall business environment, with the host country's efforts to reduce these costs instilling trust and making it an attractive destination for FDI [5].

Moreover, the political environment, as part of institutional quality, plays a key role in determining FDI inflows. A stable political environment, coupled with fair elections, can enhance fiscal policies and stabilize diplomatic relationships, making the country more appealing to foreign investors. Conversely, a poor institutional quality marked by corruption can increase the costs of doing business, dissuading foreign investors [6]. In addition to that [7], pointed out that countries with weaker institutions perform poorly in comparison to the countries having better institutional quality. This underscores the importance of analyzing political and institutional factors influencing FDI inflows, as these determinants vary based on political, socio-economic, and geographical dynamics.

Thus, it is imperative to analyse the political and institutional factors influencing the FDI inflows. The determinants of FDI vary from country to country depending on the political, socio-economic and geographical dynamics. Macroeconomic factors such as, financial development, inflation, per capita GDP, international trade and exchange rate also play a significant role in determining FDI inflows in an economy [[8], [9], [10]]. Though the determinants of FDI have been extensively researched, the studies on the impact of institutional quality on FDI inflows has been limited due to lack of data on the quantification of institutional quality. In this study, institutional quality, gross domestic product per capita, exchange rate, inflation, financial development, capital formation and international trade are taken as the determining factors of FDI.

The objective of this study is to assess the impact of institutional quality on the level of FDI inflows in the South Asian and Southeast Asian countries. The rationale behind choosing these regions for the study is that they constitute a significant portion of the world economy as fast-growing economies. These regions constitute both developing and developed countries with a large market size. They represent rich geo-political, ethnic and linguistic diversity, and varied religious and socio-cultural practices. Along with that, these regions share common characteristics such as, geographical location, abundance of natural resources, high regional integration, population growth and economic growth.

The performance of an economy is determined by its macroeconomic performance comprising of GDP growth rate, exchange rate, rate of inflation etc. It reflects the position of a country and its ability to execute policies. Table 1 shows the current scenario of the macroeconomic status in South Asia and Southeast Asia. The South Asian region comprises of eight emerging economies, namely, Afghanistan, Bangladesh, Bhutan, India, Maldives Nepal, Pakistan and Sri Lanka. With a population of more than 1.81 billion, they comprise 3.8% of the total global economy [11]. This region is one of the fastest growing economic regions with an annual growth rate of 6.67 %. The Southeast Asian region comprises of ten countries, namely, Brunei Darussalam, Cambodia, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand, and Vietnam. This region also registers a moderate growth rate of 5.25%. The growth was driven by strong investment mainly in Singapore, Indonesia and Vietnam. These three countries received more than 80% of FDI inflows in Southeast Asia in 2019 [12]. Southeast Asia received a major share of FDI inflows at 6.01%, in comparison to 1.44% in South Asia. Both the regions comprising of emerging countries registered an impressive economic growth.

Table 1.

Stylised facts: Macroeconomic scenario of South Asia and Southeast Asia (2019).

	South Asia	Southeast Asia
GDP (current US$) (trillion)	3.44	2.93
GDP annual growth (%)	6.67	5.25
Gross savings (% of GDP)	29.9	32.9
FDI inflows (% of GDP)	1.44	6.01
Trade (% of GDP)	42.11	108.17
Inflation	3.67	2.56
Population (billion)	1.81	0.65

Source: United Nations Conference on Trade and Development (UNCTAD), 2019 [13].

The findings of our study would focus on drawing attention of the governments of emerging economies in formulating their policies to strengthen their stand on the global economy. The focus would be on the betterment of the institutional quality of the governments so as to attract more investment and to achieve higher economic growth. The rest of the paper is organised as follows: Section 2 reviews the existing theoretical literature and examines the available empirical findings from other countries and regions on the determinants of FDI and its links with other factors; Section 3 provides a theoretical framework of the link between institutional quality and FDI inflows. Section 4 presents the data and methodology where information regarding the data source and model specification is provided. Section 5 presents the empirical results and analysis. Finally, section 6 presents the conclusion with some policy implications of the study.

2. Review of literature

The impact of institutional quality on FDI inflows is well addressed in the theoretical as well as empirical literature. However, the results are quite inconclusive. Previous studies have found evidence of mixed effects of the impact of institutional quality on FDI inflows. While some studies have found a positive association between institutional quality and FDI inflows, on the other hand many studies have reported a negative association. Furthermore, many studies have also reported a weak or insignificant association. Also, since the indices associated with institutional quality can be varied, we need to examine the different parameters associated with institutional quality and their effects on FDI inflows.

Researchers have concluded that countries having improved governance infrastructure, strong judicial system and being highly ranked with various indices of institutional quality, tend to receive significantly more FDI inflows. Some of these studies are discussed in detail [14]. examined the impact of governance on FDI on a broad sample of 144 developing and developed countries over the years 1995-97. They concluded that governance infrastructure, constituting of legal, political, and institutional factors is a key determinant of both FDI inflows and outflows [15]. studied the effect of democracy on FDI inflows in a set of 98 developing countries over the period 1984–2004. Their study inferred that the democratic developing countries with better institutional quality attracted more FDI inflows [16]. found a bidirectional cointegrating relationship among FDI and institutional quality in the long run, in their study constituting of 62 developing countries over the period 1984–2003 [17]. studied the impact of institutional quality on FDI inflows on a sample of 52 countries over the period 1985–2000 and concluded that institutional quality is an important determinant of FDI inflows [18]. examined a sample of 83 countries over the period of 1984–2003 and concluded that a favourable institutional framework and political stability positively affect the FDI inflows. Similar results were reported by Refs. [[19], [20], [21]]. [22] study for Arab economies reported that by increasing government stability and decreasing investment risks, there can be an increase in the FDI inflows. Similarly [23], examined 52 countries over the period 2006–12 and found that a favourable public and private governance system affects FDI inflows positively.

Unlike these studies, there are also studies which have reported a negative association among institutional quality and FDI inflows. Mostly, the studies which have used corruption as a proxy of institutional quality [[24], [25], [26], [27]] have reported that an increase in corruption leads to a decrease in FDI inflows [28]. found evidence that democracy has a negative impact on FDI inflows in study for a sample of 53 countries over the period 1982–1995 [29]. found that an absence of democracy and political instability helps in increasing FDI inflows in four selected GCC countries over the period 2003–2010. Their study implied that a favourable institutional quality comprising of a stable political environment may have a negative impact on the FDI inflows.

Furthermore, there are also studies which have reported an insignificant or a weak effect of institutional quality on FDI inflows [[30], [31], [32]]. reported that political factors, as a determinant of institutional quality have no significant effect on FDI inflows [[33], [34], [35]]. fail to establish any significant relationship between institutional quality and FDI inflows.

Some of the recent reviews of literature covering the relationship among all the different parameters of institutional quality and FDI inflows are presented in Table 2.

Table 2.

Recent review of literature.

Authors	Coverage	Model	Findings
[36]	1985–2009;	System GMM	Law and order and democratic accountability have a positive impact on FDI inflows.
	17 MENA Countries
[37]	2002–2012;	Fixed Effect	Regulatory quality has a positive impact on FDI inflows.
	91 countries		Absence of violence, effective management and political stability has a negative impact on FDI inflows.
[38]	1990–2010;	Panel Regression	Law and order and investment profile have a positive impact on FDI inflows.
	40 Developing and Developed Countries		Bureaucratic quality has a negative impact on FDI inflows.
[39]	2001–2015;	GMM	Institutional efficacy or quality has a positive impact on FDI inflows.
	36 SSA countries
[40]	2002–2012;	Panel Regression	Institutional quality has a positive and significant impact on FDI inflows in developed countries.
	110 developing and developed countries		Institutional quality has no significant impact on FDI inflows in developing countries.
[41]	Pakistan	ARDL	Institutional quality has a positive impact on FDI inflows in the manufacturing and services sector in the long run.
			Institutional quality has no significant impact on FDI inflows in the primary sector in the long run.
			Institutional quality has no significant impact on FDI inflows in the short run.
[42]	2001–2011;	Panel Regression	Institutional quality has no significant impact on FDI inflows.
	BRICS and MINT
[43]	1996–2015;	Panel Regression	Institutional quality has a positive impact on FDI inflows.
	10 South American countries
[44]	2002–2017;	Panel Regression	Voice and Accountability has a positive impact of FDI inflows in Brazil.
	BRICS		Regulatory Quality has no significant impact on FDI inflows in Russia.
[7]	1996–2015;	GMM	Institutional quality enhaces the FDI-led economic growth only in low- and middle-income countries.
	104 countries
[45]	2007–2017;	Arrelano-Bond autocorrelation	Institutional quality has significant impact on FDI inflows.
	17 Central and Eastern Europe countries (CEECs)
[46]	1986–2016;	Panel ARDL	Institutional quality has a strong and positive impact on FDI inflows in South East Asia and has a weak impact on FDI inflows in South Asia.
	South Asia and South East Asia
[47]	1996–2018;	GMM	Lower corruption level has a positive impact on FDI inflows.
	54 developed and developing countries
[48]	2002–2017;	Causality test	Control of corruption, political stability, and rule of law cause FDI inflows in the Western Balkan countries.
	Western Balkan countries
[49]	2000–2017;	Meta analysis	Institutional quality has a positive impact on FDI inflows.
	BRICS
[50]	2002–2019;	Panel Regression	Rule of law, regulatory quality, political stability and absence of violence have a positive impact on FDI inflows in BRIC.
	BRIC and CIVETS		Control of corruption, government effectiveness, regulatory quality, political stability and absence of violence have a positive impact on FDI inflows in CIVETS.
[51]	2002–2019;	System and Difference GMM	Institutional quality along individual indicators, political stability, rule of law, and regulatory quality are found to be poor governance indicators in all panels, while voice and accountability and control of corruption are weak indicators in Belt and Road countries. Overall, institutional quality has a significant and positive impact on FDI inflows.
	107 developing countries and 39 Belt and Road Initiative countries

Note: ARDL: Autoregressive-Distributed Lag; GMM: Generalized Method of Moments; MENA: Middle East and North Africa; BRICS: Brazil, Russia, India, China and South Africa; MINT: Mexico, Indonesia, Nigeria and Turkey; CIVETS: Colombia, Indonesia, Vietnam, Egypt, Turkey and South Africa.

Source: Authors' Compilation

Thus, no conclusive evidence has been found from the previous literature. It has been observed that different indicators of institutional quality could lead to contrasting results. Therefore, we have constructed a single institutional quality index by using Principal Component Analysis (PCA) comprising of six different indicators of institutional quality. Our study stands apart from prior research in several distinct aspects. To begin with, we encompass a more extensive spectrum of variables that constitute social, political and economic dimensions, creating a holistic and comprehensive model. This approach is in line with the findings of [52], who assert that the majority of previous studies have primarily focussed on appraising the impact of market and regulatory institutions and overlooking the other institutional components despite the evidence of their relevance in shaping foreign investment determinations. Also, the previous studies have focussed on different groups of developing or developed countries in their analysis. Our study aims to fill this gap by considering a panel of 18 South Asian and Southeast countries which is a geographical conglomeration of both developed and countries over the period of 2002–2019.

3. Theoretical framework

According to the eclectic paradigm by Ref. [53], multinational enterprises tend to invest by virtue of gaining comparative advantage over local firms, with respect to ownership, location and internalization (OLI). Thereafter [54], highlighted the importance of institutions as the set of formal and informal rules, which are followed for maximization of returns and profits and for optimal resource allocation. Institutions can be defined as constraints comprising of social, political, economic, and structural issues. According to the North's institutional theory, good quality institutions increase profitability by reducing the cost of doing business, as they can reduce production, manufacturing, and transaction costs. Additionally, they lead to better utilization of FDI inflows. Hence, investors tend to invest more as they prefer a risk-free environment. On the other hand, it postulated that the countries having poor institutional quality tend to receive less FDI inflows. It could be attributed to poorly protected property rights and political risks in doing business, leading to slow economic activities. Thus, the theory postulated that institutions are an important determinant of FDI inflows. Furthermore, the concept of locational advantage was expanded by adding economic and institutional factors by Ref. [55]. The study proposed that investors' decision towards FDI outflow is highly influenced by locations offering profitability with good rate of returns along with macroeconomic stability and sound institutional quality. Based on these theories, we can formulate the relation as:

FDI = f (Institutions, Market Size, Macroeconomic Stability)

This implies that FDI inflows are dependent upon institutional quality, market size and stable macroeconomic conditions of the host country.

The institutional quality of a country comprises of political, regulatory and governance parameters [56]. categorised the parameters into three dimensions, which further segregated into six different indicators.

• a)
  the process of selection, monitoring and replacement of the government

  • i)
    Voice and Accountability (VA): It captures the perceptions of the citizens in their freedom of association, freedom of expression and freedom to be able to choose their government and a media independence.
  • ii)
    Political Stability and Absence of Violence/Terrorism (PV): It measures the risk of a removal or destabilization of power of the government in an unconstitutional or violent way.
• b)
  the government's capacity to formulate and implement policies effectively,

  • iii)
    Government Effectiveness (GE): It measures the credibility of the government by the quality of public services, competence of civil servants and quality of bureaucracy.
  • iv)
    Regulatory Quality: It measures the ability of the government in formulating sound policies and regulations and implementing them conducive to promote development.
• c)
  the trust and respect of the citizens on the institutions governing the socio-economic interactions.

  • v)
    Rule of Law (RL): It measures the perceptions on the predictability and effectiveness of the judiciary system such as the police and the court.
  • vi)
    Control of Corruption (CC): It measures the different controls of public power exercised to curb the different kinds of corruption.

These six indicators are used in our study as a measure of institutional quality.

Market size can be measured through the GDP growth rate of a country [50]. The macroeconomic stability is measured through a set of variables such as financial development, total capital formation [43], international trade [57], exchange rate [58,59] and inflation [60,61].

4. Data and methodology

The study is based on a balanced panel data comprising of 18 South Asian and Southeast Asian countries over 18 years, from 2002 to 2019. The selected time frame is based on the availability of data for all the selected variables considered in the study. The data are sourced from the World Development Indicators (WDI) database and World Governance Indicators (WGI) database provided by the World Bank (2019). A composite index of institutional quality has been prepared using Principal Component Analysis (PCA). The first principal component of the six institutional quality indicators were extracted using factor analysis [14,60] and has been used as Institutional Quality Index (IQ). The description of the variables, measurement and their data source are presented in Table 3.

Table 3.

Variable description and data sources.

Variables	Symbols	Measurement	Data Sources
Dependent variable
Foreign Direct Investment	FDI	Foreign direct investment, net inflows (% of GDP)	WDI
Independent Variables
Economic growth	GDPGR	GDP growth (annual %)	WDI
Financial development	FD	Domestic credit to private sector by banks (% GDP)	WDI
Total capital formation	GCF	Gross capital formation (% of GDP)	WDI
Trade	TRD	Trade (% GDP)	WDI
Exchange Rate	EXCHR	Official exchange rate (LCU per US$, period average)	WDI
Inflation	INFL	Inflation, consumer prices (annual %)	WDI
Control of Corruption	CC	Control of corruption estimate	WGI
Government Effectiveness	GE	Government effectiveness estimate	WGI
Political Stability	PS	Political stability and absence of violence/terrorism estimate	WGI
Regulatory Quality	RQ	Regulatory quality estimate	WGI
Rule of Law	RL	Rule of law estimate	WGI
Voice and Accountability	VA	Voice and accountability estimate	WGI
Institutional Quality	IQ	Institutional quality index prepared by PCA using six indicators (CC, GE, PS, RQ, RL and VA)	WGI

Source: Authors' compilation of data sources from WDI (2019) and WGI (2019)

Based on the review of literature and theoretical understanding, we consider the following model specification

FDI = f (IQ, GDPGR, FD, TRD, GCF, INFL, EXCHR)

The above variables are specified in the form of following equation (eq. (1)) and estimated with panel data techniques.

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_{1}i{q}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (1)

As the primary objective is to look at the impact of institutional quality on the FDI inflow, we further respecified the model by including each component of IQ individually, namely CC, GE, PS, RQ, RL, VA in equations (eq. (2)) to (eq. (7)) respectively, in order to check the robustness of the result. The new equations many be specified as follows,

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_1{cc}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (2)

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_1{ge}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (3)

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_1{pv}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (4)

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_1{rq}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (5)

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_1{rl}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (6)

$$\begin{array}{c}fd{i}_{i,t}={\beta }_0+{\beta }_1{va}_{i,t}+{\beta }_{2}gdpg{r}_{i,t}+{\beta }_{3}f{d}_{i,t}+{\beta }_4{trd}_{i,t}+{\beta }_5{gcf}_{i,t}+{\beta }_6{infl}_{i,t}+{\beta }_7{exchr}_{i,t}{+v}_{i,t}\end{array}$$ (7)

The empirical analysis began with estimating Pooled Ordinary Least Square (OLS), Fixed Effect (FE) and Random Effect (RE) Models. The choice among these approaches depends upon homogeneity or heterogeneity of data structure. Applying various specification test such as F test, Breusch Pagan LM Test and Hausman specification test, we have concluded that FEM is appropriate. However, under heteroskedasticity and autocorrelation problem, the classical least square approach of estimating FEM will be inappropriate [62].

However, the panel regression models in general and FEM in particular, are often characterized with problems of serial correlations, cross-sectional dependence and heteroskedasticity. Two alternative approaches usually followed to address these issues. The first approach is to use OLS estimator with a robust standard error, which should be robust to serial correlations and heteroskedasticity [[63], [64], [65]]. The most commonly used method for this is the use of clustered standard errors [66]. The second approach is the use of Generalized Least Square (GLS) and Feasible Generalized Least Square estimator. The GLS considers serial correlation, cross-sectional dependence and heteroskedasticity in the estimation directly. The Feasible Generalized Least Square (FGLS) estimation takes into consideration the clustering problems present in fixed effects panel [67].

In this study, we have applied Iterated-GLS (IGLS) which allows the estimation of the coefficient and standard error in the presence of AR [1] autocorrelation within panels and heteroskedasticity and cross-sectional correlation across panels. To apply IGLS, the following general model (eq. (8)) may be specified

$${\mathrm{y}}_{\text{it}}={\mathrm{x}}_{\text{it}}\mathrm{\beta }+{ϵ}_{\text{it}}$$ (8)

where i = 1,2 … m is the number of panelsand t = 1,2 … T time period, Y is dependent variable and X is explanatory variable. Thus, GLS result is provided by

$${\hat{\beta }}_{GLS}={\left(X^{\prime }{\hat{\Omega }}^{-1}X\right)}^{-1}{X}^{\prime }{\hat{\Omega }}^{-1}y$$

$${\widehat{Var}\hat{\beta }}_{GLS}={\left(X^{\prime }{\hat{\Omega }}^{-1}X\right)}^{-1}$$

where the ꭥ matrix is written as the Kronecker product

$$\Omega ={\Sigma }_{m\times m}\otimes I_{T_i{\times T}_i}$$

In GLS the known variance covariance matrix of error term denotes as Σ is used, where as in FGLS the estimated variance-covariance matrix will be obtained by substituting the estimator $\hat{\Sigma }$ for Σ.

where,

$${\hat{\Sigma }}_{ij}=\frac{{\widehat{{\epsilon }_i}}^{\prime }\widehat{{\epsilon }_j}}{T}$$

The OLS residuals are used to estimate Σ. In case of iterated estimation process, residuals AR obtained from the last fitted model. The GLS estimates and their associate standard error are calculated using ${\hat{\Sigma }}^{-1}$. When heteroskedasticity, serial correlations and cross-sectional dependence are present, the FGLS estimator is more efficient than the OLS. In I-GLS the process iterates over the estimated disturbance covariance matrix and estimates the parameter until the parameter estimates converges.,

However, in a dynamic panel data model, usually correlation exists between the lagged value of the dependent variable and the time invariant error component. Consequently, the estimators perform poorly with strictly exogeneous regressors [68]. found that when GLS is applied to a dynamic panel data model, with large N, it usually produces inconsistent estimates. On the other hand, when GLS is applied to an augmented panel model, it produces a constant estimator, owing to initial conditions (69). For short time series with each cross-section units, it provides consistent estimates [68,69,70]. However, FLGS of [71] provides consistent estimator in the augmented model even if asymptotic unknown variance-covariance matrix depends on distribution of error. For large sample, the result of two step FGLS estimator of Blundell and Bond is inconclusive, as it depends on the first-round regression estimator. Therefore, to produce a consistent estimator, iterated FGLS can be used. For a fixed effects models, with exogeneous variable as well as arbitrary intertemporal error variances, Iterated FGLS results in conditional maximum likelihood estimate [72]. For small sample (N = 5), I-FGLS has superior performance over other estimators like GMM. It also provides consistent estimator irrespective of homoscedastic or heteroscedastic error [72]. As our sample size is small, the I-FGLS provide both unbiased as well as consistent estimator of the parameter.

5. Empirical results and analysis

5.1. Descriptive statistics

This section begins with the results of descriptive statistics present to gain some insights into the characteristics of the data for the variables used in our study. The descriptive statistics provides a preliminary understanding of data.

Table 4 provides summary statistics for the variables used in the empirical analysis. We note that the dependent variable FDI, varies between −1.32 and 28.6, having a mean value of 3.9 and standard deviation 5. However, the institutional quality index (IQ) has a very low variation from −2.57 to 2.66. On the other hand, the other variables such as FD, GCF, GDPGR, TRD, EXCHR and INFL have high variations. The descriptive statistics of the individual institutional variables have also been reported. The standard deviation of none of the variable is zero, and in many cases, they are very high. The null hypothesis of normality is rejected by appropriate probability value of Jarque-Bera statistics and conclude that data are not normally distributed. This indicates wide disparity in terms of variability of data, which is good for panel study.

Table 4.

Descriptive statistics of the variables.

Variables	Mean	Std. Dev.	Min.	Max.	Skewness	Kurtosis	Jarque-Bera stat.	Prob. values of JB Stat.	Coeffcient of Variation (C.V)
FDI	3.9	5	−1.32	28.6	2.574	10.2	998.622	0.000	1.2821
IQ	2.94E-09	0.97	−2.57	2.66	0.244	2.425	7.239	0.027	3.30e+091
FD	47.07	34.84	1.11	133.14	0.902	2.674	42.824	0.000	0.7402
GCF	27.31	10.1	7.02	69.53	1.329	6.056	209.128	0.000	0.3698
GDPGR	5.84	3.56	−13.13	26.11	0.376	10.256	678.456	0.000	0.6096
TRD	98.76	81.15	0	437.33	2.095	8.039	547.711	0.000	0.8217
EXCHR	2407.15	5098.36	1.25	22602.05	2.305	7.592	539.834	0.000	2.1180
INFL	5.83	6.64	−18.11	57.07	3.150	20.377	4356.204	0.000	1.1389
CC	−0.4	0.9	−1.67	2.33	1.357	4.66	129.116	0.000	−2.2500
GE	−0.14	0.88	−1.62	2.44	0.84	3.643	41.569	0.000	−6.2857
PV	−0.55	1.12	−2.81	1.62	0.0120	2.112	10.052	0.006	−2.0364
RQ	−0.31	0.87	−2.34	2.26	0.67	4.001	35.817	0.000	−2.8065
RL	−0.35	0.8	−1.9	1.84	0.598	3.451	20.829	0.000	−2.2857
VA	−0.65	0.62	−2.23	0.46	−0.332	2.543	8.299	0.015	−0.9538

Source: Authors' calculation

5.2. Corelation matrix

Table 5 reports the results of the correlation matrix. The correlation matrix provides a preliminary understanding about the relationship of each variable among themselves as well as with dependent variable. It helps to understand the possible endogenous effects of the variables and mitigate with possible multicollinearity and over parameterization issue in the regression model. The results show that FD and TRD are positively correlated with FDI and are statistically significant. IQ, GDPGR and EXCHR are also positively correlated with FDI but lack significance. However, GCF and INFL are negatively correlated with FDI and are also statistically significant.

Table 5.

Correlation matrix.

Variables	VIF	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)
(1) FDI		1.00
(2) IQ	1.14	0.01	1.00
(3) GDPGR	2.969	0.09	−0.06	1.00
(4) FD	2.666	0.42*	0.05	−0.17*	1.00
(5) TRD	1.254	0.79*	−0.03	−0.03	0.66*	1.00
(6) GCF	4.325	−0.12*	0.04	0.09	0.09	−0.04	1.00
(7) INFL	1.528	−0.16*	−0.19*	0.18*	−0.29*	−0.23*	−0.04	1.00
(8) EXCHR	1.387	0.08	0.04	0.08	0.13*	0.07	0.12*	0.05	1.00
(9) CC	16.88	0.47*	0.13*	−0.19*	0.53*	0.70*	0.31*	−0.35*	−0.19*	1.00
(10) GE	18.817	0.48*	0.09	−0.24*	0.67*	0.73*	0.13*	−0.40*	−0.14*	0.91*	1.00
(11) PV	4.325	0.50*	0.09	−0.06	0.44*	0.59*	0.30*	−0.36*	0.15*	0.71*	0.68*	1.00
(12) RQ	11.114	0.51*	0.14*	−0.33*	0.63*	0.69*	−0.09	−0.47*	−0.16*	0.80*	0.91*	0.59*	1.00
(13) RL	23.688	0.44*	0.13*	−0.25*	0.64*	0.67*	0.23*	−0.40*	−0.17*	0.94*	0.96*	0.69*	0.88*	1.00
(14) VA	2.333	0.01	0.11	−0.21*	0.27*	0.15*	0.16*	−0.23*	−0.32*	0.43*	0.47*	0.02	0.49*	0.52*	1.00
	Mean VIF = 7.044

* shows significance at the 0.05 level.

In multiple linear regression models, sometimes the covariates are correlated with one another, leading to multicollinearity, causing the parameter estimates to be inaccurate. Multicollinearity in the regression function is identified when the simple correlation coefficient becomes equal to or more than 0.85. Another method is to calculate the Variance Inflation Factor (VIF) or value of tolerance. If the VIF exceeds 10 or the value of tolerance is less than 0.1, then it shows that problem of multicollinearity exists between the independent variables [73]. However, in case of panel data, if average VIF is less than 10, then multicollinearity is not a problem. In case of our study, some individual governance indicators such as GE, RQ and RL have high correlation and are statistically significant. Also, the VIF exceeds 10 in case of the individual governance indicators which confirms the presence of multicollinearity, but average VIF is 7.044 which is less than 10. To have more robust result and solve the problem of multicollinearity, Principal component analysis (PCA) is used. This method reduces the covariate dimensions by using the analyzation and manipulation of data matrices and simultaneously maximizing the amount of variation [74]. Therefore, we consider IQ index which has been constructed using PCA.

5.3. Stationary check

Before proceeding with empirical estimation, it is imperative to check the stationarity of the variables to avoid spurious regression. We have employed ADF-Fisher test in this study. The advantage of using this method is that it allows for heterogeneity across units as much as possible. Table 6 reports the unit root statistics of the variables used in the study in three categories: individual intercept, individual intercept and trend and none, in level and first difference. The results show that all the test rejected the null hypothesis of non-stationarity when the variables are used in first difference. Also, some of the variables are stationary at level, although it cannot be generalized. Thus, the variables are integrated of order 1.

Table 6.

Unit root test(ADF-Fisher).

Variables	Individual Intercept		Individual Intercept and Trend		None
	Level	First Difference	Level	First Difference	Level	First Difference
	Statistic (P-Value)	Statistic (P-Value)	Statistic (P-Value)	Statistic (P-Value)	Statistic (P-Value)	Statistic (P-Value)
FDI	84.35*** (0.00)	200.53*** (0.00)	78.45*** (0.00)	91.115*** (0.00)	29.11 (0.79)	164.91*** (0.00)
IQ	50.62** (0.05)	191.09*** (0.00)	37.59 (0.40)	118.383*** (0.00)	95.17*** (0.00)	201.22*** (0.00)
FD	21.98 (0.97)	98.22*** (0.00)	39.01 (0.34)	126.40*** (0.00)	10.22 (1.00)	268.40*** (0.00)
GCF	50.79** (0.05)	181.30*** (0.00)	49.70** (0.05)	35.88*** (0.00)	15.37 (1.00)	59.05*** (0.00)
GDPGR	121.76*** (0.00)	198.50*** (0.00)	100.88*** (0.00)	100.46*** (0.00)	41.07 (0.26)	148.00*** (0.00)
INFL	109.25*** (0.00)	241.83*** (0.00)	85.16*** (0.00)	228.12*** (0.00)	91.55*** (0.00)	286.57*** (0.00)
EXCHR	29.77 (0.76)	95.73*** (0.00)	19.30* (0.08)	228.12*** (0.00)	4.83 (0.96)	286.57*** (0.00)
TRD	40.68 (0.27)	105.64*** (0.00)	19.30* (0.08)	228.12*** (0.00)	4.83 (0.96)	286.50*** (0.00)
CC	64.73*** (0.000)	175.89*** (0.00)	19.3087* (0.08)	228.12*** (0.00)	4.83 (0.9634)	286.57*** (0.00)
GE	53.84** (0.03)	196.75*** (0.00)	19.30* (0.0813)	228.12*** (0.00)	4.832 (0.96)	286.57*** (0.00)
PV	79.86*** (0.00)	190.68*** (0.00)	19.30* (0.08)	228.1*** (0.00)	4.83 (0.96)	286.57*** (0.00)
RQ	67.13*** (0.00)	197.64*** (0.00)	19.30* (0.0813)	228.12*** (0.00)	4.83 (0.96)	286.57*** (0.00)
RL	49.01* (0.07)	183.52*** (0.00)	19.30* (0.08)	228.12*** (0.00)	4.83 (0.9634)	286.57*** (0.00)
VA	35.44 (0.49)	151.95*** (0.00)	19.3087 (0.08)	228.12*** (0.00)	4.83 (0.96)	286.57*** (0.00)

Source: Authors' calculation

5.4. Fixed effect model

The pooled model considers individual homogeneity across panel and can be estimated by ordinary least squares method (OLS). However, if the countries’ unobservable individual effects such as geographical location, colonial history, political regime, religions etc. are not considered, then pooled OLS gives measurement error in the estimated parameter [75], and biased results [76]. Table 5 reports the pooled OLS regression results. To capture heterogeneous characteristic of individual countries across panels, the FE or RE is more suitable. The Pooled OLS, FE and RE are applied with suitable diagnostic checking along with Heteroscedasticity and Autocorrelation Consistent (HAC) in standard errors. We find the existence of the problems of heteroskedasticity and autocorrelation from the Breusch-Pagan/Cook-Weisberg test for heteroskedasticity and Wooldridge test for autocorrelation respectively.

The choice between POLS with FE and POLS with RE are tested using F statistics as well as Breusch-Pagan LM test respectively. The results select FE and RE over POLS. Further, the Hausman test [77] has been used to examines the appropriateness of a FE or RE. The Hausman test assumes the null hypothesis of absence of correlation between unobserved individual effects with the independent variables, against the alternative of existence of correlation. The null hypothesis (above 0.05) is that the model is random effect, which implies that rejecting null hypothesis infers that the model is fixed effect. The details of these results are reported in Table 7 along with the fixed effect regression result with Hausman test statistics and the result of the Pesaran cross-sectional dependence test. Generally, there is a possibility of presence of common correlation and shocks among the countries, while dealing with panel data analysis. This occurs when the countries in consideration are associated regionally or globally [78]. studied this situation and concluded that when the economies are strongly interrelated, the hypothesis of cross-sectional independence is invalid. A cross-sectional dependency analysis by Ref. [79] has been used to check this association. The results in Table 7 show the presence of cross-sectional dependence among the variables. However, with diagnostic checking we found that Fixed Effect Model shows the presence of heteroscedasticity, autocorrelation and cross-sectional dependence. Therefore, to solve the problem we have applied iterated GLS (I-GLS) method.

Table 7.

Fixed Effect Model: Dep variable FDI.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)
GDPGR	0.082**	0.081**	0.079**	0.080*	0.083**	0.086**	0.082**
	(0.040)	(0.041)	(0.040)	(0.041)	(0.041)	(0.040)	(0.041)
FD	0.056***	0.056***	0.054***	0.055***	0.056***	0.052***	0.056***
	(0.010)	(0.010)	(0.010)	(0.010)	(0.010)	(0.010)	(0.010)
TRD	0.016***	0.016***	0.017***	0.016***	0.017***	0.017***	0.016***
	(0.006)	(0.005)	(0.006)	(0.006)	(0.006)	(0.005)	(0.006)
GCF	−0.004	−0.003	−0.008	−0.004	−0.004	−0.007	−0.004
	(0.023)	(0.023)	(0.023)	(0.023)	(0.023)	(0.023)	(0.023)
INFL	−0.007	−0.010	−0.009	−0.009	−0.006	−0.002	−0.009
	(0.022)	(0.022)	(0.021)	(0.022)	(0.022)	(0.022)	(0.022)
EXCHR	−0.000**	−0.000**	−0.000***	−0.000**	−0.000**	−0.000***	−0.000**
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
IQ	0.084
	(0.133)
CC		0.011
		(0.705)
GE			1.275
			(0.798)
PV				0.132
				(0.396)
RQ					0.401
					(0.589)
RL						1.485*
						(0.871)
VA							0.067
							(0.559)
Constant	0.282	0.259	0.700	0.364	0.349	1.064	0.333
	(0.895)	(0.995)	(0.933)	(0.955)	(0.905)	(1.009)	(1.120)
No. of Obs.	306	306	306	306	306	306	306
R-squared	0.151	0.149	0.157	0.150	0.151	0.158	0.149
adjusted R2	0.0780	0.0767	0.0850	0.0770	0.0782	0.0861	0.0767
F-Stat(7,281)	7.114	7.046	7.475	7.065	7.124	7.535	7.049
Prob F	7.82e-08	9.39e-08	2.97e-08	8.93e-08	7.61e-08	2.53e-08	9.33e-08
Diagnostic Checking for FEM
BP-LM Test for Independence Chi-Square(153)	314.010 0.000	322.351 0.000	286.319 0.000	320.094 0.000	313.707 0.000	297.312 0.000	321.503 0.000
Modified Wald test for groupwise heteroskedasticity	10555.90 0.0000	12988.57 0.000	12367.65 0.000	10332.11 0.000	11856.48 0.000	9824.14 0.000	13114.81 0.000
LM Test for Serial Correlation	135.595 0.000	136.226 0.000	132.553 0.000	135.738 0.000	135.958 0.000	133.496 0.000	136.140 0.000
Pesaran CSD	2.408 0.016	2.228 0.026	2.385 0.017	2.449 0.014	2.438 0.015	2.462 0.014	2.260 0.024
POLS Vs FEM vs REM
F Stat(17, 281), with p value (POLS vs FE)	17.85 0.000	17.90 0.000	18.05 0.000	16.87 0.000	17.97 0.000	18.34 0.000	17.81 0.000
Hausman Test FE vs RE	19.47 0.000	39.51 0.000	41.03 0.000	38.04 0.000	43.92 0.000	43.08 0.000	38.27 0.000
Breusch and Pagan LM Test (POLS Vs RE)	183.21 0.000	181.03 0.000	174.37 0.000	182.62 0.000	185.95 0.000	186.44 0.000	194.61 0.000

Standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1.

5.5. Fixed effect iterated GLS

We proceed with our analysis by using Iterated generalized least squares (I-GLS) in the fixed effect model. This method allows estimation in the presence of AR [1] autocorrelation within panels and cross-sectional correlation and heteroskedasticity across panels. Since the estimated FEM is having heteroskedasticity, autocorrelation and cross-sectional dependence, the I-GLS deems fit for our model. Table 8 reports the I-GLS regression results. The results show that institutional quality has a positive and significant impact on FDI inflows in South Asian and Southeast Asian countries. This implies that when the quality of institutions is better, it attracts more FDI. The reason behind this association may be because of, better institutional quality helps in raising business profitability, enhancing efficiency in production and also encourages better allocation of productive resources. On the other hand, poor institutions accepting corruption can lead to increase in investment costs and a reduction in profits. Also, political risks, violence and terrorism act as hindrances for foreign investment as the investors are reluctant to invest in an unstable economy. The foreign investors prefer the countries having open and transparent legal and regulatory administration, open markets, effective delivery of government services and non-corrupt institution. Our results are consistent with [43] in their study on South America during the year 1996–2015 and with [40] for their study on developed countries for the period 2002–12. On the contrary, in the same study [40] found no significant impact of institutional quality on FDI inflows in case of developing countries.

Table 8.

I-GLS Regression results.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Dep variable: FDI
GDPGR	0.1299*** (0.0000)	0.1665*** (0.0000)	0.1441*** (0.0000)	0.1319*** (0.0000)	0.1291*** (0.0000)	0.1375*** (0.0000)	0.1709*** (0.0002)
FD	0.0174*** (0.0000)	0.0233*** (0.0000)	0.0117*** (0.0000)	−0.0037*** (0.0000)	0.0523*** (0.0000)	0.0308*** (0.0000)	0.0271*** (0.0000)
TRD	0.0522*** (0.0000)	0.0320*** (0.0000)	0.0378*** (0.0000)	0.0444*** (0.0000)	0.0284*** (0.0000)	0.0328*** (0.0000)	0.0057*** (0.0001)
GCF	−0.0363*** (0.0000)	−0.0022*** (0.0000)	0.0128*** (0.0000)	−0.0231*** (0.0000)	−0.0028*** (0.0000)	0.0078*** (0.0000)	−0.0461*** (0.0000)
INFL	−0.0577*** (0.0000)	−0.0427*** (0.0000)	−0.0422*** (0.0000)	−0.0433*** (0.0000)	−0.0440*** (0.0000)	−0.0314*** (0.0000)	−0.0891*** (0.0000)
EXCHR	−0.0001*** (0.0000)	−0.0000*** (0.0000)	−0.0001*** (0.0000)	−0.0003*** (0.0000)	−0.0002*** (0.0000)	−0.0001*** (0.0000)	−0.0002*** (0.0000)
IQ	0.1499***
	(0.0000)
CC		0.2990***
		(0.0000)
GE			0.2140*** (0.0000)
PV				0.0346*** (0.0000)
RQ					0.2891*** (0.0007)
RL						−0.5820*** (0.0002)
VA							−0.2534*** (0.0019)
_cons	−1.4884*** (0.0002)	−0.3311*** (0.0001)	−1.1156*** (0.0002)	−1.4471*** (0.0000)	−0.8652*** (0.0014)	−1.7718*** (0.0003)	0.8170*** (0.0087)
Obs.	306	306	306	306	306	306	306
Wald chi square	2.80E+09***	4.50E+10***	7.18E+09***	4.28E+10***	7.83E+07***	1.66E+09***	6838224***
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000
Breusch-Pagan LM test of independence: chi2(153)	340.657***	308.864***	290.267***	318.215***	340.76***	322.095***	278.53***
P value	0.000	0.000	0.000	0.000	0.000	0.000	0.000
Modified Wald test for groupwise heteroskedasticity	21905.12***	6368.61***	16879.62***	18483.36***	5437.9***	17368.61***	81565.65***
Prob > chi2 (18)	0.000	0.000	0.000	0.000	0.000	0.000	0.000

Standard errors are in parenthesis.

***p<0.01, **p<0.05, *p<0.1.

The results indicate that FDI has been positively impacted by GDP growth, which is backed by several studies such as [50,61,80,81]. It implies that with an increase in GDP growth, the level of foreign investment rises. This may be due to the fact, that the foreign investors see potential of growth in these economies, and hence an increase in their profits. Even [55] in the eclectic paradigm, mentioned that market size (proxied by GDP growth in our study) helps in attracting FDI inflows. However, our result is in contrast with studies by Refs. [82,83] who found a negative and significant relationship between FDI inflows and GDP growth.

The variables trade and financial development are positive and significant. This implies that the open economies have incentive to attract more market-seeking FDI. The foreign investors prefer investing in countries with sizable trade volume [84]. This finding is consistent with that of [61]. However, gross capital formation, exchange rate and inflation have a negative effect on FDI inflows. This can be explained as high inflation rates tend to discourage the FDI inflows. The companies who cannot bear loss may decrease their foreign investment to evade any loss due to exchange rate volatility, inflation and increase in domestic investment [85]. However, the coefficient of the interaction of exchange rate with institutional quality is positive and significant, suggesting that, jointly with effective institutional infrastructure, exchange rate can help in attracting FDI inflows.

We also find that, two of the individual governance indicators such as rule of law and voice and accountability show that they negatively and significantly affect FDI inflows as compared to the institutional quality index which affects the FDI inflows positively and significantly. This implies that if we consider each of the institutional indicators alone, then that is not sufficient to attract FDI. Therefore, we need to consider all the variables together to study their effect on FDI inflows. Thus, in an economy to attract FDI, it should take into consideration all the factors of governance as the foreign investors looking to invest in a country will investigate all the parameters of institutional quality before making their investment.

6. Conclusion and policy implications

This study examines the effect of institutional quality on FDI inflows in South Asian and Southeast Asian countries by controlling the effects of macroeconomic instabilities during the period 2002–2018. We have used six governance indicators as a measure of institutional quality using PCA. We find that institutional quality affects FDI inflows in this region positively. Furthermore, we have investigated the impact of each of the six variables separately on FDI inflows and found that except for the 2 variables, that is, rule of law and voice and accountability, all the other institutional variables have a positive impact on FDI inflows. Also, it was found that GDP growth, international trade and financial development have a positive and significant impact on FDI inflows in the South Asian and Southeast Asian countries. On the other hand, gross capital formation, inflation and exchange rate have a negative and significant impact on FDI inflows.

The results of this study have several implications for the policymakers. For the host economies to benefit from FDI inflows, they need to implement proper rule and regulations in the market. The regulations should be aimed towards investor protection, business facilitation and simultaneously, enhancing productivity. There should be strict measures conducive to control corruption. Measures should be taken towards building a better and an efficient institutional system that encourages foreign investment. The political systems should be democratised and political risks should be reduced. Also, political hindrances should be removed to pave way for bureaucratic procedures such as, business entries and streamlining licenses.

Another important implication is to strengthen the judicial system of the economies. It needs to be structured in a way to provide protection to the foreign investors and also, make the investing procedure free of complications. A transparent and impartial legal system is essential for ensuring the control of corruption and stability of the government. It is also important to introduce structural reforms and apply a stabilization program to reduce political risk in the economies. Thus, an ideal governance system would help the South Asian and Southeast Asian countries in attracting FDI and thus would have positive spillovers to other economic activities that are vital for economic growth and development.

The study is however subject to certain limitations such as the availability of data. The data for the institutional quality index for consecutive years are available from 2002 onwards. The time frame of the study has been taken considering the availability of all the variables taken in the study so as to make the dataset a balanced panel dataset. Further, future research can focus on the effect of the ongoing coronavirus pandemic and its effect on institutions and FDI inflows. In this study, we have considered six indicators of institutional quality, further research can be carried out by taking other factors such as economic freedom. Also, as per the development of new empirical methodology, advanced methodologies can be applied for the empirical analysis.

Data availability statement

The datasets generated and/or analysed during the current study are available in the World Development Indicators.

Link: https://databank.worldbank.org/source/world-development-indicators

CRediT authorship contribution statement

Padmaja Bhujabal: Writing – review & editing, Writing – original draft, Data curation. Narayan Sethi: Supervision, Resources. Purna Chandra Padhan: Validation, Software, Methodology, Formal analysis.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.