Does Capital Account Liberalization Spur Entrepreneurship: The Role of Financial Development

Abstract

Entrepreneurship is regarded as the cornerstone of the sustainable development of a society. In this study, we empirically investigate the possible economic impacts of capital account liberalization on entrepreneurship. Using a panel dataset of 103 countries and regions and the system generalized method of moments (GMM), this paper demonstrates a positive relationship between capital account liberalization and entrepreneurship in developed economies whereas a negative relationship in developing economies. Furthermore, domestic financial development plays an important moderating role in the relationship between capital account openness and entrepreneurship. Specifically, the negative impact of capital account liberalization in developing economies is mitigated by a high degree of domestic credit and equity market development, the continuous deepening of finance and better financial inclusion. Our findings imply that domestic financial development is an essential prerequisite for the opening of a country’s capital account, especially for developing countries.

1. Introduction

Since the 1990s, the critical role of entrepreneurship has been gradually recognized. Looking back at the development history of all enterprises in the world, it is not difficult for us to find such a general phenomenon, that is, those corporations with strong entrepreneurial spirit are always capable of achieving better growth and obtaining better corporate returns than those with a weak entrepreneurial spirit. China’s 40 years of reform and opening up has also cultivated a large body of entrepreneurial talent specializing in discovering and making use of market opportunities and bearing business risks. Entrepreneurship has long been viewed as the endogenous driving force for promoting technology innovation [1,2], increasing employment opportunities [3,4,5], boosting economic growth [6,7,8,9] and enhancing sustainable development [10,11,12] of a nation.

Meanwhile, economic globalization and financial liberalization have gradually become the new engine for world economic growth since the 1980s. The world’s major economies, including China, have benefited from economic globalization. In the 1990s, with the speeding up of international capital flows, developed countries gradually canceled capital controls and implemented the reform of capital account liberalization. The liberalization of trade in services and financial liberalization reform further promoted the process of global financial liberalization. Therefore, developing countries followed the footsteps of financial reform in developed countries, lifting restrictions on international capital inflows and joined the waves of global financial liberalization. Nevertheless, some developing countries still experienced varying degrees of financial crises and economic recessions after capital account liberalization, and these countries imposed strict restrictions on capital flows again. After the Asian financial crisis, emerging economies and some developed countries came to be aware of the negative impacts and great challenges ahead of capital account liberalization.

In recent years, the literature on capital account liberalization and entrepreneurship has grown considerably. Researchers have paid increasing attention to the critical role played by capital account liberalization and its macro impacts on economic growth [13,14], aggregate productivity [15,16], innovation [17,18] and the income distribution inequality [19,20,21]. A growing number of scholars have also explored the potential determinants of entrepreneurship at the macro level [22,23,24]. Studies devoted to entrepreneurship in developing and emerging countries are a useful supplement to existing research [25,26,27,28,29,30,31,32,33]. The capital account liberalization and entrepreneurship nexus, however, is inconclusive and inconsistent in the literature. Vlachos and Waldenström [34] find in their research that capital account liberalization increases the growth rates of firm creation among externally dependent industries. Audretsch et al. [35] argue that open conditions can promote entrepreneurs’ risk-bearing ability. Bekaert [15] argues that the growth effect of financial liberalization is permanent, which comes from the development of the securities market and the banking sector and the improvement of institutions’ quality. Hartwell [36] argues that If governments want to promote entrepreneurship, they should avoid capital controls and encourage the improvement of the investment environment. Gregory [37] uses a panel data set of 62 countries over the period from 1995 to 2013 and finds that capital controls have a negative effect on entrepreneurship in emerging market countries, while it has a positive effect on entrepreneurship in developed markets. Therefore, the implication of existing literature is that capital account liberalization could reduce the cost of raising capital, relieve financial constraints, increase the efficiency of capital allocation and diversify potential risk for entrepreneurs. For those reasons, most of the studies tend to support the positive relationship between financial liberalization and entrepreneurship. However, most of the literature often ignores the potential heterogeneity between developed and developing countries, so the empirical evidence has not conclusively demonstrated that the influencing patterns of capital account liberalization in developed countries are the same as that of developing countries.

Subsequently, a growing large body of literature explores the effects of financial development on entrepreneurship [18,38,39,40,41,42]. Financial development is an important concept to measure the strength of the domestic financial system of a country or region. Expanding the scale of the domestic financial industry, perfecting the function and operation of the financial system and improving the level of domestic financial development can promote economic development [43]. Financial development could promote the development of entrepreneurship through a series of possible channels, in which it could support entrepreneurial innovation [38] and alleviate the credit constraints for entrepreneurs [40,44,45,46]. Thus, the core function of finance is to select and fund the most innovative entrepreneurs. However, we barely find literature integrating capital account liberalization, financial development and entrepreneurship into a unified research framework. Bayar et al. [47] tried to investigate the influencing factors, such as financial sector development, foreign direct investment inflows and trade and financial openness, on entrepreneurship, using information from 15 upper middle income and high-income countries over the period of 2001–2015. However, their research only covers a small sample of developed countries and ignores the heterogeneous effects that may exist in developing economies.

Is there a stable relationship between capital account liberalization and entrepreneurship? Does capital account liberalization significantly hinder or spur entrepreneurship? To answer these questions and resolve the existing disputes, after a rigorous review of the literature, we try to concentrate on the potential heterogeneity between developed and developing countries and the critical role financial development plays in the relationship between capital account liberalization and entrepreneurship. The purpose of this study is to examine the role of financial development when choosing the degree of capital account liberalization, where the objective is to maximize the entrepreneurship of a nation.

Unlike previous studies, this paper makes several marginal contributions to the existing literature: first, from the perspective of capital account openness, we try to explore other essential macro factors that may significantly influence a country’s overall entrepreneurship. Compared to some studies focusing exclusively on a single country, we include a rich sample of many countries the regions and use the cross country empirical evidence to disentangle the impact of capital account liberalization on entrepreneurship; second, this paper is probably the first paper trying to include financial development into the research framework of capital account liberalization and entrepreneurship, which could enable us better understand the important role domestic financial development plays in the relationship between capital account liberalization and entrepreneurship; third, we divide financial development into different financial market sectors, namely, the credit market development and the equity market development and so on, which could lead us to explore how different financial market development has a heterogeneous effect on capital account liberalization and entrepreneurship.

The remainder of this paper is organized as follows: Section 2 makes the theoretical analysis and builds the corresponding research hypotheses; Section 3 introduces the empirical strategy, variable measurements and data; Section 4 conducts a comprehensive empirical analysis and carries out the in-depth mechanism and heterogeneity analysis; Section 5 provides the robustness checks and deals with the endogeneity issues; and Section 6 details the final conclusions and outlines potential policy implications.

2. Theoretical Analysis and Research Hypotheses

2.1. The Relationship between Capital Account Liberalization and Entrepreneurship

Capital account liberalization comes from the theory of financial liberalization, which originates in the research on financial repression, financial liberalization and financial deepening from McKinnon [48] and Shaw [49]. They both find in their research that the financial markets in developing countries have serious irrational phenomena of distortions of financial development (credit controls, maximum interest rate restrictions, low exchange rate policies, restrictions on the free flow of capital and barriers to entry into the financial industry). To eliminate the above phenomenon of financial repression, a series of financial liberalization reforms should be adopted, such as freeing the maximum restrictions on interest rates and exchange rates, lifting the credit control, removing the restrictions on entry barriers of the financial industry, allowing the development of the private financial market, realizing self-regulation of the banking industry and freeing the capital account control, etc. Nevertheless, the empirical literature on capital account liberalization reveals that it is generally difficult to find robustly positive evidence that opening up a country’s capital account leads to higher economic growth [30].

Theoretically, the concept of entrepreneurship is quite complex, dynamic and evolving over time. There are basically three different representative schools of entrepreneurship research, namely, the Schumpeter entrepreneurship, the new classical entrepreneurship, and the Austrian school of entrepreneurship. First, Schumpeter, the founder of innovation theory, has put forward creatively the innovative spirit of entrepreneurs [50]. According to Schumpeter’s point of view, the essence and basic function of entrepreneurs is “innovation”. Additionally, entrepreneurs’ innovative activities include the following: (1) adopting a new product or new features of a product; (2) employing a new production method; (3) acquiring a new market; (4) control of a new source of supply of raw materials or intermediate goods; (5) implementing a reorganization of industry, and so on. Knight [51] and Schultz [52] are the representatives of the new classical school, who pay more attention to the risk-bearing abilities and adventurous spirits of entrepreneurs. Mises [53] and Kirzner [54] are the representatives of the Austrian school, who contend that entrepreneurs are alert to market opportunities and could make full use of market information that is beneficial to both sellers and buyers. Austrian school’s conception about entrepreneurship actually implies that rational allocation of resources can be implemented only by entrepreneurs.

It is worth noting that Schumpeter also put forward the idea that the core function of finance is to screen entrepreneurs with an innovative spirit and provide them with adequate credit support so as to help entrepreneurs recombine various factors of production and carry out innovative activities. It can be seen that Schumpeter also has an important thought that deserves great attention: the essence of finance is to provide enough financial support for entrepreneurs. King and Levine [31] also point out that a country’s financial system should provide a series of financial services to enhance the innovative activities of entrepreneurs, which mainly include: (1) evaluating and selecting the most promising entrepreneurs; (2) mobilizing savings in order to finance entrepreneurial projects with the greatest potential for success; (3) dispersing the risks of entrepreneurial innovation activities; (4) estimating the potential returns of innovative activities, etc.

Some scholars study the impact of capital account liberalization on a country’s entrepreneurship, and their theoretical analysis and empirical research prove that: the higher the degree of openness of a country’s capital account, the fewer restrictions on international capital flow, the more financing or investment opportunities and better financial services available to entrepreneurs, which is conducive to the continuous creation of more entrepreneurs engaging in entrepreneurial or innovative activities. What is innovative about these studies is that they link capital account openness with entrepreneurship and analyze the critical role of capital account openness in fostering entrepreneurship; however, they tend to ignore the significant heterogeneity between developed and developing countries in terms of economic foundation and institutional background, including the different economic effects of capital account openness.

There is, however, an important reality that we cannot ignore. The financial systems in developing countries lag far behind those of developed countries, leading to insufficient capital account openness or even failure of liberalizing their capital accounts. Market distortion resulting from capital controls in developing countries makes it more difficult for them to obtain finance or investment support, which seriously stifle entrepreneurship and restrict business expansion. The failure of capital account reform in developing economies inevitably causes the inequality of credit support, the low efficiency of capital allocation and a scarcity of resources to be allocated, including the special resource of entrepreneurship. In this mode, the inadequate risk diversification mechanism in developing countries makes entrepreneurial activities riskier and often leads to a higher failure rate. What’s worse, the relatively poor economic situation, lower degree of economic openness, fragile political and legal system and complex business environment of developing countries exacerbate the influence of an underdeveloped financial system on entrepreneurship. Accordingly, this study proposes the following Hypothesis 1:

Hypothesis 1.

The impact of capital account liberalization on entrepreneurship in developed economies is positive, while the impact of capital account liberalization on entrepreneurship in developing economies is negative.

2.2. The Mediating Role of Domestic Financial Development

Subsequently, we try to propose and investigate a possible spillover effect from domestic financial development, especially for the developing countries. The financial systems of developing countries suffer from two major problems: financial constraints and inefficient allocation of capital.

Financial constraints or financial frictions fail to stimulate innovation and foster entrepreneurship. Financial constraints (or credit constraints) refer to the inability of potential entrepreneurs to obtain sufficient start-up capital for new businesses. Many researchers have found that financing constraints are a dominant factor inhibiting entrepreneurial activities [36,54,55,56,57,58,59,60]. Entrepreneurs with innovative potential are often excluded from starting businesses due to the lack of sufficient capital. The unequal access to financial markets in developing countries greatly reduces the generation of entrepreneurial activities.

The allocation efficiency of capital refers to whether capital is efficiently and cheaply provided for entrepreneurs in need of funds in a timely manner and the degree of capital market in allocating scarce resources to the most efficient entrepreneurs. Due to the imperfect financial system in developing countries, there is serious inefficiency of capital allocation, which is not conducive to the cultivation of entrepreneurship and the growth of entrepreneurs. Consequently, ineffective allocation of capital is always associated with slow spillover and diffusion of knowledge and technology, diseconomies of scale for corporations and low firm start-up rates.

In the case of capital account liberalization reforms in developing countries, when the domestic financial development is relatively inadequate, financing constraints and inefficiency in capital allocation make foreign capital and domestic capital lack effective channels to connect each other, which hinder the free flow of knowledge, technology, and management that accompanies capital flows. Meanwhile, a large sudden inflow or outflow of capital in a short period of time in emerging markets will make it impossible for domestic entrepreneurs to obtain effective and timely financial support. Consequently, emerging market countries often experience domestic financial crises due to large capital outflows or inflows. To make matters worse, due to the lack of perfect investment or financing channels in the domestic financial market, for example, the underdevelopment of credit market and equity market restricts the entry of foreign capital; thus, foreign capital often cannot find suitable investment projects or make serious investment failure, which, in turn, leads fewer domestic entrepreneurs to engage in entrepreneurship and innovation.

Inversely, when the level of financial development is high in some developing countries, it could effectively help to break the financing constraints and improve the efficiency of capital allocation. Sound credit and equity markets are conditionally open to foreign capital, which could let domestic entrepreneurs have more financing channels and capital investment resources. At the same time, the effective flow of capital at home country and abroad will also bring about the free flow of technology, talent and management etc., which are conducive to the development of domestic entrepreneurship. A well-developed domestic financial system could prevent huge economic fluctuations associated with capital flows, and foreign capital can be traded in appropriate domestic financial markets. Based on the above analysis, we propose the following Hypothesis 2:

Hypothesis 2.

Domestic financial development plays an important moderating role in the relationship between capital account liberalization and entrepreneurship, and it is an essential prerequisite for the opening of a country’s capital account, especially for developing economies.

3. Model, Methodology and Data

This study builds an econometric model to analyze the effect of capital account liberalization on entrepreneurship and on this basis and analyze the crucial role financial development plays in the relationship between financial development and entrepreneurship.

3.1. Model and Methodology

We employ the dynamic panel model developed by Gregory [37] and the two-step system generalized method of moments (GMM) to study the relationship between capital account liberalization and entrepreneurship, and the baseline model is shown as below:

$${\mathrm{NBD}}_{\mathrm{i},\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{NBD}}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\phi }}_1{\mathrm{Kaopen}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\phi }}_2{\mathrm{FD}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\xi }}_{\mathrm{i}}+{\mathsf{\eta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

where ${\mathrm{NBD}}_{\mathrm{i},\mathrm{t}}$ is a valid measure of entrepreneurship; ${\mathrm{NBD}}_{\mathrm{i},\mathrm{t}-1}$ denotes the first-order lagged term of entrepreneurship; ${\mathrm{Kaopen}}_{\mathrm{i},\mathrm{t}}$ is a valid measure of capital account liberalization; ${\mathrm{X}}_{\mathrm{i},\mathrm{t}}$ represents a vector of control variables that could possibly affect entrepreneurship; ${\mathsf{\xi }}_{\mathrm{i}}$ is a proxy for the unobserved country specific effect; ${\mathsf{\eta }}_{\mathrm{t}}$ is the time-specific effect; ${\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$ is the residual term; i and t denote country and year, respectively; ${\mathsf{\alpha }}_0$ is the constant term; ${\mathsf{\alpha }}_1,{\mathsf{\phi }}_1{\mathrm{and} \mathsf{\phi }}_2$ are the corresponding coefficients. The lagged term of capital account liberalization is added into the regression equation to reflect the dynamic process of capital account liberalization and to eliminate the potential impact of uncontrolled factors.

In particular, if the marginal effect of capital account liberalization is not constant and is non-linear, a simple linear model will incur the omitted variable bias, that is to say, the effect of capital account liberalization on entrepreneurship could possibly be influenced by other factors, such as the degree of domestic financial development. Thus, we add the interaction term between capital account liberalization and financial development to study the possible moderating effect of financial development, and the dynamic panel model is shown as below:

$${\mathrm{NBD}}_{\mathrm{i},\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{NBD}}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\phi }}_1{\mathrm{Kaopen}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\phi }}_2{\mathrm{FD}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\phi }}_3{\mathrm{Kaopen}}_{\mathrm{i},\mathrm{t}}{\ast \mathrm{FD}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\xi }}_{\mathrm{i}}+{\mathsf{\eta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

For dynamic panel data, as the lagged independent variable is related to individual effects and error terms on each cross-section, both OLS estimation and random effect estimation of the panel data are biased and inconsistent. Anderson and Hsiao [61] first suggest using a first-order difference equation to eliminate the influence of individual effect, but there is still a negative correlation between lagged independent variables and error terms in the transformed model, resulting in a deviation of the estimated coefficient from the true value. Then, Anderson and Bond [62] use Monte Carlo simulation to find that, compared to the estimation of fixed effect and the estimation using differential instrumental variables, the autoregressive coefficient of differential GMM estimation has the smallest error and variance. However, differential GMM cannot estimate the parameters of variables that do not change at any time point. Meanwhile, when the time span is relatively small, the past value of the variable can only convey less information to the future value. If the horizontal lag value is taken as the instrumental variable, the problem of the weak instrumental variable will occur, which will affect the effect of model estimation. For this reason, Blundell and Bond [63] suggest combining the level equation to the differential equation and using more moment conditions, namely the system GMM estimation.

The system GMM estimation method integrates the advantages of differential GMM and horizontal GMM estimation methods. On one hand, it effectively avoids the problem of overidentification of instrumental variables that may occur in differential GMM estimation, and on the other, it achieves more efficient estimation than horizontal GMM because instrumental variables are not correlated with the disturbance terms. Meanwhile, the estimation results of system GMM are more reliable than OLS and fixed-effects models. Therefore, we adopt a two-step system GMM method to estimate the parameters of all variables and use the Stata software to implement the empirical analyses in our study.

3.2. Variables and Data

Given the concerns for data availability and empirical reliability, we use a sample of 103 countries and regions (for simplicity, hereinafter referred to as countries) from 2007 to 2018. Variable selections and data sources are shown as below.

3.2.1. Dependent Variable

Entrepreneurship. In an empirical study, scholars usually choose different proxies for entrepreneurship according to different research questions. However, due to the limitations of different proxies themselves, there is still no unified measure for entrepreneurship. At present, scholars mainly study the connotation of entrepreneurship from the perspective of starting a new business and innovation, and basically, it has the following kinds of proxies for entrepreneurship: self-employment rate [64,65,66], the ownership ratio [67], density of new business entry [68,69], the relative market share of small businesses [35], number of inventions and patents [70], Global Innovation Index, Total Entrepreneurial Activity (TEA) index from the Global Entrepreneurship Monitor [71]. In this paper, we take the density of new business entry as the proxy for entrepreneurship. To measure this entrepreneurial activity, annual data is collected directly from 155 company registrars on the number of newly registered firms in the World Bank’s entrepreneurship survey and database. The survey dataset is available from 2006 and renews annually.

3.2.2. Independent Variables

Key independent variable—capital account liberalization. We use the Chinn–Ito kaopen index as the proxy for capital account liberalization. The Chin–Ito index is an index measuring a country’s degree of capital account openness, which was initially introduced by Chinn and Ito [72]. Chinn and Ito [72] use the first principal component of the original variables pertaining to the tabulation of restrictions on cross-border financial transactions reported in the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER) to construct the comprehensive capital account openness index. The newest dataset contains the Chinn and Ito annual index series for the time period of 1970–2018 for 182 countries.

Moderator variable—Financial Development. We include financial development as the moderator variable to study the moderating effect in our model. As financial development is a holistic concept that basically reflects the improvement of the overall financial system and the revolutionary progress of the financial industry, there is always no unified measure in the previous literature. However, we could break down the financial development of a nation into different market sectors. As for the credit market, domestic credit to the private sector (PC) [38] and domestic credit provided by financial institutions and banks are available in the WDI database. However, the latter only contains a few countries, so here, we choose the domestic credit to the private sector as the proxy for credit market development. Domestic credit to the private sector refers to financial resources provided to the private sector by financial corporations, which could better reflect the efficiency of banking financial institutions and the credit support provided to private enterprises. As for the stock market development, the total value of stocks traded [73] and market capitalization of listed domestic companies (SMC) [18,39] are chosen to reflect the development of the equity market. Here, we use SMC as the proxy for stock market development as the total value of stocks traded has a large number of missing data. At the same time, broad money (M2) is also considered to look into the development of the money market because the ratio of M2 to GDP is mostly used as the proxy for the degree of a country’s financial deepening [48] in the literature. Last but not least, we also include the degree of financial inclusion (FI) as a proxy for domestic financial development to see how financial products and services are accessible and affordable to businesses and entrepreneurs.

3.2.3. Control Variables

To alleviate the endogeneity problems caused by omitted variables, a series of control variables are included in our model: trade openness refers to the ratio of total imports and exports to GDP, labeled as “Trade”; life expectancy refers to the log value of a nation’s average life expectancy, labeled as “Life”; economic development refers to the log value of real GDP per capita, labeled as “RGDP”; country-level political risk refers to the political risk rating of a country, labeled as “PRR”. The data of political risk rating comes from the International Country Risk Rating (ICRG) database, which assesses the political stability of a country on a comparable basis with other countries by assessing risk points for each of the component factors of government stability, socioeconomic conditions, investment profile, internal conflict, external conflict, corruption, military in politics, religious tensions, law and order, ethnic tensions, democratic accountability and bureaucracy quality. Political risk ratings range from a high of 100 (least risk) to a low of 0 (highest risk). The other data all comes from the World Bank’s World Development Index (WDI) database.

The detailed data descriptions and measurements of the variables used in our model are presented in

Table 1. Variables and measurement.

Variable Type	Variable Name	Symbol	Measurable Indicator
Dependent Variable	Entrepreneurship	NBD	New business density (new registrations per 1000 people ages 15–64)
Core Independent Variable	Capital Account Liberalization	Kaopen	The Chinn-Ito index (first principal component)
Moderating Variable	Financial Development	PC	Domestic credit to private sector (% of GDP)
		SMC	Market capitalization of listed domestic companies
		M2	Broad money (% of GDP)
		FI	Commercial bank branches (per 100,000 adults)
Control Variables	Trade Liberalization	TRA	Total import and export (% of GDP)
	Life Expectancy	LE	Log value of a nation’s average life expectancy
	Economic Development	GDPP	Log value of real GDP per capita
	Political Risk	PRR	The political risk rating from ICRG

Note: GDP denotes gross domestic product.

. According to the data availability of databases this paper uses, we first delete economies that are not individual countries. In order to avoid the influence of outliers on the empirical results, all variables are further reduced by 1% on both sides. We then obtain a sample of 103 countries, including 32 developed countries and 71 developing and emerging markets. The sample countries participating in the baseline regression and their classifications are detailed in the Appendix

Table A1. Country codes of Sample countries and regions and classification.

Developed Economies
AUS	AUT	BEL	CYP	CZE	DNK
EST	FIN	FRA	DEU	GRC	HKG
ISL	IRL	ISR	ITA	JPN	KOR
LVA	LTU	MLT	NLD	NZL	NOR
PRT	SGP	SVK	SVN	ESP	SWE
CHE	GBR
Developing and Emerging Economies
ALB	ARE	ARG	ARM	AZE	BFA
BGD	BGR	BHR	BLR	BRA	BWA
CHL	CIV	COL	CRI	DOM	DZA
GAB	GHA	GIN	GTM	HRV	HTI
HUN	IDN	IND	JAM	JOR	KAZ
KEN	KWT	LBR	LKA	MAR	MDA
MDG	MEX	MLI	MMR	MNG	MWI
MYS	NAM	NER	NGA	OMN	PAK
PAN	PER	PHL	POL	QAT	RUS
SAU	SEN	SLE	SLV	SUR	TGO
THA	TUN	TUR	TZA	UGA	UKR
URY	VNM	ZAF	ZMB	ZWE

. The baseline regression data cover the period from 2007 to 2018 because most of the data of new business entry density are only available from 2007, and the kaopen index is only updated to 2018. Most of the variables are extracted from the World Development Index (WDI) database of the World Bank, except for the data of capital account liberalization and country political risk. The summary statistics of all variables in the baseline regression are presented in

Table 2. Descriptive statistics of the baseline regression.

Variable	Obs.	Mean	SD	Min	Max
NBD	103	3.771	5.076	0.008	39.040
Kaopen	103	0.903	1.563	−1.920	2.334
PC	103	69.773	50.885	3.121	255.310
TRA	103	95.521	63.792	0.167	442.620
LE	103	4.296	0.119	3.840	4.442
GDPP	103	9.142	1.402	5.929	11.431
PRR	103	68.916	11.472	42.380	93.670

Note: Obs. denotes number of countries in the baseline model. SD denotes standard deviation.

. In this paper, based on different research aims in each section, we conduct different regressions in the empirical analysis. Given space limitations, the descriptive statistics of the other regressions are not presented but are available from authors upon request.

The pairwise correlation matrix among the variables of the baseline model is presented in

Table 3. The correlation matrix of the variables.

Variable	NBD	Kaopen	PC	TRA	LE	GDPP	PRR
NBD	1.000
Kaopen	0.398	1.000
PC	0.595	0.418	1.000
TRA	0.488	0.315	0.316	1.000
LE	0.350	0.576	0.603	0.304	1.000
GDPP	0.457	0.649	0.675	0.333	0.832	1.000
PRR	0.476	0.634	0.649	0.381	0.728	0.846	1.000

. When the correlation coefficient of all variables is less than 0.85, it is generally considered that there is no problem of multicollinearity in the model [74]. As can be seen from Table 3, the correlation coefficients of all variables are not more than 0.85; therefore, the model of this paper is not affected by the multicollinearity problem. We also calculate the correlation coefficients of other regression variables, and they are all less than 0.85. The descriptive statistics of the correlation matrix of the variables for developed and developing countries individually are presented in the Appendix

Table A2. The correlation matrix of the variables in the developed economies.

Variable	NBD	Kaopen	PC	TRA	LE	GDPP	PRR
NBD	1.000
Kaopen	−0.001	1.000
PC	0.444	0.437	1.000
TRA	0.358	0.195	0.125	1.000
LE	0.055	0.524	0.591	0.148	1.000
GDPP	−0.030	0.466	0.470	−0.061	0.804	1.000
PRR	0.103	0.356	0.185	−0.003	0.219	0.584	1.000

and

Table A3. The correlation matrix of the variables in the developing economies.

Variable	NBD	Kaopen	PC	TRA	LE	GDPP	PRR
NBD	1.000
Kaopen	0.369	1.000
PC	0.406	0.198	1.000
TRA	0.273	0.372	0.410	1.000
LE	0.152	0.336	0.483	0.365	1.000
GDPP	0.387	0.440	0.479	0.345	0.658	1.000
PRR	0.388	0.310	0.389	0.366	0.501	0.573	1.000

.

4. Empirical Analysis

4.1. Effect of Capital Account Liberalization on Entrepreneurship: The Baseline Model

The effect of overall capital account liberalization on entrepreneurship is shown in

Table 4. The impact of capital account liberalization on entrepreneurship.

Variable	Full Sample				Developed Economies	Developing Economies
Lagged NBD	0.830 ***	0.791 ***	0.772 ***	0.762 ***	0.846 ***	1.093 ***
	(560.7)	(454.8)	(389.8)	(286.0)	(28.90)	(393.4)
Kaopen	0.867 ***	0.780 ***	0.556 ***	0.556 ***	0.425 ***	−0.0563 ***
	(128.7)	(86.17)	(49.11)	(31.35)	(3.236)	(−3.984)
PC	0.00497 ***	0.00663 ***	0.00530 ***	0.00578 ***	0.00618 **	−0.00427 ***
	(29.50)	(20.68)	(14.61)	(13.71)	(1.977)	(−6.831)
LE	−0.851 ***	−1.127 ***	−10.16 ***	−10.05 ***	−5.307	2.763 ***
	(−4.226)	(−5.171)	(−28.38)	(−29.01)	(−0.404)	(7.757)
TRA		0.00727 ***	0.0102 ***	0.0101 ***	0.00231	0.00222 ***
		(44.44)	(45.58)	(46.33)	(0.997)	(6.016)
GDPP			0.961 ***	0.489 ***	−1.852 *	−0.124 ***
			(26.61)	(11.08)	(−1.729)	(−4.492)
PRR				0.0648 ***	0.0963 ***	0.0125 ***
				(23.18)	(2.592)	(6.190)
Constant	2.632 ***	4.028 ***	33.22 ***	33.13 ***	32.92	−11.54 ***
	(3.119)	(4.439)	(26.77)	(25.72)	(0.628)	(−8.439)
AR (1) Testing p value	0.008	0.008	0.010	0.011	0.028	0.006
AR (2) Testing p value	0.214	0.221	0.240	0.247	0.232	0.170
Sargan Testing Statistic	90.362	92.041	92.691	90.940	23.532	57.367
Sargan Test p value	0.861	0.831	0.818	0.851	1.000	0.708
Time fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Country fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Number of countries	103	103	103	103	32	71

Note: The values in parentheses are the p-values. ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

. In our baseline model, we include 33 developed economies and 72 developing and emerging economies. The detailed samples of countries and their classifications are displayed in the Appendix Table A1. First of all, we control both the time fixed effect and unobserved country-specific effect to exclude factors that may change over time and over the country so as to make our empirical results more reliable. Second, for all the baseline regressions, some of the first-order auto-correlation tests (AR (1) testing) are significant at the 1% level, but the second-order auto-correlation tests (AR (2) testing) are not significant, indicating that there is no second-order auto-correlation for the model residual. The p value of Sargen testing equals 1; thus, there is no overidentification problem, and the chosen instrument variables are effective. All the tests indicate that system GMM is an appropriate econometric method to construct in this paper.

For the full sample consisting of both the developed and developing economies, the results of the regressions all show that both capital account liberalization and financial development (measured by domestic credit to the private sector (% of GDP)) have a positive effect on entrepreneurship, as the corresponding coefficients are positive and significant at the 1% level. This implies that capital account liberalization and financial development could improve entrepreneurship from a global perspective. However, when we group our sample into developed and developing economies and conduct the corresponding regressions, different grouping empirical results occur. We find a positive relationship between capital account liberalization and entrepreneurship in developed economies and a negative relationship in developing economies, which indicates that in developing economies, on the contrary, capital account liberalization hinders the development of domestic entrepreneurship. Such inconsistencies between developed and developing economies are probably due to the incomplete financial system, low level of financial development and unreasonable capital controls in developing countries. Therefore, Hypothesis 1 is confirmed, and the full sample regression probably ignores this heterogeneity effect between developed and developing economies. Our baseline regression results are consistent with research conducted by Gregory [37].

Last but not least, the coefficient signs of the control variables in the baseline regression are basically consistent with other empirical studies. For the full sample, although a nation’s average life expectancy has a negative impact on entrepreneurship, trade liberalization and economic development impose a positive effect on entrepreneurship. Further, if a country experiences a low political risk (high political risk rating), it usually has stronger entrepreneurship. However, for the subsample of developing economies, the effect of life expectancy is positive and significant. It is perhaps due to the inefficient and inadequate social insurance system in developing economies, which will force people to find sources of income with aging. We also add more control variables in the subsequent robustness checks, which also implies that our main conclusion is not affected by the change in the control variables.

4.2. Mechanism Verification and Heterogeneity Analysis: The Role of Financial Development

As is discussed in the section of model specification, the marginal effect of capital account liberalization is probably not constant and non-linear; thus, the effect of capital account liberalization on entrepreneurship could be influenced possibly by the degree of domestic financial development. Thus, we include the interaction term of capital account liberalization and financial development to look deeper into the moderating effect of financial development.

Beck and Levine [43] propose that financial development is relatively a holistic concept, reflecting not only the improvement of the overall financial system but also the progress of the entire financial industry. That is to say, financial development is not merely a quantitative increase but also a qualitative change, namely, the systematic improvement and refinement of the financial institutions, the financial markets and financial efficiency. Inspired by top researches in the financial development field, instead of just using one single proxy variable, this paper reflects the connotation of financial development through a wide range of indicators. We first investigate the influencing mechanisms of credit markets and equity markets development. Further, it is necessary to include the degree of domestic financial deepening and the level of a country’s financial inclusion into the research framework of financial development.

4.2.1. Mechanism Verification and Heterogeneity Analysis: The Role of Credit and Equity Markets Development

Table 5. The impact of capital account liberalization on entrepreneurship: The role of credit market development.

Variable	Full Sample	Developed Economies	Developing Economies
Kaopen * PC	0.000517 ***	0.0106 **	−0.00613 ***
	(2.666)	(2.482)	(−11.96)
AR (1) Testing p value	0.011	0.033	0.016
AR (2) Testing p value	0.250	0.288	0.170
Control variables	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes
Country fixed effect	Yes	Yes	Yes
Number of countries	103	32	71

Note: The values in parentheses are the p-values. ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

and

Table 6. The impact of overall capital account liberalization on entrepreneurship: The role of equity market development.

Variable	Full Sample	Developed Economies	Developing Economies
Kaopen * SMC	−0.00566 ***	−0.0148	−0.00255 ***
	(−23.97)	(−1.098)	(−8.136)
AR (1) Testing p value	0.012	0.053	0.028
AR (2) Testing p value	0.136	0.227	0.967
Control variables	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes
Country fixed effect	Yes	Yes	Yes
Number of countries	65	25	40

Note: The values in parentheses are the p-values. ***, **, and * indicate significance at 1%, 5% and 10% levels, respectively.

present the main results of the situation of credit and equity market development for the full sample and the heterogeneous results for developed and developing economies correspondingly.

Specifically, for the full sample in Table 5, the interaction term between capital account liberalization and credit markets development (private credit markets) is positive, suggesting that the positive spillover effect of capital account liberalization increases with the development of the overall domestic credit market. For developed countries, the coefficient of the interaction term is also positive, so the previous conclusion is also applicable to developed countries, meaning that the impact of capital account liberalization in developed countries is more positive at high levels of financial development. For developing countries, the main effect of capital account liberalization in the baseline regression is negative, and the interaction term is also negative after the interaction term is added in the model, which indicates that the negative spillover effect of capital account liberalization in developing countries is relieved due to the development of the domestic credit market. The above empirical results are consistent with our research Hypothesis 2.

As for Table 6, the situation for equity market development is slightly different from that of the credit market. The coefficient of the interaction term for the full sample is negative, and that of developed countries is not significant, but the interaction term coefficient of developing countries is negative. This implies that the development of domestic equity markets is more beneficial to developing countries because the negative effect of capital account liberalization in developing countries has been weakened by the development of the equity market.

4.2.2. Mechanism Verification and Heterogeneity Analysis: The Role of Financial Deepening and Financial Inclusion

Table 7. The impact of overall capital account liberalization on entrepreneurship: The role of financial deepening.

Variable	Full Sample	Developed Economies	Developing Economies
Kaopen * M2	0.00153 ***	−0.0164	−0.00135 **
	(2.711)	(−0.396)	(−2.414)
AR (1) Testing p value	0.023	0.180	0.017
AR (2) Testing p value	0.706	0.704	0.279
Control variables	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes
Country fixed effect	Yes	Yes	Yes
Number of countries	81	14	67

Note: The values in parentheses are the p-values. ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

and

Table 8. The impact of overall capital account liberalization on entrepreneurship: The role of financial inclusion.

Variable	Full Sample	Developed Economies	Developing Economies
Kaopen * FI	0.223 ***	−0.0135	−0.149 ***
	(12.51)	(−0.0365)	(−9.278)
AR (1) Testing p value	0.006	0.030	0.008
AR (2) Testing p value	0.226	0.242	0.223
Control variables	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes
Country fixed effect	Yes	Yes	Yes
Number of countries	99	32	67

Note: The values in parentheses are the p-values. ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

show the outcomes of the cases of financial deepening and financial inclusion. The empirical results of financial deepening and financial inclusion are relatively similar. The coefficient of the interaction term for the full sample is positive, and the interaction term coefficient of developed countries is not significant, but the interaction term coefficient of developing countries is negative. This shows that financial deepening and financial inclusion will be more beneficial to developing countries. The negative effect of capital account liberalization in developing countries has been weakened by financial deepening and the continuous development of financial inclusion. The above empirical results are also consistent with our research hypothesis.

5. Robustness Checks and Endogeneity Issues

In order to verify whether a series of conclusions in this paper are reliable, this paper conducts a series of robustness tests from the following aspects: changing the estimation methods, choosing different proxy variables and adding more control variables.

5.1. Robustness Checks

5.1.1. Different Estimation Methods

In the previous section, we use the system GMM method to alleviate the endogeneity issues of a dynamic panel. For comparison, in this section, we use the static model and the related estimation strategies to re-examine the relationship between capital account liberalization and entrepreneurship. These estimation strategies include OLS, fixed effect (FE), random effect (RE) and pooled effect (PE) methods.

5.1.2. Different Proxy Variables

Due to the limitations of the proxies that we use in the baseline regression, in this section, we exploit different proxies for the core explanatory and explained variables. First, the EWN index created by Lane and Milesi-Ferretti [75] is often used in the literature to measure the level of capital account openness by country, so we choose the EWN index to undergo a robust test. Second, we take the number of inventions and patents as the proxy for entrepreneurship as it is a reasonable measure for entrepreneurs’ innovative spirit.

Further, we use the comprehensive index proposed by Svirydzenka [76] as the other proxy variable of financial development. He constructs the overall financial development index from two aspects: financial institutions and financial markets. At the same time, he measures the quantitative value of each first-level sub-index from three aspects: Depth, access, and efficiency. Svirydzenka [76] then uses a series of original variables to calculate all the six second-level sub-indexes. On this basis, after having these six indexes, the author could calculate two first-level sub-indexes, which represent the development of the financial institution and financial market. Lastly, the comprehensive financial development index is calculated to indicate the overall financial development.

Last but not least, as it is still controversial to specify an appropriate proxy for entrepreneurship, a more convincing representative measure of entrepreneurship at the macro-level is probably the aggregated entrepreneurial orientation (EO). Following Mthanti and Ojah [77], we use the principal component analysis (PCA) method to construct the EO at the aggregate level. Mthanti and Ojah [77] compute EO from three dimensions that covary: innovativeness, proactiveness and risk-taking. The first-level sub-index of innovativeness is calculated by three second-level indicators: a country’s innovative input, scientific output and technological output. The sub-index of proactiveness is measured comprehensively by foreign direct investment (FDI), exports of goods and services (%GDP), royalty and license fees and the number of internet users. Finally, non-agricultural value-added (%GDP), gross investment to GDP ratio and the domestic savings rate are employed to calculate the level of a country’s risk-taking.

5.1.3. Adding More Control Variables

As the entrepreneurship of a country at the macro level is probably affected by other important driving factors, we include more control variables to test whether the main results of the paper will be influenced by changing the control variables. We subsequently control the human capital (refers to the secondary school enrollment rate) and the gross fixed capital formation (% of GDP) in our model.

5.1.4. Taking Consideration of the Financial Crisis

Due to the relatively short time span of our study, it is quite difficult to run a truly longitudinal analysis. Therefore, following the existing literature, we add the financial crisis dummy variable into our model to indicate the frequency of a banking or currency crisis.

After the above robustness test, it is found that the main conclusions of this paper are still valid and robust.

5.2. Endogeneity Issues

In order to deal with the possible endogeneity issues, we try to tackle this problem from the following econometrical techniques: first of all, we construct a dynamic panel model rather than a static one in model setting and control the lagged terms of entrepreneurship. Compared to the traditional cross-sectional or time-series model, the dynamic panel model can improve the validity of econometric models and alleviate endogeneity issues. Further, we use the system GMM method developed by Blundell and Bond [63] to effectively deal with the endogeneity problem. In the section of empirical analysis, we can see that all the regressions have passed the Sargan test, so there is no problem of overidentification, and the chosen instrument variables are effective. Last but not least, we also add the interaction term between capital account liberalization and financial development, which ensures that the problem of omitted variable bias is relieved.

6. Conclusions and Recommendations

Entrepreneurship has been deemed crucial to aggregate job creation, economic development, productivity growth and social stability. No government can afford to ignore the prominent role of entrepreneurship when formulating national policies. Meanwhile, the government should take a comprehensive look at the likely determinants of entrepreneurship.

International experience also shows that in the process of economic reform and opening up, capital account liberalization is probably the most difficult, most controversial and most risky reform in the emerging market countries. Although some countries enjoy the dramatic economic growth effects of capital account liberalization, there are still many developing countries experiencing great volatility in financial markets, currency crisis and a series of great recessions after opening a capital account. Therefore, developing economies that wish to promote entrepreneurship should revisit the decisive prerequisites for capital account liberalization.

By constructing an econometric model, we use the system generalized method of moments (GMM) method and a rich cross-country dataset of 103 countries in 2007–2018 to investigate empirically the possible economic influence of capital account liberalization on entrepreneurship. The results show that the impact of capital account liberalization on entrepreneurship in developed economies is positive, while in contrast, that of the developing economies is negative. Furthermore, domestic financial development plays an important moderating role in the relationship between capital account openness and entrepreneurship. Specifically, the negative impact of capital account liberalization in developing economies is mitigated by a high degree of domestic private credit and equity market development, the continuous deepening of finance, and better financial inclusion of a country.

Therefore, this study provides far-reaching institutional inspiration and policy suggestions for developing countries to liberalize their capital account and cultivate national entrepreneurship. The profound practical implication for government policy makers is that when the aim of liberalizing a country’s capital account to strengthen domestic entrepreneurship, one of the central institutional contexts cannot be ignored, which is the development of domestic financial markets and financial systems. That is to say, domestic financial development is an essential prerequisite for capital account liberalization, especially for developing countries.

Policies or policy combinations that our research can provide for developing countries include the followings: developing domestic credit markets (particularly private credit markets) and equity markets to minimize the financing constraints faced by entrepreneurs; continuing to deepen the reform of domestic financial markets and improving the degree of domestic financial inclusion, so that more individuals and enterprises could have access to a wide range of financial support to start a business and make innovation; finally, gradually liberalizing capital account and carefully reducing restrictions on international capital flows, which further provide more financing or investment opportunities for entrepreneurs.

However, this study is not without limitations, and future work may explore the following issues. First of all, when conducting cross-country research, researchers have limited access to cross-country data sources, and many cross-country data lack a unified measurement standard. Because the WDI database currently offers relatively large samples of cross-country data, so our research could not obtain more sample countries from other channels or sources except for the WDI database. In the future, with the increasing access to international data, we hope that we can continue to increase the number of countries in the sample and make a truly longitudinal analysis in order to make our research conclusions more reliable. Second, our paper is limited to the perspective of domestic financial development when discussing the preconditions of capital account liberalization. In fact, in addition to financial development, there may be many other prerequisites for a country to open its capital account to cultivate its entrepreneurship, such as a stable business environment and fair legal system, etc. Future studies could try to explore the possible moderating effects from the above factors. Third, the impact of global taxation agreements, treaties and policies on capital liberalization and the role of international corporations could also be an interesting line of inquiry for further study stemming from this research.