How Does the Sustainable Investment Climate Affect Firm Geographic Diversification in China? Managerial Discretion as a Mediator

Abstract

Although existing research has explored the main effects of the sustainable investment climate, including political and legal environments, on the performance of firms’ geographic diversification, most studies neglect the role of strategic leadership, especially the CEO. The concept of managerial discretion provides a theoretical lens to fill this void. Using a large database from the Investment Climate Survey conducted by The World Bank, which covers 12,400 firms from 120 cities among China’s 30 provinces, we empirically prove that the investment climate, including governmental, legal, and financial environments, significantly influence the degree of CEO managerial discretion. Moreover, CEO managerial discretion is positively associated with firm geographic diversification and mediates the relationship between the investment climate and firm geographic diversification. Specifically, the greater the CEO managerial discretion, the lower degree of firm investment within the city and the province in which it is headquartered and the higher level of the firm’s internationalization. Implications for firms’ geographic diversification and constructing a sustainable investment climate for emerging markets are finally given.

1. Introduction

The investment climate is the institutional, policy, and regulatory environment in which firms operate—factors that influence the link from sowing to reaping [1]. Sustainable investment climate becomes more and more important for firms to win competitive advantages and achieve sustainable development [2]. In the past two decades, The World Bank has promoted improving business environments as a key strategy for development, which has led to a significant effort in collecting surveys of the investment climate at the firm level across countries [3]. An essential measure is launching a Doing Business Report on measuring the regulations that enhance business activity and those that constrain it [4]. According to The World Bank’s definition in the Doing Business Report 2020, Doing Business is founded on the principle that economic activity benefits from clear rules that cover 12 areas of regulations and allow voluntary exchanges between economic actors, set out strong property rights, facilitate the resolution of commercial disputes, and provide contractual partners with protections against arbitrariness and abuse. Such rules are much more effective in promoting growth and development when they are efficient, transparent, and accessible to those for whom they are intended. These rules will create a sustainable business environment where new entrants with drive and innovative ideas can get started in business and where productive firms can invest, expand, and create new jobs. For example, in the daily operations of small and medium-size domestic firms, the role of government policy is a central focus of the Doing Business data, because businesses can thrive only if regulation is efficient, transparent, and easy to implement. As shown in Figure 1, China’s business climate has made great progress from the year of 2008 to 2020.

However, the sustainable degree of business climate among provinces in China has great differences. In turn, subnational institutional environments have attracted more and more attention in recent studies [5,6,7,8]. Without a sustainable business climate, firms tend to expand or even transfer their operating geographical areas to alleviate risk from the region in which it is headquartered. For instance, focusing on the most important risk in banking: credit risk, scholars find that allowing for geographical diversification could reduce banks’ credit risk by 1.1% on average, with risk reduction ranging from negligible up to 8% [9].

Existing research on firm geographic diversification mainly focus on its consequences, such as performance [10,11,12], risk [13,14,15], bank lending [16], or even the relationship between production and geographic diversification [17,18,19], while few of them examines its antecedents. Since the institutional-based view was proposed to understand specific organizational strategic choices including geographic diversification in emerging markets [20,21,22], a growing body of literature has discussed the direct impact of institutions, especially subnational institutions, on firm strategies, including geographic diversification [5,23,24,25]. However, most studies have neglected the significance of strategic leadership or upper echelons, especially a CEO, in the process of strategic decision-making [26,27], which leads to difficulties in clarifying the mechanism of how external institutions affect firm strategic choices.

How strategic leaders influence their organizations is an important question that has assumed greater importance in the search for factors that contribute to firms’ sustainable competitive advantage. Do CEOs matter? To debate with doubtful views of institution theory [28] and population ecology theory [29], the concept of managerial discretion, meaning the latitude of action, was introduced [30] and confirmed as one of the most important moderators in strategic leadership research [31,32].

It is widely recognized that the institutional environment can shape the roles played by the strategic leaders [33], and different institutional arrangements may also affect how company leaders shape company strategies and performance [34,35]. Specifically, managerial discretion could also act as a mediator between national institutions and firm performance through both theoretical derivation and empirical study.

Until now, the research on strategic leadership as a whole has been deeply rooted in the United States, its economy, corporate evolution, and major dilemmas, as decades of strategic leadership research has employed the United States as the source of theories (e.g., upper echelon theory, agency theory, and resource dependence perspective), and the dominant empirical setting. Recent studies have offered evidence that in China, strategic leaders such as board chairs have a greater impact in shaping firm strategies and performance than their counterparts in the United States [36]. Given that both formal and informal institutions in China are fundamentally different from those in the United States, the question becomes whether theories on strategic leadership developed in the United States will apply in China. Specifically, does the mediating effect of managerial discretion exist in China? Furthermore, if yes, besides static outcome such as firm performance, will it mediate the effects of the institutional environment on the firm dynamic strategic process rather than static outcome such as performance?

To answer the above questions, this paper aims to examine the mediating effect of managerial discretion on the relationship between investment climate and firm geographic diversification in the Chinese context. This paper contributes to the existing literature and practice in three aspects as follows. Firstly, we develop the Institution-Based View by examining the effects of investment climate on firm geographic diversification and proving the determinant role of institutional environments in firm sustainable development in a transitional context. Secondly, this paper contributes to the Upper Echelons Theory by empirically confirming the mediating role of managerial discretion between investment climate and firm geographic diversification, which bridges the macro and micro domains. Thirdly, we identify the regional boundary of Chinese institutional disparity at the province level rather than the city level by exploring the institutional antecedents of firm geographic diversification, which provides new inspiration to further academic research and policy reformation in the Chinese context.

The remainder of the research is organized as follows. The study hypotheses regarding the link among investment climate, CEO managerial discretion, and firm geographical diversification are developed in Section 2, research techniques including data and empirical methodology are discussed in Section 3, the empirical findings are summarized and analyzed in Section 4, the robust check is conducted in Section 5, and lastly, conclusions and suggestions are presented in Section 6.

2. Theory and Hypotheses

2.1. Sustainable Investment Climate and CEO Managerial Discretion

2.1.1. Institution-Based View and Firm Geographic Diversification

Institution-Based View (IBV) treat formal and informal institutions as the causes of diversification strategy [20]. Given that both formal and informal institutions in China are fundamentally different from those in the United States, it is necessary to clarify what kinds of subnational institutions drive or constrain Chinese firms’ geographic diversification. Regarding the antecedents of geographical diversification, prior studies in the United States context emphasize more on resources, transition costs, and agency factors, while their Chinese counterparts pay more attention to asset portfolio, governmental policies, and institutional factors.

Then, what kind of institutional factors can shape firm geographic diversification in China? Previous studies have shown that firm diversification depends on the development of the formal and informal financial system and the efficiency of resource allocation [37]. Additionally, government intervention accelerates the process of diversification of listed firms that are directly controlled by governments. Moreover, legal institutions and property rights protection rules significantly affect regional exports. Higher-quality legal systems in an exporter’s location are associated with higher export performance [38]. Additionally, both national policy and capital market strongly shape listed firms’ diversification. Specifically, in terms of the administrative level, companies held by central, or country and township governments show a smaller degree of diversification, while firms owned by provincial, or city-level governments experience a higher degree of diversification. In terms of geographical location, coastal firms diversify more than inland ones do.

Comprehensively, some elements of the business environment, for instance, the “grabbing hands” of the government, seem to matter a great deal for most economies. Other elements, such as infrastructure and contracting institutions (that is, courts and access to finance), hinge on their initial status and the size of the market [3]. Using comparable enterprise surveys that cover more than 100,000 firms in 123 countries enterprise surveys data in developing countries, the literature provides evidence that a good business climate including regulation, security, and finance variables favors growth by encouraging investment [39]. In conclusion, we can confirm that government intervention, legal environment, and financial development have affected firm diversification in the Chinese context.

2.1.2. Institutional Environment and Managerial Discretion

The Institution-Based view emphasizes external institutional environments, but does not pay adequate attention to the role of strategic leadership. In this regard, it is important to combine institutional environments with strategic leadership to study the antecedents of firm geographic diversification.

Early research takes governmental regulation as an important constraint of CEO managerial discretion [30,40,41]. The following studies treat the degree of government regulation as a key component of managerial discretion [42]. Recently, researchers began exploring the systematic differences in CEO managerial discretion at the national level. For example, the degree of a CEOs influence on firm performance is greater in U.S. firms than their counterparts in Germany or Japan. The degree of CEO managerial discretion in these countries is highly related to the legal tradition, firm ownership disparity, board governance, and cultural values [34]. Further studies show that the effects of certain informal and formal national institutions—individualism, tolerance of uncertainty, cultural looseness, dispersed firm ownership, employer flexibility, and common-law legal origin—are associated with the degree of CEO managerial discretion available to CEOs. Moreover, due to deregulation, CEO managerial discretion is greater in U.S. and U.K. firms than those in northern Europe and East Asia [35]. Moreover, with the ongoing deregulation in many countries, CEO managerial discretion has increased rapidly [43].

In China’s context, a set of studies have shown that government intervention, legal justice, and financial development are the top three kinds of environment for firms doing business. In turn, they will be associated with the degree of CEO managerial discretion of those firms located in the focal regions.

Therefore, we predict that:

Hypothesis 1 (H1).

The more sustainable investment climate including the lower level of governmental intervention (H1a), and the greater the level of local legal justice (H1b) and financial development (H1c) in a region, the higher the discretion available to CEOs of firms headquartered in that region.

2.2. Managerial Discretion and Firm Geographic Diversification

2.2.1. Upper Echelons Theory and Managerial Discretion

According to the Upper Echelons Theory, corporate executives have a great influence on firm decisions and performance, and firm decisions can be largely reflected by executives’ experiences, beliefs, and values [27]. Managerial discretion measures the latitude of actions available to managers. In other words, if managers have relative greater discretion in the process of strategic choices, firm decisions and outcomes would be reflected more closely by their characteristics. In contrast, if managerial discretion is constrained greatly, managers’ characteristics may not be reflected in firms’ decisions and strategic choices [31]. The degree of CEO managerial discretion will significantly influence a firm’s cross-region development.

2.2.2. Managerial Discretion and Firm Geographic Diversification

Most studies on the effects of strategic leadership on firm diversification usually focus on top managers’ demographical factors. For instance, executive education background, number of served firms, male executives, or those with a technological background, are positively related to the degree of firm diversification. For those top managers with financial backgrounds, the degree of diversification of their firms will decrease. Moreover, executive age has a U-shape relationship with the degree of diversification.

These personal factors are used as the measurement of CEO managerial discretion in early studies. Subsequent research combined these personal factors with firm-level ones. For example, economic and personal motivation can affect the listed firms’ choice of diversified business models. In addition, factors such as firm size, ownership structure, entry time, and even industry also significantly influence the decision and mode of firms’ diversification. Further exploration focused on firms’ internal and external governance structures. In the internal aspect, based on agency theory and resource dependence theory, the result shows that the board structure of small and medium-sized enterprises (SMEs) significantly affects firm diversification strategy. In the external aspect, under the condition of regional institutions disparity, firms located in a province that has greater CEO managerial discretion will have a higher degree of internationalization. Moreover, according to the definition of CEO managerial discretion in management [30], a great degree of latitude of action would help a CEO break through local constraints and also gain more resources from outside via geographic diversification.

Following the above logic, we propose that:

Hypothesis 2 (H2).

The greater CEO managerial discretion in a region, the higher degree of firm geographic diversification.

2.3. The Mediating Role of Managerial Discretion

CEO managerial discretion conventionally serves as a moderator in the existing research [31]. Recently, its mediating role has attracted increased attention [44]. For example, CEO managerial discretion of public firms in a country mediates the relationship between certain national informal and formal national institutions—individualism, tolerance of uncertainty, cultural looseness, dispersed firm ownership, a common-law legal origin, and employer flexibility—and firm performance [35]. Different from that, this paper aims to examine its mediating effect between subnational institutions and firm strategy in China.

We have argued that regional differences in the investment climate in China will be associated with differences in CEO managerial discretion in H1. In turn, we have argued that discretion will be associated with the higher degree of firm geographic diversification in H2. Logically, this implies that discretion occupies a mediating role between investment climate and geographic diversification. Contextual constraints, arising from government, legal, and marketing environments, will affect CEOs’ latitudes of actions, which in turn will affect firm geographic diversification. Thus:

Hypothesis 3 (H3).

CEO managerial discretion mediates the relationship between sustainable investment climate and firm geographic diversification.

3. Method

3.1. Sample

Our sample derives from the World Bank Investment Climate Survey. The dataset is the newest available one for this research and is widely used by studies published in recognized academic journals [38,45,46,47,48]. The survey covers 12,400 firms among 120 cities across 30 provinces in mainland China, except for Tibet, for the lack of sufficient time, resources, and enterprises meeting survey criteria to include Tibet. Cities especially capital cities are included from all 30 provinces. The inclusion of additional cities for a particular province depends on provincial GDP. Among the 120 cities, only four huge cities, Beijing, Shanghai, Tianjin, and Chongqing, contain 200 firms, and other 116 cities contain 100 firms each. The 120 cities included in this survey account for 70–80 percent of China’s GDP.

All firms are from 21 manufacturing industries (

Table 1. Industrial distribution of survey sample.

Code	Industry	#	%
13	Farm and Side-line Food Processing	969	7.81
14	Food Manufacturing	243	1.96
15	Beverage Manufacturing	178	1.44
16	Tobacco Product	46	0.37
17	Textile Manufacturing	952	7.68
18	Clothing, Shoes and Hats Manufacturing	206	1.66
19	Leather Furs Down and Related Products	139	1.12
20	Timber Processing Bamboo Cane Palm Fiber and Straw Products	141	1.14
21	Furniture manufacturing	55	0.44
22	Papermaking and Paper Products	235	1.90
23	Printing and Record Medium Reproduction	62	0.50
24	Cultural Educational and Sports Goods	41	0.33
25	Petroleum Refining and Coking	182	1.47
26	Raw Chemical Materials and Chemical Products	1441	11.62
27	Medical and Pharmaceutical Products	426	3.44
28	Chemical Fiber	47	0.38
29	Rubber Products	21	0.17
30	Plastic Products	329	2.65
31	Nonmetal Mineral Products	1299	10.48
32	Smelting and Pressing of Ferrous Metals	491	3.96
33	Smelting and Pressing of Nonferrous Metals	345	2.78
34	Metal Products	366	2.95
35	Ordinary Machinery	1077	8.69
36	Special Purpose Equipment	486	3.92
37	Transport Equipment	989	7.98
39	Electric Equipment and Machinery	864	6.97
40	Electronic and Telecommunication Equipment	598	4.82
41	Instruments Meters Cultural and Clerical Machinery	60	0.48
42	Handicrafts and other Manufacturing	109	0.88
43	Recycled Material Manufacturing	3	0.02
	Total	12,400	100

Source: World Bank [47].

) without service industries, since some services (e.g., financial services) are prone to greater regulation, and the inclusion of higher numbers of such service businesses in some cities (e.g., Beijing, Shenzhen, Shanghai) could distort survey results. For each city, the top 10 industries in terms of sales revenue are drawn. The process of sample selection by the World Bank appropriately considered firms’ industrial distribution and, in turn, reduced the potential influence of the industry to the most extent. Regarding the size of these firms, large, medium, and small sizes are equally distributed within each industry, and each accounting for 1/3 of total industry revenue. Then, from each of the three types of firms, an equal number of firms are drawn. As for the ownership structure of these firms, 8%are registered as majority state-owned; 28% as foreign invested; and 64% as domestic non-state. Each firm has at least 10 employees. We delete some missing or abnormal observations and 12,301 firms left in the sample.

3.2. Measurement

The World Bank asked each surveyed firm questions on different aspects of investment climate, and then aggregated firm-level responses into city-level indexes for those firms located in the same city, using the different standardized methods so that they can be compared.

Table 2. Measurement of variables.

Variables	Definitions
Independent Variables
gov_int	Governmental intervention. The survey asks how many days that firms have to interact with four government departments including Taxation, Public Security, Environment, and Labor and Social each year in total.
leg_jus	Legal justice. In commercial or other legal disputes, the percent of cases the company’s legal contracts or properties were protected.
fin_dev	Financial development. If your company has loans from banks or other financial institutions, the collateral value of a recent overdraft or loan is what percentage of the total loan?
Mediating Variables
ceo_dis	Managerial discretion. How much discretion does the CEO have over the production, investment, and employment, respectively (i.e., not intervened by the government, 1 = 0–19%, 2 = 20–39%, 3 = 40–59%, 4 = 60–69%, 5 = 70–79%, 6 = 80–89%, 7 = 90–99%, 8 = 100%). We counted the average of three aspects as CEO managerial discretion.
Dependent Variables
city	Percentage of sales within the city in 2004.
inprv	Percentage of sales within the province (autonomous region, central municipality) in 2004.
inter	Percentage of overseas (including HK, Macau, Taiwan) sales in 2004.
Control Variables
tenure	CEO tenure. 2005 minus the number of years the current CEO has held the position.
pol_app	CEO political connection. Is the CEO appointed by the government? 1 = no, 2 = yes.
duality	CEO duality. Is CEO also the president of the Board of Directors? 1 = no, 2 = yes. Firms without board will also be coded as 1.
sal_gap	Salary gap. How many times more is the CEO’s annual income than that of the mid-level managers (annual income includes salary and bonus. Middle managers, i.e., department or branch manager)? (1) < 2 times, (2) 2–3 times, (3) 3–4 times, (4) 4–6 times, (5) >6 times.
firm_age	Firm age. 2004 minus firms’ established year.
state_own	State ownership. Percentage owned by state among the firm’s total assets.
for_own	Foreign ownership. Percentage owned by foreigners among the firm’s total assets.
non_soe	Private ownership. Percentage owned by private among the firm’s total assets.
lnsales	Firm size. Natural logarithm of the average of core business income during 2002–2004.
per_gdp	City economical scale. GDP per capita from 2002–2004.

shows the measurement of all variables.

3.2.1. Independent Variable: Investment Climate

Government Intervention. The survey asks how many days that firms have to interact with four kinds of government departments including Taxation, Public Security, Environment, and Labor and Social each year in total. The firm-level index equals the respondent days divide by 365.

Legal Justice. In commercial or other legal disputes, the percent of cases, the company’s legal contracts, or properties were protected.

Financial Development. As loans are a key resource for Chinese firms, especially non-SOEs, we use the percentage of firms that received loans from banks in the city to measure the Financial Environment.

3.2.2. Mediating Variable: CEO Managerial Discretion

The data of CEO managerial discretion originates from the survey, which asks how much discretion the CEO has over the product, investment, and employment, respectively. The choices range from 1 to 8, which stands for CEO managerial discretion from low to high. The index is reported by the CEO of the firm. We counted the average of three aspects as CEO managerial discretion.

3.2.3. Dependent Variable: Geographic Diversification

The survey asks the CEO the percentage of company sales in corresponding regions in 2004. The regional scale contains: (1) Within the city; (2) Within the province (autonomous region, central municipality, excluding the sales within the city); (3) Out of the province; (4) Overseas (including HK, Macau, Taiwan). We take the percentage of company sales on the scale of (1) as the firm development level in the city, the sum of (1) and (2) as the development degree in the province, and (4) as the degree of firm internationalization.

3.2.4. Control Variables

Previous systematic reviews show that CEO managerial discretion does not occur by happenstance, rather it is derived from three sets of factors, including task environment, internal organization, and managerial characteristics [26,44]. This paper aims at examining the effects of institutional environments, so individual-, firm- and city-level factors can be controlled. Considering the availability of data, we control individual factors of strategic leaders such as tenure, whether political appointment, duality, and salary gap, which are reported by company managers. Five firm-level control variables indexes cover firm age, three types of firm ownership structure, and size, which are reported by company accountants. At the city level, the city economical scale provided by The World Bank is controlled.

4. Results

4.1. Correlation Analysis

Table 3. Correlation of IV, DV, and MV.

VAR	Obs.	Mean	Std. Dev.	Min	Max	1	2	3	4	5	6
1. city	12,300	23.17	31.94	0	100
2. inprv	12,300	44.05	37.31	0	100	0.748 ***
3. inter	12,300	16.49	31.58	0	100	−0.325 ***	−0.500 ***
4. ceo_dis	12,223	7.194	1.523	1	8	−0.046 ***	−0.042 ***	0.043 ***
5. gov_int	12,301	0.042	0.014	0.006	0.089	0.076 ***	0.044 ***	−0.093 ***	−0.129 ***
6. leg_jus	12,301	0.635	0.167	0.269	0.982	−0.154 ***	−0.092 ***	0.131 ***	0.140 ***	−0.381 ***
7. fin_dev	12,301	0.600	0.141	0.19	0.92	−0.128 ***	−0.071 ***	0.023 **	0.040 ***	0.034 ***	0.349 ***

Note: *** p < 0.01, ** p < 0.05.

shows the Pearson correlation of investment climates, CEO managerial discretion, and firm geographic diversification. The degree of government intervention is negatively related to CEO managerial discretion (r = −0.129, p < 0.01), while the value of both legal justice (r = 0.140, p < 0.01), and financial development (r = 0.040, p < 0.01) are positively correlated with CEO discretion. Based on this, H1 is pre-proved. In the aspect of CEO managerial discretion and firm geographic diversification, the result shows that CEO discretion is significantly negative correlated with firms’ development in the local city (r = −0.046, p < 0.01) and province (r = −0.042, p < 0.01), while significantly positive with the degree of firm internationalization (r = 0.043, p < 0.01). Therefore, H2 also receives primary verify.

4.2. Regression Analysis

The ordinary least squares (OLS) model is used to test H1 and H2. Following the method by the previous studies on nation-level institutional indexes and managerial discretion [35], we first add each institution index into each regression model, after controlling multi-layer variables including individual-, firm- and city-level factors. The results are shown in

Table 4. Regression analysis.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
Variables	ceo_dis	ceo_dis	ceo_dis	ceo_dis	city	inprv	inter	city	inprv	inter
gov_int	−10.30 ***			−6.593 ***				121.8 ***	103.2 ***	−175.3 ***
	(1.137)			(1.253)				(24.14)	(29.37)	(24.72)
leg_jus		1.006 ***		0.792 ***				−8.245 ***	1.544	8.102 ***
		(0.0969)		(0.113)				(2.188)	(2.662)	(2.240)
fin_dev			0.273 **	−0.0383				−10.25 ***	−4.527	3.118
			(0.116)	(0.124)				(2.382)	(2.899)	(2.440)
ceo_dis					−0.999 ***	−1.153 ***	0.784 ***	−0.761 ***	−1.067 ***	0.515 **
					(0.208)	(0.251)	(0.213)	(0.208)	(0.253)	(0.213)
tenure	0.011 ***	0.009 **	0.012 ***	0.009 **	−0.240 ***	−0.244 ***	0.332 ***	−0.183 ***	−0.228 ***	0.283 ***
	(0.004)	(0.004)	(0.004)	(0.004)	(0.068)	(0.082)	(0.070)	(0.068)	(0.083)	(0.069)
pol_app	−0.300 ***	−0.316 ***	−0.332 ***	−0.299 ***	−0.405	1.833	−1.232	0.827	1.569	0.649
	(0.064)	(0.064)	(0.064)	(0.064)	(1.236)	(1.496)	(1.266)	(1.230)	(1.497)	(1.260)
duality	0.165 ***	0.166 ***	0.169 ***	0.164 ***	−0.371	−0.917	2.262 ***	0.263	0.866	2.168 ***
	(0.035)	(0.035)	(0.035)	(0.035)	(0.671)	(0.812)	(0.687)	(0.667)	(0.812)	(0.683)
sal_gap	0.005	0.010	0.011	0.006	−1.413 ***	−1.735 ***	0.595 **	−1.290 ***	−1.657 ***	0.456 *
	(0.013)	(0.013)	(0.013)	(0.013)	(0.258)	(0.312)	(0.264)	(0.257)	(0.313)	(0.263)
firm_age	0.0003	0.0003	0.0003	0.0001	0.020	0.077 **	−0.056 **	0.015	0.072 **	−0.047 *
	(0.001)	(0.001)	(0.001)	(0.001)	(0.027)	(0.032)	(0.027)	(0.027)	(0.032)	(0.027)
state_own	−0.002 ***	−0.002 ***	−0.002 ***	−0.002 **	0.019	0.012	−0.033 **	0.009	0.008	−0.025 *
	(0.001)	(0.001)	(0.001)	(0.001)	(0.014)	(0.017)	(0.014)	(0.014)	(0.017)	(0.014)
non_soe	0.003 ***	0.003 ***	0.003 ***	0.003 ***	−0.031 ***	−0.006	0.022 **	−0.023 ***	0.003	0.016 *
	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.009)	(0.010)	(0.009)	(0.009)	(0.010)	(0.009)
for_own	0.002 ***	0.002 ***	0.002 ***	0.001 **	−0.113 ***	−0.214 ***	0.404 ***	−0.108 ***	−0.213 ***	0.394 ***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.011)	(0.014)	(0.012)	(0.011)	(0.014)	(0.012)
lnsales	0.007	−0.020 **	0.015	0.016	−3.194 ***	−3.447 ***	0.667 ***	−2.985 ***	−3.421 ***	0.588 ***
	(0.010)	(0.010)	(0.010)	(0.010)	(0.187)	(0.226)	(0.191)	(0.189)	(0.230)	(0.194)
per_gdp	−0.049 *	−0.067 **	−0.087 ***	−0.048 *	3.818 ***	2.028 ***	5.301 ***	3.349 ***	1.739 ***	6.000 ***
	(0.027)	(0.027)	(0.027)	(0.027)	(0.512)	(0.620)	(0.525)	(0.515)	(0.626)	(0.527)
Constant	7.920 ***	7.163 ***	7.718 ***	7.383 ***	35.81 ***	77.18 ***	−53.70 ***	41.54 ***	75.99 ***	−56.08 ***
	(0.234)	(0.242)	(0.238)	(0.245)	(4.794)	(5.804)	(4.912)	(4.958)	(6.034)	(5.078)
Observations	8572	8572	8572	8572	8571	8571	8571	8571	8571	8571
F	31.06 ***	33.47 ***	23.90 ***	30.65 ***	55.96 ***	80.89 ***	221.95 ***	51.93 ***	64.75 ***	184.47 ***
R-squared	0.038	0.041	0.030	0.044	0.067	0.094	0.222	0.078	0.096	0.232

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

. We test H1 by Model 1, Model 2, and Model 3, and use Model 5, Model 6, and Model 7 to test H2.

4.2.1. Investment Climate and CEO Managerial Discretion

H1 proposes that a sustainable investment climate positively influences CEO managerial discretion. Model 1 shows that the impact of government intervention on CEO discretion is significantly negative (β = −10.30, p < 0.01), and the result proves H1a. Hypothesis 1b proposes that a higher degree of legal justice and financial development will increase CEO discretion of firms headquartered in the region. Model 2 and Model 3 show that the impact of legal justice (β = 1.006, p < 0.01) and financial development (β = 0.273, p < 0.01) on CEO managerial discretion are significant and positive, which strongly support H1b and H1c. Moreover, we also examined the impact of all institution indexes on CEO discretion in Model 4, and the effects of government intervention (β =−6.593, p < 0.01) and legal justice (β = 0.792, p < 0.01) correlate with our results one by one in Model 1 and Model 2, respectively. Therefore, H1 is fully supported.

4.2.2. CEO Managerial Discretion and Geographical Discretion

Hypothesis 2 predicts that CEO managerial discretion will lead to firm geographic diversification. As shown in Model 5 and Model 6, after controlling the potential impact of regional economic scales, the degree of CEO managerial discretion is still significantly and negatively related to the degree of firm development in the local city (β = −0.999, p < 0.01) and province (β = −1.153, p < 0.01), while significantly and positively correlated to the degree of firm internationalization (β = 0.784, p < 0.01) in Model 7. When we consider the effects of all institution variables and CEO discretion on each variable of firm geographic diversification in Model 8 (β = − 0.761, p < 0.01), Model 9 (β = −1.067, p < 0.01), and Model 10 (β = 0.515, p < 0.05), respectively, the results keep consistent with Model 5, Model 6, and Model 7. Accordingly, the empirical result provides empirical support to Hypothesis 2.

4.2.3. Mediating Effects of CEO Managerial Discretion

We test the mediating effects of CEO managerial discretion using the Sobel test method [49]. Considering the extremely large sample, there is no need to use the bootstrap method [50]. As we can see from the result in

Table 5. Sobel test: The mediating effects.

Indexes	city	inprv	inter
gov_int	10.525, 3.886, p < 0.01	12.271, 3.872, p < 0.01	−8.950, −3.38, p < 0.01
leg_jus	−0.663, −2.706, p < 0.01	−0.917, −3.159, p < 0.01	0.658, 2.705, p < 0.01
fin_dev	−0.366, −3.155, p < 0.01	−0.407, −3.069, p < 0.01	0.372, 3.183, p < 0.01

Note: Sobel β, Z-value, and p-value, respectively in each group.

, CEO managerial discretion mediates the nine pairs of the relationship between three kinds of investment climates (i.e., government intervention, legal justice, and financial development) and the three levels of firm geographic diversification. Therefore, Hypothesis 3 is confirmed.

5. Robustness Check

In this part, we use firm perceived institutional environments to replace the above city-level institutions to further test the robustness of the relationship of investment climate, managerial discretion, and geographic diversification. Moreover, distinct methods are applied to test the mediating effect of managerial discretion.

5.1. Measurement

Besides objective institutions used in the above studies, we apply the subjective institutional environment that is perceived by firms to test hypotheses, considering the argument that firm-perceived sustainable investment climate more accurately reflects the objective reality. Those who hold this argument believe that even for firms located in the same region, their perceived institutional environments may be different. For example, the survey asks firm CEOs about how long they spent dealing with officers, rather than inquiring as to what extent bureaucracy blocks firm development. Dependent variables and control variables are the same as those listed in Table 1.

5.2. Perceived Institutions and CEO Managerial Discretion

Table 6. Robust regression analysis.

	(11)	(12)	(13)	(14)	(15)	(16)	(17)
Variables	ceo_dis	ceo_dis	ceo_dis	ceo_dis	city	inprv	inter
gov_int1	−0.046 ***			−0.035 ***	−0.087	−0.434	0.712 ***
	(0.013)			(0.013)	(0.250)	(0.305)	(0.256)
leg_jus1		0.006 ***		0.005***	−0.020	0.002	0.016
		(0.001)		(0.001)	(0.013)	(0.016)	(0.014)
fin_dev1			0.039 ***	0.016	−0.441	−1.090 ***	1.440 ***
			(0.015)	(0.015)	(0.289)	(0.353)	(0.296)
ceo_dis					−0.829 ***	−1.001 ***	0.742 ***
					(0.218)	(0.266)	(0.223)
tenure	0.012 ***	0.011 ***	0.012 ***	0.011 ***	−0.236 ***	−0.252 ***	0.298 ***
	(0.004)	(0.004)	(0.004)	(0.004)	(0.070)	(0.085)	(0.071)
pol_app	−0.324 ***	−0.309 ***	−0.309 ***	−0.299 ***	−0.082	2.052	−0.432
	(0.064)	(0.065)	(0.064)	(0.065)	(1.252)	(1.528)	(1.282)
duality	0.178 ***	0.181 ***	0.166 ***	0.185 ***	0.155	−0.496	1.625 **
	(0.035)	(0.036)	(0.035)	(0.036)	(0.694)	(0.847)	(0.711)
sal_gap	0.014	0.009	0.010	0.009	−1.404 ***	−1.669 ***	0.533 *
	(0.014)	(0.014)	(0.014)	(0.014)	(0.267)	(0.326)	(0.273)
firm_age	−0.0001	−0.0003	−0.0001	−0.00003	0.026	0.087 ***	−0.049 *
	(0.001)	(0.001)	(0.001)	(0.001)	(0.027)	(0.033)	(0.028)
state_own	−0.002 ***	−0.002 ***	−0.002 ***	−0.002 **	0.015	0.010	−0.037 ***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.014)	(0.017)	(0.014)
non_soe	0.003 ***	0.003 ***	0.003 ***	0.003 ***	−0.030 ***	−0.005	0.019 **
	(0.000)	(0.000)	(0.000)	(0.000)	(0.009)	(0.011)	(0.009)
for_own	0.002 ***	0.003 ***	0.002 ***	0.002 ***	−0.102 ***	−0.196 ***	0.394 ***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.012)	(0.015)	(0.012)
lnsales	−0.010	−0.014	−0.018 *	−0.020 *	−3.112 ***	−3.396 ***	0.451 **
	(0.010)	(0.010)	(0.010)	(0.010)	(0.197)	(0.241)	(0.202)
per_gdp	−0.085 ***	−0.086 ***	−0.077 ***	−0.076 ***	3.829 ***	2.337 ***	5.049 ***
	(0.027)	(0.027)	(0.027)	(0.028)	(0.530)	(0.646)	(0.542)
Constant	7.920 ***	7.419 ***	7.728 ***	7.462 ***	35.86 ***	76.02 ***	−55.36 ***
	(0.238)	(0.244)	(0.237)	(0.250)	(5.069)	(6.186)	(5.191)
Observations	8492	8100	8296	7797	7797	7797	7797
F	23.99 ***	28.37 ***	23.53 ***	23.75 ***	39.46 ***	55.57 ***	156.50 ***
R-squared	0.030	0.037	0.030	0.038	0.066	0.091	0.220

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

reports the results of robust regression analysis. To test the robustness of H1, we add perceived institution indexes into the regression model one by one, which can be seen from Model 11 to Model 13. In addition, we regressed CEO managerial discretion on all perceived institutional factors, and the result is shown in Model 14.

Regarding the relationship between government environment and CEO managerial discretion proposed by Hypothesis 1a, Model 11 shows that the greater degree of government intervention perceived by firms and the lower level of the discretion available to CEOs of the firms in the region (β = −0.046, p < 0.01). However, the proportion of CEO managerial discretion is significantly and positively correlated to the degree of firms’ confidence in legal justice (β = 0.006, p < 0.01) in Model 12, and to the ease of firm apply for the loan from legal financial and legal institutions (β = 0.039, p < 0.01) in Model 13. Moreover, the result reported in Model 14 by adding all institutions correlates with the above. Therefore, Hypothesis 1 is fully supported.

5.3. Mediating Effect of CEO Managerial Discretion

To test the mediating effect of CEO managerial discretion between perceived investment climate and geographic diversification, we still use the Sobel test method. According to the result shown in

Table 7. Robust Sobel test: The mediating effects.

Indexes	city	inprv	inter
gov_int1	0.050, 3.472, p < 0.01	0.055, 3.377, p < 0.01	−0.046, −3.354, p < 0.01
leg_jus1	−0.005, −4.115, p < 0.01	−0.005, −3.728, p < 0.01	0.005, 4.027, p < 0.01
fin_dev1	−0.024, −2.198, p < 0.05	−0.023, −2.054, p < 0.05	0.024, 2.199, p < 0.05

Note: Sobel β, Z-value, and p-value, respectively in each group.

, CEO managerial discretion mediates the relationship between perceived investment climate and firm geographic diversification, as H3 predicts.

Furthermore, we use the Casual Steps Approach [51] to test the mediating effect of CEO managerial discretion. As stated in the literature review, the main effects of institutions on geographic diversification have received tons of supporting evidence in the Chinese context. Moreover, the relationship between perceived institutions and CEO managerial discretion is proved in the first three models of Table 5, and the relationship between CEO managerial discretion and firm geographic diversification is verified in the last three models of Table 5.

Additionally, Model 15, Model 16, and Model 17 of Table 6 regress geographic diversification on all perceived investment climate, CEO managerial discretion, and control variables. The regression coefficients of CEO managerial discretion on the three indexes of geographic diversification are all significant (β_{gov_int1} = −0.829, p < 0.01; β_{leg_jus1} = −1.001, p < 0.01; β_{fin_dev1} = 0.742, p < 0.01). In this condition, according to the Casual Steps Approach, when the regression coefficient of an institutional index is not significant, then the relationship between such institution index and geographic diversification is fully mediated by CEO managerial discretion. When the regression coefficient is significant, the relationship between institution and diversification is partially mediated. That is to say, either fully or partially mediated, the mediating effect of CEO managerial discretion has been confirmed, as Hypothesis 3 proposes.

6. Conclusions, Implications, and Discussion

6.1. Conclusions

The study’s goal is to see how investment climate affects firms’ geographic diversification. Combining the Institution-Based View with Upper Echelons Theory using a large sample firm-level dataset from The World Bank, we empirically study the mediating role of CEO managerial discretion between investment climate and firm geographic diversification, and draw three conclusions as follows. First, government intervention constrains CEO managerial discretion available to CEOs whose firms are headquartered in the region, while legal justice and financial development improve CEO discretion of local firms. Second, the greater degree of CEO managerial discretion, the lower the proportion of firm development within the local city and province and a higher proportion of internationalization. Third, CEO managerial discretion mediates the relationship between investment climate and firm geographic diversification.

The above findings show the mechanism of the institutional environment affecting firm geographic diversification by the mediating effect of CEO managerial discretion, which further expands the role of managerial discretion from the moderator to the mediator. More importantly, these findings strongly support the analytical framework of “institutional environment—managerial discretion—geographic diversification”, which sheds light on future exploration on the mechanism of other strategic choices besides geographic diversification, and more explorations on bridging macro and micro domains [52,53].

According to the above explanations, the conclusions regarding the analytical framework are shown in Figure 2.

6.2. Implications

Since the investment climate has a great impact on CEO managerial discretion and, in turn, determines the firm geographic diversification, top managers should comprehensively consider the external institutional environment and their firms’ stage within the life cycle. Specifically, we prove that local government intervention generally constrains managerial discretion. However, it does not mean that the local government’s role in firm development is always negative. Inversely, at the starting stage, firms need various kinds of resources, therefore it is helpful for firms to win governmental protection to survive relatively fierce competition. At the growth stage, firms should avoid over-relying on local protection to build core competitiveness. After entering the maturity stage, firms could enter the provinces with a relatively sustainable investment climate to reduce the potential risk of the portfolio. Furthermore, top executives could promote the entrepreneurial spirit to force the reformation of macro institutions.

Given the importance of investment climate for both domestic private investment and foreign direct investment, and in turn for regional economic development, policy-makers should make every effort to improve the investment climate, including governmental, legal and financial environment. Specifically, the local governments of China should reconstruct state-owned banks, promote commercialization of banks and marketization of interest rates, deepen reformation of the capital market, perfect laws and regulations, and improve new established private firms’ access to long-term risk capital. For the central government, since investment climate varies across China, countermeasures could be taken from three aspects to construct a uniform domestic market. First, to continually exert the central governments’ macro control function, optimize each province’s resource allocation, narrow down inter-provincial development gap. Second, to increase the development discretion of provinces. Third, to consider a re-division of the administrative area at province level in order to decrease the powerful position of the local government.

6.3. Limitations and Directions

First, this study proves the crowding-out effect of government intervention, and the spill over effects of legal justice and financial development. Further exploration could pay attention to the “helping hands” of government instead of “grabbing hands” [3], using the Inducement-Constraint Perspective [45], which will provide a more objective angle to understand the role of the government environment [54].

Second, incoming research could make efforts to explore the effects of informal institutions (e.g., trust) on firm geographic diversification by developing suitable measurements in the context of China [55]. Then, it is worthy to compare the substitute effect with the complementary effect of the formal and informal institutions [56].

Lastly, on basis of the framework we constructed, future studies could examine the effects of institutions on entrepreneurship [57,58], or corporate governance (e.g., the governance of inter-organizational business dispute) [59], strategic human resource management [60], and multilevel organizational research [61], through the mediating role of strategic leadership to contribute to bridging macro and micro domains.