Table of Contents

Internet development and entrepreneurship

Jia Guo a,b,* , Zhiming Cheng b,c , Ben Zhe Wang d

a) School of Business, University of the Chinese Academy of Social Sciences, Beijing, China

b) Department of Management, Macquarie Business School, Macquarie University, Sydney, New South Wales, Australia

c) Social Policy Research Centre, The University of New South Wales, Sydney, New South Wales, Australia

d) Department of Economics, Macquarie Business School, Macquarie University, Sydney, New South Wales, Australia

ARTICLE INFO

Keywords:
Internet development
Entrepreneurship
China

ABSTRACT

We examine the relationship between Internet development and entrepreneurship in China, using survey data from the China Family Panel Studies on individual entrepreneurship and administrative data on new firm registration. Employing instrumental variable (IV) and difference-indifferences (DID) approaches to address the endogeneity of Internet development, we find that Internet development increases the likelihood of individual entrepreneurship and the number of new firm registrations. Our IV model suggests that a one standard deviation increase in the Internet development index increases the likelihood of becoming an entrepreneur by 0.588 standard deviations. Our DID model finds that the Broadband China Program – the inaugural national broadband infrastructure initiative – increases annual new firm registrations by 14,358 in the selected prefectures. These results are robust to a battery of robustness checks. We find that improved access to information and finance are the key mechanisms through which Internet development affects entrepreneurship.

Introduction

Entrepreneurship requires a supportive ecosystem to develop and thrive (Baker & Nelson, 2005; Chen et al., 2020; Ramoglou & Zyglidopoulos, 2015). Macro-environmental forces, which vary across sectors, locations, times, and socio-demographic groups, can facilitate or inhibit entrepreneurial activity (Davidsson et al., 2020). The theory of external enablers describes the influence of these macro-environmental changes on entrepreneurial behaviours (Davidsson, 2015). These external enablers are crucial as they define the context in which businesses operate and evolve, often dictating market dynamics, competitive pressures, and strategic opportunities (Davidsson, 2015; Davidsson et al., 2020). This paper examines how Internet development, as a crucial external enabler, shapes entrepreneurial activities. Using survey and administrative data on household ventures and new business creation, we find that a higher level of Internet development facilitates entrepreneurship. Our results suggest that better access to information and financial resources are the main channels through which Internet development drives entrepreneurial activities. We further explore the non-linear relationship between the level of Internet development and entrepreneurship, particularly how the Internet can foster a winner-takes-all dynamic that might impede entrepreneurial growth in sectors highly vulnerable to digitalisation. Our study contributes to the literature on the impacts of Internet development on socioeconomic development. Previous research has explored various dimensions of the impacts of Internet development, including its effects on social capital (Arnott & Bridgewater, * Corresponding author. E-mail addresses: jia.guo3@hdr.mq.edu.au (J. Guo), zhiming.cheng@mq.edu.au (Z. Cheng), ben.wang@mq.edu.au (B.Z. Wang). Contents lists available at ScienceDirect China Economic Review journal homepage: www.elsevier.com/locate/chieco https://doi.org/10.1016/j.chieco.2024.102280 Received 28 February 2024; Received in revised form 28 June 2024; Accepted 8 September 2024 China Economic Review 88 (2024) 102280 Available online 11 September 2024 1043-951X/© 2024 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

2002; Batjargal, 2007; DiMaggio et al., 2001; Wang et al., 2022), inequality (Anderson et al., 1997; Compaine, 2001; LaRose et al., 2007), culture (Danet & Herring, 2007; Porter, 2013), organisation and economic structures (Batjargal, 2007; Wang et al., 2022), and political participation (Agre, 2002; Farrell, 2012). Our study contributes to the literature by demonstrating how Internet access drives entrepreneurship, thereby fostering socioeconomic development in China. Our paper also contributes to the literature on the determinants of entrepreneurship. Previous studies have investigated a wide range of factors, including personal characteristics such as age (Bonte ¨ et al., 2009; Fischer et al., 1993), gender (Fischer et al., 1993), marital status(Parker, 2008), religion (Zelekha et al., 2014), language proficiency (Wei et al., 2019), cognitive abilities (Hafer & Jones, 2015), and psychological attributes (Frese & Gielnik, 2014), as well as personal experiences (Awaworyi Churchill et al., 2023; Cheng et al., 2021; Yu et al., 2022), family influences (Aldrich et al., 1998), and macro environment such as regional culture (Doepke & Zilibotti, 2014), public policies (Chen et al., 2023; Hayward et al., 2022, 2023), industry regulations, and property rights (Pathak et al., 2013). We contribute to this line of research by examining the Internet as a critical factor in facilitating entrepreneurship. In addition, we contribute to the literature by providing evidence of a causal relationship between Internet development and entrepreneurship. One existing study related to ours is Tan and Li (2022), who show that personal Internet use increases household propensity for entrepreneurship. Although using lagged personal Internet use as their key independent variable of interest, Tan and Li (2022) remain prone to potential endogeneity issues that make causal inferences challenging. Our study differs from Tan and Li (2022) in that we focus on regional Internet development, which is external to individuals and firms. In addition, we use instrumental variable (IV) and difference-in-differences (DID) approaches to address endogeneity. We also conduct a series of robustness checks of our IV and DID models.

2. Internet development and entrepreneurship

The Internet can diminish entry barriers for new businesses and facilitate access to information and financial instruments, which have traditionally been monopolised by well-capitalised entities (Davidsson, 2015). The access enabled by the Internet allows nascent enterprises to leverage global data repositories, sophisticated analytical tools, and comprehensive market intelligence. These resources are integral for discerning and exploiting emergent market opportunities (Davidsson et al., 2020). As a result, the Internet mitigates information asymmetry between established corporations and new entrants, thereby enhancing competitive parity across sectors. The Internet also democratises access to educational resources and entrepreneurial training programs (Mack et al., 2017). Online courses, webinars, and virtual mentorship programs give aspiring entrepreneurs the knowledge and skills necessary to start and grow their businesses (Gentile et al., 2020). This widespread availability of educational content reduces the barriers to acquiring critical business acumen, which was once confined to formal institutions and exclusive networks. Moreover, the proliferation of online forums and professional networks facilitates knowledge sharing and collaboration among entrepreneurs globally, fostering a culture of continuous learning and innovation. As entrepreneurs gain access to best practices and cutting-edge industry insights, they are better equipped to tackle challenges and seize opportunities, further enhancing the overall entrepreneurial ecosystem.

3. The Internet and access to information

Information accessibility is paramount in contemporary business operations (Evans & Wurster, 1997). Even in sectors traditionally deemed less information-oriented, information comprises a significant portion of the cost framework. It serves as the backbone for opportunity recognition, informed decision-making, and strategic planning (Shane, 2000). Entrepreneurs rely on comprehensive, accurate, and timely information to navigate market complexities, comprehend consumer needs, monitor competitors, and identify potential innovation opportunities (Ardichvili et al., 2003). Internet access has democratised information acquisition, rendering it more equitable (Autio et al., 2018), timely, and cost-effective in replicating and disseminating information (Arnott & Bridgewater, 2002). Before the Internet, obtaining business-related information was prohibitively expensive for most entrepreneurs, who relied heavily on libraries, trade publications, formal market research, and personal networks to gain insights into market trends and business opportunities (Hisrich et al., 2017). Real-time information is crucial for business success as it enables tailoring offerings to meet evolving consumer needs effectively (Asongu & Moulin, 2016). This involves accessing extensive market data and consumer feedback and processing and utilising this information efficiently, thus enhancing the adaptability and competitiveness of businesses (Brynjolfsson & McAfee, 2014). This is particularly important in rapidly evolving markets, helping businesses stay attuned to consumer trends and technological advancements, and enabling a more dynamic and responsive approach to business strategy and operation (Kaplan & Haenlein, 2010). Without Internet access, the environment favoured those with more financial, social, and otherwise resources as they could more easily afford the costs associated with accessing critical information and leveraging personal networks to their advantage (Burt, 1992). Internet access empowers entrepreneurs to build and engage with their networks through digital platforms and social media. These platforms are crucial in facilitating global connections among entrepreneurs, linking them with peers, mentors, investors, and customers (Ellison et al., 2007). This enhanced connectivity significantly contributes to the development of social capital. This social capital can be an important asset for entrepreneurs, enabling them to access advice, financial resources, and support networks that were previously beyond their reach (Awaworyi Churchill et al., 2023).

The Internet and access to finance

Recognised as the most commonly cited obstacle to entrepreneurship, financial difficulty plays a crucial role in the initiation of new J. Guo et al. China Economic Review 88 (2024) 102280 2 business ventures (Andersen & Nielsen, 2012; Cai et al., 2018; Field et al., 2010; Holtzeakin et al., 1994; Ma et al., 2019; Michelacci & Silva, 2007). Entrepreneurial capital predominantly originates from self-owned resources or financing through loans. Given the risk, entrepreneurs are often discriminated against by formal lending channels through stringent credit assessments (Brush & Cooper, 2012; Ekpe et al., 2011; Muravyev et al., 2009) and higher interest rates (Muravyev et al., 2009). Internet technology, such as big data and cloud computing, reduce information asymmetry for lenders in the financial market by portraying the credit image of customers (Asongu & Moulin, 2016), which helps those traditionally excluded from formal financial institutions access capital resources. Internet infrastructure also enhances information transparency and service convenience, intensifying market competition and significantly reducing the cost of obtaining credit. For example, digital lending platforms use advanced algorithms to assess credit risk more accurately, thereby reducing the need for expensive traditional credit assessments and collateral requirements (Moro et al., 2015). Internet infrastructure enables extensive data analytics to offer diverse services tailored to potential entrepreneurs’ nuanced financial needs (Yang et al., 2022). This capability to provide personalised financial solutions is important in addressing the specific challenges and opportunities entrepreneurs face. Digital finance innovation, including peer-to-peer lending and crowdfunding, has introduced new financing models that cater more effectively to the entrepreneurial sector, broadening the spectrum of available financial services (Gomber et al., 2017). The rise of FinTech, propelled by the Internet, revolutionisesthe financial sector by injecting liquidity through alternative financing channels such as peer-to-peer lending platforms, online crowdfunding, and advanced digital payment systems (Guo et al., 2016). These financial innovations decrease the cost of capital for startups and reduce their reliance on conventional financial intermediaries, thus facilitating a more efficient capital allocation within the economy (Brynjolfsson & McAfee, 2014). The Internet’s role, therefore, goes beyond mere access to information. It redefines the resource allocation mechanism, bolstering entrepreneurial efficiency and innovation. This enhancement of allocative efficiency and reduction in transaction costs underpin a more dynamic, competitive, and resilient economic ecosystem, making it more conducive to entrepreneurship. Internet development also provides entrepreneurs with access to other digital finance resources, such as mobile transactions, digital banking services, and digital payment platforms, making financial services more accessible to startups previously only available to established firms (Guo et al., 2016), which enhances the efficiency and effectiveness of entrepreneurial activities.

5. Data and methods

5.1. Data

We utilise data from multiple sources. The first is the China Family Panel Studies (CFPS), which is administered by Peking University. The CFPS is an annual longitudinal social survey designed to facilitate research on a wide array of socioeconomic phenomena in contemporary China, of which the first wave was collected in 2010. In this paper, we use the 2010, 2012, 2014, 2016, and 2018 waves of the CFPS. Our analytical sample includes interviews with 27,858 individuals across 30 provinces. The second source of our data is Tianyancha, a business data provider in China, from which we obtain the data on the annual number of company registrationsin each prefecture. For control variables such as province and prefecture characteristics, we use data from the Chinese Research Data Services (CNRDS) and the National Bureau of Statistics of China (NBSC), depending on the exact specifications. 5.2. Methods We estimate the following model to examine the relationship between Internet development and entrepreneurship, entreijt = α + βintjt + Xʹ itΥ + Zʹ jtΦ + μi + ϑj + νt + ϵit (1) where entreijt is a binary variable equalling one if respondent i is an entrepreneur in province j in year t, and zero otherwise (i.e., the respondent is an employee).1 The key independent variable of interest, denoted as intjt, represents the degree of Internet development in province j in year t. This is measured by an Internet development index on a scale from 0 to 1, which integrates the Internet penetration rate, the count of mobile Internet users, and the count of broadband Internet users. A higher value of the index indicates a higher level of Internet development. Appendix B presents the detailed methodology for constructing the index using the entropy weight method. Fig. 1 shows the level of Internet development in China over a period from 2010 to 2018. Over the years, Internet development has been increasing across the country. By 2018, a significant part of China appears to have a development level of 0.4 or higher. The eastern coastal provinces exhibit the highest level of Internet development. However, back in 2010, most of the country appeared to have an Internet development level below 0.3, especially in the central and western regions of China. The estimated coefficient, β, for intjt denotes the effect of having a one-unit increase in the Internet development index on the probability of being an entrepreneur. Of the remaining variables, Xʹ it is a vector of individual characteristics, including marital status, schooling, and residential area (i.e., urban vis-a-vis ` rural area). Zʹ jt is a vector of provincial characteristics, including GDP per capita, ratio of general fiscal expenditure, ratio of private sector employees, and so on. ϵit is the error term.

Eq. (1) is estimated using two-way fixed effects (TWFE), controlling for both individual fixed effects (μi ) to mitigate the impacts of time-invariant individual characteristics and unobservables, such as family backgrounds and initial capital, which may influence entrepreneurial decisions, as well as time (wave) fixed effects (νt) to address the unobserved influence of secular changes and common shocks over time. In addition, we control for province fixed effects ( ϑj ) to address the impact of unobservables, such as culture and institutions that are common within a province. Table 1 provides a detailed description and descriptive statistics of the variables used in our analyses. The final total sample size from CPFS is 95,378 involving 27,858 respondents. Considering retirement ages, we use a sample of males aged between 16 and 60 and females aged between 16 and 50. The level of Internet development can be endogenous, with possible sources of endogeneity stemming from measurement error, reverse causality, and omitted variables. For example, measurement errors, which are well known to attenuate estimated coefficients (Buonaccorsi, 2010), may arise from inaccurate data. Reverse causality could occur if the entrepreneurial activities influence other factors, such as economic growth (Valliere & Peterson, 2009; Stel, Carree, & Thurik, 2005; Wong et al., 2005) and technology innovation (Stel, Carree, & Thurik, 2005), which then feedback into Internet development. Omitted variables, such as socioeconomic factors or infrastructure quality, may also affect both Internet development and entrepreneurship. To address the endogeneity of Internet development, we employ an IV approach. We use two IVs. Specifically, we utilise an interaction term between historical landline telephone penetration rate and the contemporary prevalence of the Internet as our IV. Our IV strategy shares similarities with Dettling (2017), which employs an interaction term between the percentage of a state’s population living in multiple-dwelling units (e.g., apartments and units), where Internet service providers (ISPs) can upgrade broadband wiring from telephone lines more efficiently and cost-effectively for multiple customers simultaneously, and a vector of year fixed effects to allow for the identification of flexible trends in Internet diffusion across states in the United States. Specifically, we use an interaction term between landline coverage in 1990 and the proportion of Internet users accessing the Internet through Internet caf´es in each survey year within a province as an IV in our study. We gauge historical landline coverage as measured by the ratio of landline telephone subscriptions to a province’s population in 1990, obtained from the NBSC. This selection of 1990 landline coverage aligns with the period when dial-up Internet access relied on telephone lines. In addition, the introduction of high-speed broadband Internet in the late 1990s and early 2000s prompted substantial infrastructure investments by ISPs. This involved retrofitting existing phone and cable lines and deploying new switches and servers. The installation of residential high-speed Internet often necessitated upgrading the wiring connecting indoor infrastructure to the ISP, while the existing premises’ wiring typically remained untouched. This situation favoured regions with dense landline coverage for installation efficiency, feasibility, and profitability for ISPs, contrasting with areas with fewer and less concentrated landlines. In regions with denser landline coverage, each upgraded wiring length catered to multiple customers, facilitating economies of scale, and enabling a cost-effective provision of Internet access to potential consumers. Considering these dynamics, we propose that provinces with a higher prevalence of landlines likely gained earlier and broader access to dial-up and, subsequently, high-speed broadband Internet compared to those with fewer landlines. This insight underscores the influence of historical infrastructure on shaping contemporary Internet accessibility patterns. For contemporary Internet prevalence, we use data from the NBSC to measure the proportion of Internet users accessing the Internet through Internet caf´es in each survey year within a province. This choice is motivated by the fact that Internet caf´es serve as

Table 1: Summary Statistics

Variables	Descriptions	Mean	S.D.
Outcome variables
Entrepreneurship	Respondent’s employment status (solo or employer entrepreneur = 1; employee = 0)	0.111	–
No. of new firm	Annual number of registered new firms in a prefecture	46,290	55,922
Key independent variables
Internet development index	On a scale from 1 to 0. See Appendix B for estimation methods.	0.309	0.179
Broadband China Program (BCP)	Selected by the Chinese government as a BCP pilot prefecture (yes = 1; no = 0)	0.165	–
Individual controls
Married	Marital status (married = 1; unmarried = 0)	0.821	–
Years of education	Years of formal education	8.495	4.403
Urban area	Residential area (urban = 1; rural area = 0)	0.455	0.498
Province controls
GDP per capita	In RMB	43,046	22,613
Unemployment rate	Registered urban unemployment rate	3.277	0.591
Ratio of private sector employees	Ratio of the number of private sector employees to state sector employees	0.908	0.329
Ratio of general fiscal expenditure	Ratio of general fiscal expenditure to GDP	0.227	0.090
Ratio of sci/tech/edu expenditure	Ratio of fiscal expenditure on science, technology, and education to GDP	0.042	0.015
Ratio of tertiary industry outputs	Ratio of tertiary industry outputs to GDP	0.438	0.100
Instrumental variables
1990 landline coverage × ratio of Internet café users	An interaction term of the following: Ratio of landline telephone subscriptions in provincial population in 1990 and ratio of Internet users via Internet cafés in each survey year	0.356	0.403
Min. distance to the Eight Verticals and Eight Horizontals network × ratio of Internet café users	An interaction term of the following: Minimum distance from the centroid of each province to the backbone city of the national Eight Verticals and Eight Horizontals network and ratio of Internet users via Internet cafés in each survey year	0.004	0.003
Moderators (Self-rated on a 5-point scale)
Relational competence	Getting along with others (lowest = 0; highest = 10)	5.622	1.561
Social popularity	Popular with others (lowest = 0; highest = 10)	6.006	1.545
Social status	Social status in local area (very low = 1; very high = 5)	2.780	0.717
Mediators (Yes = 1; no = 0)
Access to online information	The respondent has obtained information from the Internet	0.822	–
Access to entrepreneurial funding	The respondent has obtained entrepreneurial funding from the following source
– Personal fund		0.799	–
– Relative and friend		0.094	–
– Investor		0.031	–
– Venture capital		0.001	–
– Commercial loan		0.042	–
– Policy support loan		0.001	–
– Other sources		0.032	–
Denial of entrepreneurial funding	The respondent has been denied entrepreneurial funding from the following source
– Relative and friend		0.695	–
– Formal institution		0.177	–
– Informal institution		0.018	–

accessible and cost-effective hubs for individuals, especially in provinces where personal Internet connections may be limited or costly, and that the popularity of Internet caf´es reflects exogenous and flexible time trends in Internet development within a province. The fundamental assumption underpinning the interpretation of our IV estimate as causal is that the provincial landline subscription in 1990 would not have been systematically correlated with subsequent trends in entrepreneurship had Internet development and diffusion not occurred. This assumption does not imply that the level of landline coverage must be devoid of correlation with differences in entrepreneurship levels across provinces. It is recognised that provinces with higher landline rates in 1990 also tend to exhibit higher average incomes and greater population densities in subsequent survey years, factors that may relate to disparities in entrepreneurship levels. Assuming these correlations are time-invariant, they are absorbed by the province fixed effects. In addition, we control for a series of provincial characteristics. In essence, exogeneity is upheld if a province’s landline coverage in 1990 is not correlated with changes in entrepreneurship within the province once adjustments are made for changes in measured socioeconomic conditions within the province over time and provincial fixed effects. One may still be concerned that landline coverage in 1990 may be correlated with time-variant unobservables that cannot be absorbed by province fixed effects and control variables. To address this, we employ an alternative IV by interacting the minimum distance between a province’s geographic centroid (i.e., the average location of all the points within a province) and the main optical cable in 1998 with the proportion of Internet users accessing the Internet through Internet caf´es in each survey year within a province. In 1998, China completed the so-called ‘Eight Verticals and Eight Horizontals’ fibre-optic grid, which consists of eight major northsouth (vertical) and eight east-west (horizontal) optical cable routes, forming a grid-like structure that ensures extensive coverage and connectivity (see, Chen (2023), for a review). Given that the optical cable branches within provinces are extended from the interprovincial optical cable mainlines, the farther a city is from the mainline, the higher the costs of broadband network deployment and the slower the diffusion of the Internet. This alternative IV leverages the geographical proximity to crucial telecom infrastructure,

which influences Internet access in a way that is plausibly exogenous to entrepreneurship. Conley et al. (2012) present practical methods for performing inference while relaxing the exclusion restriction for IV. When the instrument deviates from perfect exogeneity, these methods can be used to examine the sensitivity of the IV results. Specifically, the union of confidence intervals (UCI) approach is employed to identify the union of intervals for the effect of Internet development on individual entrepreneurship when the IV is not perfectly exogenous, i.e., when the effect of the IV on entrepreneurship (γ) is not equal to zero. Conley et al. (2012) suggest that γ should lie within a specific support interval [ − δ; +δ]. Following Madsen et al. (2018), δ is estimated by including the direct effect of IV on outcome variable in Eq. 1. Confidence regions for β are then estimated for any γ in that interval. The marginal effects are statistically significant when the upper and lower bounds of confidence intervals are both either above or below zero. One may still be concerned that we cannot directly test the exclusion restrictions of our IVs. To address this, we employ a DID approach. Specifically, we use the rollout of the Chinese Government’s Broadband China Program (BCP) since 2013 to identify the effect of Internet development on entrepreneurship. The BCP is a national initiative to boost broadband infrastructure to improve nationwide Internet accessibility. It focuses on expanding high-speed Internet through fibre-optic networks and upgrading existing telecom infrastructure. In 2014, 2015, and 2016, the Chinese Government designated 120 cities as BCP pilot cities. We estimate the effect of the BCP on entrepreneurship using TWFE,

$\text{new firm}_{kt} = \alpha + \beta \text{BCP}_{kt} + \mathbf{P’} \mathbf{y}_{k} + \mu_{k} + \nu_{t} + \epsilon_{kt} \quad (2)$

where new firmkt represents the annual number of company registrations in city k in year t. BCPkt represents whether city k is selected as a BCP pilot city in year t. Pʹ kt is a vector of city-level covariates, including GDP per capita, ratio of private sector employees, ratio of general fiscal expenditure, ratio of sci/tech/edu expenditure, ratio of tertiary industry outputs. μk denotes city fixed effects, νt indicates time fixed effects, and ϵkt is the error term. The TWFE-DID estimator is essentially a weighted average of all possible two-group/two-period estimators that compare one group changing its treatment status to another group that does not. Bias in the TWFE-DID estimator may arise when already-treated units act as the control group if the treatment effects are heterogeneous across different times. In the context of the BCP, the issue could become more prominent when the cities that are later treated are of great numbers. To address this, we first employ the Goodman-Bacon (2021) decomposition to calculate the average effect and total weight of the three types of treatment: earlier/later treated vs. never treated, earlier treated vs. later untreated, and later treated vs. earlier treated. The last pair (i.e., later treated vs. earlier treated) is considered a ‘bad’ control group. To address this, we employ the approach proposed by Callaway and Sant’Anna (2021), which allows the treatment effect to be heterogeneous across different times. The framework also relies on the existence of groups that never receive treatment, aligning well with our research on BCP where such groups are present. Additionally, our study does not involve scenarios of policy exits, making this approach ideal for robustly assessing the treatment effects without the complications that such conditions would entail. 5.3. Mediators Since we focus on access to information and funding as crucial channels through which Internet development fosters entrepreneurship, we will test several mediators in our analysis. The first mediator, information accessibility, is indicated by a binary variable that equals one if the respondent obtains information from the Internet.2 We test a set of seven binary variables representing different sources of entrepreneurial funding, including personal funds, relatives and friends, investors, venture capital, commercial loans, policy support, and other sources. Each variable is assigned a value of one if the funding is from that source and zero otherwise. Additionally, we examine three binary variables related to loan application/request rejections, categorised by who rejected the application, such as relatives and friends, formal financial institutions, and informal financial institutions. Each variable is assigned a value of one if the rejection came from that specific source and zero otherwise. 5.4. Moderators Entrepreneurs get access to resources either by themselves, through the market system, or a combination of both (Kuada, 2009; Peng & Luo, 2000). In an imperfect market, personal networks could help entrepreneurs overcome the problem of business information asymmetry and gain greater access to resources (Barnett et al., 2019; Chen et al., 2015). As illustrated in Section 2, the Internet may have substantial potential to enhance resource accessibility, providing critical access to information and financial tools. For those with robust networks, however, the incremental benefits in terms of accessing entrepreneurial resources might be less pronounced (Ellison et al., 2007). For individuals already possessing better social networks and higher social status, the Internet mainly serves to maintain and marginally expand their network (Ellison et al., 2011). This maintenance can be crucial for entrepreneurial activity by ensuring the flow of information and resources but does not fundamentally alter the landscape of opportunities available to them (Burt, 2000; 2 In the 2010 wave, we identify this variable as equal to one if a respondent answered ‘Internet’ to the question: ‘What is your main way to get information?’; otherwise, it is equal to zero. For the 2014, 2016, and 2018 waves, we assign the binary variable a value of one if the respondent’s response to the question about the importance of the Internet in obtaining information, measured on a five-point scale (1 = never use; 5 = very important), is one or above; otherwise, it is zero.

\text{new firm}_{kt} = \alpha + \beta \text{BCP}_{kt} + \mathbf{P’} \mathbf{y}_{k} + \mu_{k} + \nu_{t} + \epsilon_{kt}

(2) where new firmkt represents the annual number of company registrations in city k in year t. BCPkt represents whether city k is selected as a BCP pilot city in year t. Pʹ kt is a vector of city-level covariates, including GDP per capita, ratio of private sector employees, ratio of general fiscal expenditure, ratio of sci/tech/edu expenditure, ratio of tertiary industry outputs. μk denotes city fixed effects, νt indicates time fixed effects, and ϵkt is the error term. The TWFE-DID estimator is essentially a weighted average of all possible two-group/two-period estimators that compare one group changing its treatment status to another group that does not. Bias in the TWFE-DID estimator may arise when already-treated units act as the control group if the treatment effects are heterogeneous across different times. In the context of the BCP, the issue could become more prominent when the cities that are later treated are of great numbers. To address this, we first employ the Goodman-Bacon (2021) decomposition to calculate the average effect and total weight of the three types of treatment: earlier/later treated vs. never treated, earlier treated vs. later untreated, and later treated vs. earlier treated. The last pair (i.e., later treated vs. earlier treated) is considered a ‘bad’ control group. To address this, we employ the approach proposed by Callaway and Sant’Anna (2021), which allows the treatment effect to be heterogeneous across different times. The framework also relies on the existence of groups that never receive treatment, aligning well with our research on BCP where such groups are present. Additionally, our study does not involve scenarios of policy exits, making this approach ideal for robustly assessing the treatment effects without the complications that such conditions would entail.

5.3. Mediators

Since we focus on access to information and funding as crucial channels through which Internet development fosters entrepreneurship, we will test several mediators in our analysis. The first mediator, information accessibility, is indicated by a binary variable that equals one if the respondent obtains information from the Internet.2 We test a set of seven binary variables representing different sources of entrepreneurial funding, including personal funds, relatives and friends, investors, venture capital, commercial loans, policy support, and other sources. Each variable is assigned a value of one if the funding is from that source and zero otherwise. Additionally, we examine three binary variables related to loan application/request rejections, categorised by who rejected the application, such as relatives and friends, formal financial institutions, and informal financial institutions. Each variable is assigned a value of one if the rejection came from that specific source and zero otherwise.

5.4. Moderators

Entrepreneurs get access to resources either by themselves, through the market system, or a combination of both (Kuada, 2009; Peng & Luo, 2000). In an imperfect market, personal networks could help entrepreneurs overcome the problem of business information asymmetry and gain greater access to resources (Barnett et al., 2019; Chen et al., 2015). As illustrated in Section 2, the Internet may have substantial potential to enhance resource accessibility, providing critical access to information and financial tools. For those with robust networks, however, the incremental benefits in terms of accessing entrepreneurial resources might be less pronounced (Ellison et al., 2007). For individuals already possessing better social networks and higher social status, the Internet mainly serves to maintain and marginally expand their network (Ellison et al., 2011). This maintenance can be crucial for entrepreneurial activity by ensuring the flow of information and resources but does not fundamentally alter the landscape of opportunities available to them (Burt, 2000; 2 In the 2010 wave, we identify this variable as equal to one if a respondent answered ‘Internet’ to the question: ‘What is your main way to get information?’; otherwise, it is equal to zero. For the 2014, 2016, and 2018 waves, we assign the binary variable a value of one if the respondent’s response to the question about the importance of the Internet in obtaining information, measured on a five-point scale (1 = never use; 5 = very important), is one or above; otherwise, it is zero.

Table 2: Internet Development and Individual Entrepreneurship

	(1) TWFE	(2) IV-FE
Outcome variable: individual entrepreneurship
Internet development	0.057*	1.031***
	(0.032)	(0.313)
First stage results
1990 landline coverage × ratio of Internet café users		0.032***
		(0.001)
The first stage F statistic		486.442
Control variables	Yes	Yes
Individual fixed effects	Yes	Yes
Province fixed effects	Yes	Yes
Year fixed effects	Yes	Yes
N	95,378	95,378

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level (in parentheses). Full results are available from the authors upon request.

Putnam, 2000).

To further understand the role of self-perceived social attributes in examining Internet development and entrepreneurship, we use several moderators as proxies. The first moderator we utilise is self-rated relational competence, gauged on a ten-point scale where zero denotes ‘very difficult to get along with others’ and ten signifies ‘very easy to get along with others’. The second moderator is selfrated social popularity, rated on a ten-point scale, with zero indicating oneself being ‘very unpopular for others’ and ten indicating ‘very popular for others’. The third moderator we employ is self-rated social status within one’s community, assessed on a five-point scale ranging from ‘very low’ to ‘very high’.

Main results

Table 2 presents the TWFE and IV estimates regarding the relationship between the Internet development index and individual propensity for entrepreneurship. Model 1 shows that a one standard deviation increase in the Internet development index is associated with a 0.033 standard deviation increase in one’s probability of being an entrepreneur [(0.057 × 0.179)/0.314 = 0.033]. The IV results in model 2 show that a one standard deviation increase in the Internet development index corresponds to a 0.588 standard deviation increase in one’s probability of being an entrepreneur [(1.031 × 0.179)/0.314 = 0.588]. The F statistics for the IV exceed the rule of thumb value of 10 or even higher values (e.g., 50 or 104.7) suggested by more recent literature (Lee et al. (2022). In sum, both the TWFE and IV estimates support that Internet development has a positive impact on individual propensity for entrepreneurship. In Table 3, model 1 presents the results of the effects of the BCP program on entrepreneurial activities using TWFE-DID. The results suggest that being selected as a BCP city increases the annual number of new firm registrations by an average of 15,664. Again, this supports the idea that Internet development has a positive impact on entrepreneurship. We also use an event study approach spanning 11 years, comprising six years before and five years after the initial implementation of the BCP. Fig. 2 illustrates that the impact of being selected as a BCP city on new firm registrations has only emerged, and consistently grown, over the post-implementation period, notably accentuated after the first two years of implementation. This trend indicates the time needed from selection into the BCP to develop and improve Internet infrastructure, consequently influencing entrepreneurial activities. Table 4 presents the results from the Goodman-Bacon (2021) decomposition for the DID analysis based on the BCP. The first row compares cities treated earlier with those treated later, with the latter serving as the comparison group and a weight of 5.2 %. The second row represents a ‘bad’ comparison, where cities treated earlier serve as the comparison group for cities treated later, carrying the lowest weight at 4 %. The last row compares cities that never received the treatment, serving as the comparison group, and has the highest weight at 90.7 %. The decomposition results show that the issue of bad comparison in the TWFE-DID estimator is relatively minor. Even though the problem of negative weights is not prominent in TWFE-DID, we also employ the staggered DID approach method proposed by Callaway and Sant’Anna (2021) (i.e., CSDID). We use the ‘never treated’ cities as the ‘clean’ control group and divide the pilot cities into different groups according to the year when they were first selected. Models 2–5 in Table 3 present the results from this CSDID heterogeneity robust estimator. Model 2 reports the average treatment effect on the treated (ATT) using the simple weighted average method (weighted by group size), showing that being selected as a BCP city increases the annual number of new firm registrations by 14,358. Further detailed results, including the average treatment effect by cohort and calendar year,3 are presented in Table 2 Internet development and individual entrepreneurship. (1) (2) Outcome variable: individual entrepreneurship TWFE IV-FE Internet development 0.057* 1.031*** (0.032) (0.313) First stage results 1990 landline coverage × ratio of Internet caf´es users 0.032*** (0.001) The first stage F statistic 486.442 Control variables Yes Yes Individual fixed effects Yes Yes Province fixed effects Yes Yes Year fixed effects Yes Yes N 95378 95378 Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level in parentheses. Full results are available from the authors upon request. 3 The cohort ATT first calculates the average effect for each group across all time periods and then it averages these effects together across groups to summarise the overall average effect of being selected as the BCP pilot city. The calendar time ATT first calculates the average effect of participating in the program in each time period across all treated groups and then it averages these effects across all time periods to obtain the overall average effect of being selected as the BCP pilot city.

Table 3: Broadband China Program and New Firm Registrations

Outcome variable: No. of new firms	(1) TWFE-DID	(2) CSDID Simple ATT	(3) ATT by group	(4) ATT by calendar year	(5) ATT by periods before and after treatment
BCP cities	15,663.649***	14,358.400***	11,334.670***	14,595.170***
	(5,094.392)	(4,756.092)	(3,922.372)	(4,368.540)
Before treatment					654.331
					(765.722)
After treatment					18,744.910***
					(6,035.492)
Control variables	Yes	Yes	Yes	Yes	Yes
City fixed effects	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes
N	2,106	2,106	2,106	2,106	2,106

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the city level (in parentheses). Full results are available from the authors upon request.

Table 4: Goodman-Bacon (2019) Decomposition – Broadband China Program and New Firm Registrations

Comparison	Beta	Total weight
Earlier treated vs. later treated	18,309.310	0.052
Later treated vs. earlier treated	−18,700.000	0.040
Treated vs. never treated	16,899.435	0.907

models 3 and 4, respectively, with all effects shown to be positive. Model 5 delineates the ATT for periods before and after the implementation of the BCP policy, suggesting that the treatment effect is significant only post-implementation. Fig. 3 illustrates the dynamic treatment effects using an alternative event study approach that accounts for heterogeneous treatment effects. The Fig. shows that the confidence intervals before the BCP include zero, while the post-treatment effects of BCP are positive and significant. In other words, the BCP only affected the creation of new firms after a prefecture was selected for inclusion in the program.

Robustness checks

We conduct several robustness checks of the main results. First, we perform a placebo test to assess the impact of unobservable variables on our results from model 1 in Table 3. We randomly select the distribution of the treatment group and re-estimate the model 1000 times. Fig. 4 illustrates the distribution of β-values and p-values. Most coefficients have p-values greater than 0.1, and the coefficient from the regression (15,663.649) lies well outside the distribution, indicating a large deviation. This suggests that the results are unlikely to be systematically biased by unobservables.

Second, we use the limited information maximum likelihood (LIML) approach to estimate model 2 in Table 2. Although LIML is primarily justified in the literature for its advantages over two-stage least squares (2SLS) or IV methods in cases of small sample sizes or weak instruments (Hayashi, 2011), this rationale does not directly apply to our study due to our large sample size and strong IV. Nevertheless, we utilise LIML to ensure our findings are robust to different estimation approaches. The results, presented in model 1 of Table 5, show that the outcomes from the LIML estimation are qualitatively consistent with our baseline IV-FE results in model 2 in Table 2, confirming the robustness of our conclusions.

Third, we further employ a two-step generalised method of moments (GMM) approach to estimate the effect of Internet development on entrepreneurship. The advantage of GMM is that it provides a flexible estimation method that yields consistent estimates without requiring full knowledge of the error term distribution. Model 2 in Table 5 presents the results from the GMM estimator, which

Table 5: Internet Development and Individual Entrepreneurship – Alternative Estimators

Outcome variable: entrepreneurship	(1) LIML	(2) GMM	(3) IV-FE
Internet development	1.031***	1.031***	0.296*
	(0.313)	(0.313)	(0.173)
Covariates	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes
N	95,378	95,378	95,378
Kleibergen-Paap rk Wald F statistic	486.442	486.442	2,073.886

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level (in parentheses). Full results are available from the authors upon request. The value of the “Eight Verticals and Eight Horizontals” × Ratio of Internet café users in each survey year has been scaled down by a million for presentation purposes.

are also qualitatively consistent with our baseline results from IV-FE in model 2 in Table 2. Fourth, the additional IV result is presented in model 3 of Table 5, which refects that a one standard deviation increase in the Internet development index corresponds to a 0.169 standard deviation increase in the probability of becoming an entrepreneur. In sum, while both the TWFE and IV estimates confirm the positive impact of Internet entrepreneurship, the TWFE model in model 1 of Table 2 underestimates this effect

Last, we use the Conley et al. (2012) test, which relaxes the exclusion restriction to determine if the inference remains valid. The UCI method results, shown in Table 6, assess how a direct link between the IV and the outcome variable affects the IV estimates in model 2 in Table 2. It reports the estimated bounds for 95 % confidence intervals. The marginal effects are significant if both the upper and lower bounds of these intervals exceed zero. The UCI findings indicate that the confidence interval remains above zero, strongly suggesting a positive effect of Internet development on entrepreneurship.

Results for mediators

Table 7 presents the IV-FE results of the effects of Internet development on access to information and funding, crucial factors influencing individual entrepreneurship. Model 1 shows that Internet development significantly increases the likelihood of people using the Internet to access information. This finding underscores the impact of the Internet in enhancing the accessibility and utility of information, which is crucial for informed decision-making and strategic planning in business contexts. By facilitating more accessible and affordable access to vital data and insights, the Internet enables entrepreneurs and businesses across various sectors to stay informed, respond swiftly to market changes, and harness opportunities for innovation and growth. Models 2–8 report the effects of Internet development on access to entrepreneurial funding. Model 3 suggests that Internet development has reduced the reliance on relatives and friends for entrepreneurial funding, while model 6 indicates that Internet development has significantly increased the likelihood of securing commercial loans as a means of entrepreneurial funding. A possible explanation for these findings is that historically, the absence of alternative funding options necessitated reliance on personal networks. However, with the advent of the Internet providing viable financial solutions, individuals are more inclined to seek commercial loans for their entrepreneurial endeavours. Models 9–11 report the effects of Internet development on the propensity of funding requests being rejected by various sources. Models 10 and 11 demonstrate that Internet development has significantly lowered the probability of loan rejection by formal and informal institutions, respectively. One plausible explanation is that the Internet, through digital platforms, online banking, and mobile financial applications, dismantles traditional barriers to financial inclusion, including geographical constraints, high expenses, and intricate prerequisites (Demirgüç-Kunt & Singer, 2017). This increased accessibility enables individuals and small businesses, particularly those in underserved and remote areas, to participate in the financial system (Beck et al., 2009).

Results for moderators

Table 8 presents the results of the moderators. Generally, these results show that the effect of Internet development on the probability of becoming an entrepreneur is weaker among individuals with higher self-perceived relational competence, social popularity, and social status. One potential explanation for this finding is that individuals with fewer initial advantages may benefit more from the Internet’s capacity to provide access to resources and information they would otherwise lack, thereby levelling the

Table 6: The Conley-Hansen-Rossi (2012) Test on IV Exogeneity

Plausibly bounds (95% confidence intervals)	The effects of Internet development on individual entrepreneurship
Upper bound	2.516
Lower bound	1.296

Table: Potential Mediators Between Internet Development and Individual Entrepreneurship – IV-FE Results

Outcome Variable: individual entrepreneurship	(1) Access to information via Internet	(2) Personal funds	(3) Relatives and friends	(4) Investor	(5) Venture capital	(6) Commercial loan	(7) Policy support	(8) Others	(9) Funding request rejected by Relatives and friends	(10) Formal institution	(11) Informal institution
Internet development	6.224***	1.478	−1.807*	−0.181	0.126	1.067*	0.071	−0.754	1.696	−3.087***	−0.532**
	(0.630)	(1.458)	(1.054)	(0.847)	(0.092)	(0.590)	(0.059)	(0.647)	(1.073)	(0.882)	(0.264)
Covariates	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	31,381	2,160	2,160	2,160	2,160	2,160	2,160	2,160	10,022	10,022	10,022
Kleibergen-Paap rk Wald F statistic	300.282	49.270	49.270	49.270	49.270	49.270	49.270	49.270	125.877	125.877	125.877

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level (in parentheses). Full results are available from the authors upon request.

Table 8: Potential Moderators Between Internet Development and Individual Entrepreneurship – IV-FE Results

Outcome variable: individual entrepreneurship	(1)	(2)	(3)
Internet development	4.018***	4.276***	1.533***
	(0.310)	(0.316)	(0.376)
Internet development × relational competence	−0.498***
	(0.027)
Internet development × popularity		−0.527***
		(0.030)
Internet development × social status			−0.104**
			(0.049)
Covariates	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes
N	78,856	86,997	91,666
Kleibergen-Paap rk Wald F statistic	225.612	232.204	242.646

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level (in parentheses). Full results are available from the authors upon request.

playing field and enhancing their entrepreneurial opportunities. This is consistent with the theory of underdog entrepreneurship, wherein disadvantaged individuals become entrepreneurs through resilience, resourcefulness, and other positive attributes developed in difficulties and adversities (Cheng et al., 2021; Hayward et al., 2022). In this case, the Internet provides disadvantaged individuals with the opportunity to engage in entrepreneurship by mitigating the negative impacts of their lower relational competence, social popularity, and social status.

Extensions

In many traditional product and service sectors, competitive advantage is often limited by diminishing returns to scale. Once a firm’s user base expands beyond efficient levels, no additional competitive advantage is garnered, allowing for the coexistence of competitors (Iansiti & Lakhani, 2018). Conversely, a digital platform with strong network effects, where increasing returns to scale, confers substantial advantages to a central enterprise (Iansiti & Lakhani, 2018; Verbeke & Hutzschenreuter, 2021). This contrast sets the stage for a deeper examination of the non-linear relationship between Internet development and entrepreneurship across varying levels of Internet development and within diverse industries in this section. The digital economy often witnesses the rise of the ‘winner-takes-all’ phenomenon, largely driven by the Internet’s capacity to amplify network effects. These effects occur when a product or service becomes more valuable with increased usage, favouring dominant platforms (Webster & Ksiazek, 2012). Early movers in Internet-based industries can swiftly establish dominance by leveraging their user base to enhance services or scale operations at lower costs (Evans & Schmalensee, 2005). Larger user bases generate more data, fostering innovation and solidifying market positions (Shapiro & Varian, 1999). The scalability and minimal marginal costs of digital services facilitate rapid expansion, allowing dominant firms to maintain their lead by undercutting prices or accelerating innovation (Parker et al., 2016). This makes it challenging for new entrants to compete effectively, as they lack the critical mass of users to trigger comparable network effects (Katz & Shapiro, 1985). According to Kenney and Zysman (2016), the digital platform economy enables dominant firms like Amazon and Alibaba to exploit network effects and data-driven insights, consolidating their market dominance and erecting high entry barriers for newcomers. This monopolistic trend not only limits market access for smaller ventures but also hampers innovation by concentrating economic power within a few large entities. Furthermore, Drahokoupil and Fabo (2016) emphasise that these dominant platforms leverage extensive data and customer networks, giving them significant advantages that new entrants struggle to match. This intense market consolidation diminishes diversity and undermines the competitive vitality needed for a healthy entrepreneurial ecosystem. Thus, while the Internet initially democratised market access, its evolution has paradoxically hindered the sustainable growth of entrepreneurship. The impact of the Internet on various industries varies significantly, influenced by factors such as technological adoption, industryspecific needs, and regulatory environments. Internet reshapes value chains and competitive strategies, especially in informationintensive sectors (Porter, 2001). Industries such as finance and retail, characterised as ‘highly digitisable’ are particularly susceptible to disruptions from Internet innovations, often resulting in a few dominant platforms in these markets (Bughin & Manyika, 2007). In contrast, sectors such as manufacturing and construction experience slower digital transformation (Favoretto et al., 2022). These industries are heavily reliant on physical and manual processes and constrained by geographic limitations, which prevent them from conducting business across regions. Although digital tools can improve certain operations, they do not fundamentally alter core business practices. Since industries interact with digital technologies in varied ways, we further examine the impact of the Internet on entrepreneurship across 21 industries as defined by the NBSC. These industries are classified into two highly digitisable industries, which are more prone to Internet-driven changes due to their informational intensity and digital adaptability, and lightly digitisable industries, which rely more heavily on physical processes and are less affected by digital transformations. The highly digitisable industries, encompassing transportation and warehousing, wholesale and retail trade, finance,

accommodation and food services, business services, scientific research and development, education, and cultural and recreational services, are associated with higher service orientation and knowledge intensity. The lightly digitisable industries, including mining, manufacturing, utilities (e.g., electricity and gas providers), construction and real estate, geological exploration, water services, environmental and public facility management, resident services, and other services, tend to be more production-oriented and infrastructure-intensive. Many of these sectors are involved in extracting, producing, and distributing physical goods. Based on this classification, we define two types of individual entrepreneurship: ‘entrepreneurship in highly digitisable industries’, reflecting the service-oriented first group, and ‘entrepreneurship in lightly digitisable industries’, reflecting the production and infrastructure-focused second group. These analyses suggest a potential nonlinear relationship between Internet development and entrepreneurship. The estimates are presented in Table 9, with models 1 and 2 showing TWFE estimates and models 3 and 4 displaying IV estimates. Results from Table 9 indicate an inverse U-shaped relationship between Internet development and the likelihood of individuals engaging in entrepreneurship. This suggests that while initial advancements in Internet development tend to increase the likelihood of entrepreneurship, further development beyond a certain threshold appears to decrease this likelihood. We explore whether the dual nature of Internet development, acting as both a facilitator of increased resource accessibility and a creator of market dominance, significantly influences entrepreneurial dynamics. In Table 10, models 1–2 display the impacts of Internet development on entrepreneurship in the highly and lightly digitisable industries, respectively. Model 1 indicates a negative relationship between Internet development and entrepreneurship in highly digitisable industries, which are relatively service-oriented and knowledge-intensive. The heavy reliance of these service-oriented industries on digital platforms for customer interaction and transactions makes them particularly susceptible to Internet development dynamics. This reliance can result in market concentration, favouring large platforms with superior technological capabilities, thereby reducing opportunities for new ventures. Model 2 suggests a stronger effect of Internet development on entrepreneurship in lightly digitisable industries. One possible explanation is that production-oriented industries, which focus on physical labour and tangible assets, transition less readily to digital platforms, thus leaving more opportunities for new businesses to emerge. These findings highlight the varying impacts of Internet development on entrepreneurship, contingent upon a sector’s compatibility with digital technologies.

Conclusions

We use panel data on individual entrepreneurship from CPFS and administrative data on new firm registration to test the causal impacts of Internet development on entrepreneurship. We construct an original index to measure the level of Internet development across the province. To address the endogeneity of regional Internet development, we employ IV and DID approaches by exploiting historical and geographical information related to telecommunication infrastructures, as well as a quasi-natural experiment of an Internet infrastructure program. We find that a one standard deviation increase in the Internet development index increases the likelihood of becoming an entrepreneur by 0.588 standard deviations. We also show that Internet development drives entrepreneurial activities by enhancing access to information and financial resources. Internet development can help entrepreneurs reduce their reliance on personal networks, such as friends and family, for financing, and instead secure funding through commercial loans. We find that the entrepreneurial effect of Internet development is larger among groups with lower self-perceived relational competence, social popularity, and social status. These results align with the broader consensus in the literature that the Internet levels the playing field and fosters economic inclusivity (Anderson et al., 1997; Castells, 2002). We also find that the development of the Internet in service-oriented and knowledge-intensive industries tends to create conditions that favour the emergence of ‘winner-takes-all’ dynamics, thereby concentrating market power in the hands of a few dominant players and posing significant challenges to new entrants. Our findings hold some implications for policymaking. In the context of significant financing challenges faced by small and medium-sized enterprises (SMEs) in China, financing innovations based on Internet development can enhance financial transparency and reduce costs (Huang, 2022; Huang & Liu, 2014). Policymakers could consider implementing strategies that support the expansion

Table: Nonlinear Relationship Between Internet Development and Individual Entrepreneurship

Outcome variable: individual entrepreneurship	(1) TWFE	(2) TWFE	(3) IV-FE	(4) IV-FE
Internet development	0.149***	0.184***	0.386***	0.546***
	(0.045)	(0.055)	(0.085)	(0.191)
Internet development squared	−0.076**	−0.111***	−0.102*	−0.164***
	(0.034)	(0.039)	(0.056)	(0.063)
Covariates	No	Yes	No	Yes
Individual fixed effects	Yes	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes
N	95,378	95,378	95,378	95,378
Kleibergen-Paap rk Wald F statistic	–	–	2283.528	1407.251

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level (in parentheses). Full results are available from the authors upon request.

Table 10: Internet Development and Entrepreneurship in Highly and Lightly Digitisable Industries – IV-FE Results

Outcome Variable: individual entrepreneurship	(1) Highly digitizable industries	(2) Lightly digitizable industries
Internet development	0.048	0.341**
	(0.147)	(0.147)
Internet development × Internet development	−0.086*	−0.060
	(0.050)	(0.047)
Covariates	Yes	Yes
Individual fixed effects	Yes	Yes
Province fixed effects	Yes	Yes
Year fixed effects	Yes	Yes
N	94,983	94,983
Kleibergen-Paap rk Wald F statistic	1,404.392	1,404.392

Notes: ***p < 0.01, **p < 0.05, *p < 0.1. Standard errors are clustered at the individual level (in parentheses). Full results are available from the authors upon request.

of online lending platforms, digital payment systems, and other Internet-based financial services to improve access to funding for SMEs. Our results also bring a warning to the stakeholders regarding the level of Internet development: while the Internet expansion stimulates entrepreneurial activity, it may introduce a ‘winner-take-all’ phenomenon in highly digitisable sectors, stifling new entrepreneurial entries. This underscores the need for government policies to ensure a shared benefit across both new entrants and incumbents and encourage fair competition in digital markets. This potentially includes monitoring and dismantling monopolistic practices that inhibit entrepreneurial entries. Policies aimed at reducing entry barriers and preventing ‘winner-takes-all’ dynamics could involve supporting open standards and interoperability among digital platforms to ensure that startups can compete on equal footing.

Data availability The authors do not have permission to share data. Appendix A. Internet development index In the absence of an official comprehensive index of Internet development from China, scholars have mainly used two approaches to measure Internet development levels. The first approach is constructing a comprehensive index, although there is considerable variation in the sub-indicators selected (eg. Ren et al., 2021; Wu et al., 2021). The second method uses single indicators such as the Internet penetration rate, which approximates Internet development by measuring the proportion of the population with Internet access (Fernandes et al., 2019; Freund & Weinhold, 2002, 2004).

Table A1: CFPS Questions for Identifying Entrepreneurship

CFPS wave	Question	Answer for identifying an entrepreneur
2010	What type of work unit is this?	Self-employed (excluded rural family business)
2012	Which type of work did you do?	Individual or private business operations (excluded rural family businesses)
2014	What type of work unit is this?	–
2016	What type of work unit is this?	Private company or self-employed business or other self-employed
2018	What type of work unit is this?	–

Internet penetration rate not only gauges accessibility and usage of Internet services but, as highlighted by scholars like Torero and Von Braun (2006), provides a comprehensive overview of a region’s digital accessibility and user engagement. Furthermore, as Kenny (2002) suggests, the Internet penetration rate serves as an indicator of a country’s technological infrastructure, reflecting the development of telecommunications systems and regulatory environments conducive to Internet adoption. Following existing literature, we use Internet penetration as a proxy for the level of internet development. Instead of relying solely on the Internet penetration rate to measure Internet development, we also incorporate data on mobile Internet users and broadband Internet access users. This approach allows us to capture not only the percentage of the population with Internet access but also the actual number of users actively utilising mobile and broadband services. By integrating these metrics, we achieve a more comprehensive understanding of Internet connectivity’s scale and scope. Following the literature (Ren et al., 2021; Wu et al., 2021), we employ the entropy weight method to construct an Internet development index. This method involves the following detailed steps: Step 1: Standardisation of positive indicators:

x’_{it,j} = \frac{x_{it,j} – \min(x_{it,j})}{\max(x_{it,j}) – \min(x_{it,j})}

Where $x_{it,j}$ represents the original value and $x’_{it,j}$ the standardized value of sub-indicator $j$ in province $i$ in year $t$ , respectively.

Step 2: Calculation of Proportional Values:

To calculate proportional values, you would use the following formula:

$p_{it,j} = \frac{x’_{it,j}}{\sum_{j} x’_{it,j}}$

Step 2: Calculation of Proportional Values

This step involves calculating the fraction of the standardized value of sub-indicator $j$ in province $i$ relative to the total value of sub-indicator $j$ across all provinces in year $t$ . $T$ represents the total number of years, and $N$ is the total number of provinces.

Step 3: Entropy Measurement of Each Sub-indicator

To measure the entropy value of sub-indicator $j$ , the following formula is used:

$e_j = -\frac{\sum_{t=1}^{T} \sum_{i=1}^{N} p_{it,j} \ln(p_{it,j})}{\ln(NT)}$

Where:

$e_j$ is the entropy value of sub-indicator $j$ .
$p_{it,j}$ is the proportional value calculated in Step 2.

Step 4: Calculation of Information Entropy Redundancy

To calculate the information entropy redundancy for sub-indicator $j$ :

$d_j = 1 – e_j$

Where:

$d_j$ is the information entropy redundancy of sub-indicator $j$ .
$e_j$ is the entropy value obtained in Step 3.

Step 5: Weight Determination for Each Sub-indicator

The weight $w_j$ for each sub-indicator $j$ is calculated as:

$w_j = \frac{d_j}{\sum_{j=1}^{M} d_j}$

Where:

$M$ is the total number of sub-indicators.
$w_j$ is the weight of sub-indicator $j$ .

Step 6: Compilation of the Internet Development Index

The Internet Development Index ( $idi_{it}$ ) for province $i$ in year $t$ is compiled as:

$idi_{it} = \sum_{j=1}^{M} w_j x’_{it,j}$

Where:

$idi_{it}$ is the Internet Development Index for province $i$ in year $t$ .
$w_j$ is the weight of sub-indicator $j$ determined in Step 5.
$x’_{it,j}$ is the standardized value of sub-indicator $j$ from Step 1.

Finally, the IDI for each province in each year is computed by aggregating the weighted standardized values of all sub-indicators. Table B1 illustrates specific measures of sub-indicators related to Internet penetration and their corresponding data sources. The data on the Internet penetration rate is not available for the year 2018. Following the literature (Allison, 2003; Kauffmann et al., 2022; McMillen, 2012), we performed linear interpolation to estimate missing values within each province. We first transformed the indicators into a logarithmic scale and then transformed the estimates back to their original scale. This method ensures accurate imputation and consistency in our dataset.

Table B1: The Description of Sub-indicators of the Internet Development Index

Sub-indicator	Unit	Data source
Internet penetration rate	%	Chinese Research Data Services (CNRDS)
Mobile Internet users	In ten thousand	Chinese Research Data Services (CNRDS)
Broadband Internet users	In ten thousand	Chinese Research Data Services (CNRDS)

References

Agre, P. E. (2002). Real-time politics: The internet and the political process. The Information Society, 18(5), 311–331.
Aldrich, H., Renzulli, L. A., & Langton, N. (1998). Passing on privilege: Resources provided by self-employed parents to their self-employed children. Research in Social Stratification and Mobility, 16, 291–318.
Allison, P. D. (2003). Missing data techniques for structural equation modeling. Journal of Abnormal Psychology, 112(4), 545.
Andersen, S., & Nielsen, K. M. (2012). Ability or finances as constraints on entrepreneurship? Evidence from survival rates in a natural experiment. The Review of Financial Studies, 25(12), 3684–3710.
Anderson, R. H., Bikson, T. K., Law, S. A., & Mitchell, B. M. (1997). Universal access to e-mail: Feasibility and societal implications. Educational Media International, 34(2), 86–87.
Ardichvili, A., Cardozo, R., & Ray, S. (2003). A theory of entrepreneurial opportunity identification and development. Journal of Business Venturing, 18(1), 105–123.
Arnott, D. C., & Bridgewater, S. (2002). Internet, interaction and implications for marketing. Marketing Intelligence & Planning, 20(2), 86–95.
Asongu, S., & Moulin, B. (2016). The role of ICT in reducing information asymmetry for financial access. Research in International Business and Finance, 28, 202–213. https://doi.org/10.1016/j.ribaf.2016.04.011
Autio, E., Nambisan, S., Thomas, L. D., & Wright, M. (2018). Digital affordances, spatial affordances, and the genesis of entrepreneurial ecosystems. Strategic Entrepreneurship Journal, 12(1), 72–95.
Awaworyi Churchill, S., Hayward, M., Smyth, R., & Trinh, T.-A. (2023). Crime, community social capital and entrepreneurship: Evidence from Australian communities. Journal of Business Venturing, 38(2), Article 106291.
Awaworyi Churchill, S., Smyth, R., & Trinh, T.-A. (2023). Negative life events and entrepreneurship. Journal of Business Research, 155, Article 113443.
Baker, T., & Nelson, R. E. (2005). Creating something from nothing: Resource construction through entrepreneurial bricolage. Administrative Science Quarterly, 50(3), 329–366.
Barnett, W. A., Hu, M., & Wang, X. (2019). Does the utilization of information communication technology promote entrepreneurship: Evidence from rural China. Technological Forecasting and Social Change, 141, 12–21. https://doi.org/10.1016/j.techfore.2019.01.007
Batjargal, B. (2007). Internet entrepreneurship: Social capital, human capital, and performance of internet ventures in China. Research Policy, 36(5), 605–618.
Beck, T., Demirgüç-Kunt, A., & Honohan, P. (2009). Access to financial services: Measurement, impact, and policies. World Bank Research Observer, 24(1).
Bonte, W., Falck, O., & Heblich, S. (2009). The impact of regional age structure on entrepreneurship. Economic Geography, 85(3), 269–287.
Brush, C. G., & Cooper, S. Y. (2012). Female entrepreneurship and economic development: An international perspective. Entrepreneurship and Regional Development, 24(1–2), 1–6. https://doi.org/10.1080/08985626.2012.637340
Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company.
Bughin, J., Manyika, J., & Others. (2007). How businesses are using web 2.0: A McKinsey global survey. The McKinsey Quarterly, 9.
Buonaccorsi, J. P. (2010). Measurement error: Models, methods, and applications. Chapman and Hall/CRC.
Burt, R. S. (1992). Structural holes: The social structure of competition. Harvard University Press.
Burt, R. S. (2000). The network structure of social capital. Research in Organizational Behavior, 22, 345–423.
Cai, D., Song, Q., Ma, S., Dong, Y., & Xu, Q. (2018). The relationship between credit constraints and household entrepreneurship in China. International Review of Economics & Finance, 58, 246–258.
Callaway, B., & Sant’Anna, P. H. (2021). Difference-in-differences with multiple time periods. Journal of Econometrics, 225(2), 200–230.
Castells, M. (2002). The internet galaxy: Reflections on the internet, business, and society. USA: Oxford University Press.
Chen, J. (2023). From cold war geopolitics to the crisis of global capitalism: The history of Chinese wireless network infrastructures (1987–2020). The China Quarterly, 255, 644–662.
Chen, J. J., Cui, C., Hunt, R. A., & Li, L. S.-Z. (2020). External enablement of new venture creation: An exploratory, query-driven assessment of China’s high-speed rail expansion. Journal of Business Venturing, 35(6), Article 106046.
Chen, M.-H., Chang, Y.-Y., & Lee, C.-Y. (2015). Creative entrepreneurs’ guanxi networks and success: Information and resource. Journal of Business Research, 68(4), 900–905. https://doi.org/10.1016/j.jbusres.2014.11.049
Chen, Y., Wang, H., Cheng, Z., & Smyth, R. (2023). Early-life experience of land reform and entrepreneurship. China Economic Review, 79, Article 101966.
Cheng, Z., Guo, W., Hayward, M., Smyth, R., & Wang, H. (2021). Childhood adversity and the propensity for entrepreneurship: A quasi-experimental study of the great Chinese famine. Journal of Business Venturing, 36(1), 106063. https://doi.org/10.1016/j.jbusvent.2020.106063
Cheng, Z., Tani, M., & Wang, H. (2021). Energy poverty and entrepreneurship. Energy Economics, 102, 105469. https://doi.org/10.1016/j.eneco.2021.105469
Compaine, B. M. (2001). Re-examining the digital divide.
Conley, T. G., Hansen, C. B., & Rossi, P. E. (2012). Plausibly exogenous. Review of Economics and Statistics, 94(1), 260–272.
Danet, B., & Herring, S. C. (2007). The multilingual internet: Language, culture, and communication online. Oxford University Press.
Davidsson, P. (2015). Entrepreneurial opportunities and the entrepreneurship nexus: A re-conceptualization. Journal of Business Venturing, 30(5), 674–695.
Davidsson, P., Recker, J., & Von Briel, F. (2020). External enablement of new venture creation: A framework. Academy of Management Perspectives, 34(3), 311–332.
Demirgüç-Kunt, A., & Singer, D. (2017). Financial inclusion and inclusive growth: A review of recent empirical evidence. World Bank Policy Research Working Paper, 8040.
Dettling, L. J. (2017). Broadband in the labor market: The impact of residential high-speed internet on married women’s labor force participation. ILR Review, 70(2), 451–482.
DiMaggio, P., Hargittai, E., Neuman, W. R., & Robinson, J. P. (2001). Social implications of the internet. Annual Review of Sociology, 27(1), 307–336.
Doepke, M., & Zilibotti, F. (2014). Culture, entrepreneurship, and growth. In Handbook of Economic Growth (Vol. 2, pp. 1–48). Elsevier.
Drahokoupil, J., & Fabo, B. (2016). The platform economy and the disruption of the employment relationship. ETUI Research Paper-Policy Brief, 5.
Ekpe, I., Mat, N., & Che Razak, R. (2011). Attributes, environment factors and women entrepreneurial activity: A literature review. Asian Social Science, 7(9). https://doi.org/10.5539/ass.v7n9p124
Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook “friends:” Social capital and college students’ use of online social network sites. Journal of Computer-Mediated Communication, 12(4), 1143–1168.
Ellison, N. B., Steinfield, C., & Lampe, C. (2011). Connection strategies: Social capital implications of Facebook-enabled communication practices. New Media & Society, 13(6), 873–892.
Evans, D., & Schmalensee, R. (2005). The industrial organization of markets with two-sided platforms. National Bureau of Economic Research.
Evans, P. B., & Wurster, T. S. (1997). Strategy and the new economics of information. Harvard Business Review, 75(5), 70–83.
Farrell, H. (2012). The consequences of the internet for politics. Annual Review of Political Science, 15, 35–52.
Favoretto, C., de Mendes, G. H. S., Filho, M. G., Gouvea de Oliveira, M., & Ganga, G. M. D. (2022). Digital transformation of business model in manufacturing companies: Challenges and research agenda. Journal of Business & Industrial Marketing, 37(4), 748–767.
Fernandes, A. M., Mattoo, A., Nguyen, H., & Schiffbauer, M. (2019). The internet and Chinese exports in the pre-ali baba era. Journal of Development Economics, 138, 57–76.
Field, E., Jayachandran, S., & Pande, R. (2010). Do traditional institutions constrain female entrepreneurship? A Field experiment on business training in India. The American Economic Review, 100(2), 125–129. https://doi.org/10.1257/aer.100.2.125
Fischer, E. M., Reuber, A. R., & Dyke, L. S. (1993). A theoretical overview and extension of research on sex, gender, and entrepreneurship. Journal of Business Venturing, 8(2), 151–168.
Frese, M., & Gielnik, M. M. (2014). The Psychology of Entrepreneurship. Annual Review of Organizational Psychology and Organizational Behavior, 1(1), 413–438.
Freund, C., & Weinhold, D. (2002). The internet and international trade in services. American Economic Review, 92(2), 236–240.
Freund, C., & Weinhold, D. (2004). The effect of the internet on international trade. Journal of International Economics, 62(1), 171–189.
Gentile, T. A. R., Reina, R., De Nito, E., Bizjak, D., & Canonico, P. (2020). E-learning design and entrepreneurship in three European universities. International Journal of Entrepreneurial Behavior & Research, 26(7), 1547–1566.
Gomber, P., Koch, J.-A., & Siering, M. (2017). Digital finance and FinTech: Current research and future research directions. Journal of Business Economics, 87, 537–580.
Goodman-Bacon, A. (2021). Difference-in-differences with variation in treatment timing. Journal of Econometrics, 225(2), 254–277.
Guo, F., Kong, S. T., & Wang, J. (2016). General patterns and regional disparity of internet finance development in China: Evidence from the Peking University internet finance development index. China Economic Journal, 9(3), 253–271.
Hafer, R., & Jones, G. (2015). Are entrepreneurship and cognitive skills related? Some international evidence. Small Business Economics, 44, 283–298.
Hayashi, F. (2011). Econometrics. Princeton University Press.
Hayward, M., Cheng, Z., & Wang, B. Z. (2022). Disrupted education, underdogs and the propensity for entrepreneurship: Evidence from China’s sent-down youth program. Journal of Business Research, 151, 33–39.
Hayward, M., Cheng, Z., Wang, H., & Smyth, R. (2023). Parental influence and the propensity for entrepreneurship: Evidence from the one-child policy. Journal of Business Venturing Insights, 20, Article e00428.
Hisrich, R. D., Peters, M. P., & Shepherd, D. A. (2017). Entrepreneurship. McGraw-Hill Education.
Holtzeakin, D., Joulfaian, D., & Rosen, H. S. (1994). Sticking it out—Entrepreneurial survival and liquidity constraints. Journal of Political Economy, 102(1), 53–75. https://doi.org/10.1086/261921
Huang, C., & Liu, Z. (2014). Analysis on financing difficulties for SMEs due to asymmetric information. Global Disclosure of Economics and Business, 3(1), 77–80.
Huang, S. (2022). Does FinTech improve the investment efficiency of enterprises? Evidence from China’s small and medium-sized enterprises. Economic Analysis and Policy, 74, 571–586.
Iansiti, M., & Lakhani, K. R. (2018). Managing our hub economy. Harvard Business Review, 96(1), 17.
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of social media. Business Horizons, 53(1), 59–68.
Katz, M. L., & Shapiro, C. (1985). Network externalities, competition, and compatibility. The American Economic Review, 75(3), 424–440.
Kauffmann, P. C., Takada, H. H., Terada, A. T., & Stern, J. M. (2022). Learning forecast-efficient yield curve factor decompositions with neural networks. Econometrics, 10(2), 15.
Kenney, M., & Zysman, J. (2016). The rise of the platform economy. Issues in Science and Technology, 32(3), 61.
Kenny, C. (2002). Information and communication technologies for direct poverty alleviation: Costs and benefits. Development Policy Review, 20(2), 141–157.
Kuada, J. (2009). Gender, social networks, and entrepreneurship in Ghana. Journal of African Business, 10(1), 85–103. https://doi.org/10.1080/15228910802701445
LaRose, R., Gregg, J. L., Strover, S., Straubhaar, J., & Carpenter, S. (2007). Closing the rural broadband gap: Promoting adoption of the internet in rural America. Telecommunications Policy, 31(6–7), 359–373.
Lee, D. S., McCrary, J., Moreira, M. J., & Porter, J. (2022). Valid t-ratio inference for IV. American Economic Review, 112(10), 3260–3290.
Ma, S., Wu, X., & Gan, L. (2019). Credit accessibility, institutional deficiency and entrepreneurship in China. China Economic Review, 54, 160–175. https://doi.org/10.1016/j.chieco.2018.10.015
Mack, E. A., Marie-Pierre, L., & Redican, K. (2017). Entrepreneurs’ use of internet and social media applications. Telecommunications Policy, 41(2), 120–139.
Madsen, J. B., Islam, M. R., & Doucouliagos, H. (2018). Inequality, financial development and economic growth in the OECD, 1870–2011. European Economic Review, 101, 605–624.
McMillen, D. P. (2012). Perspectives on spatial econometrics: Linear smoothing with structured models. Journal of Regional Science, 52(2), 192–209.
Michelacci, C., & Silva, O. (2007). Why so many local entrepreneurs? Review of Economics and Statistics, 89(4), 615–633. https://doi.org/10.1162/rest.89.4.615
Moro, A., Fink, M., & Maresch, D. (2015). Reduction in information asymmetry and credit access for small and medium-sized enterprises. Journal of Financial Research, 38(1), 121–143.
Muravyev, A., Talavera, O., & Schäfer, D. (2009). Entrepreneurs’ gender and financial constraints: Evidence from international data. Journal of Comparative Economics, 37(2), 270–286.
Parker, G. G., Van Alstyne, M. W., & Choudary, S. P. (2016). Platform revolution: How networked markets are transforming the economy and how to make them work for you. WW Norton & Company.
Parker, S. C. (2008). Entrepreneurship among married couples in the United States: A simultaneous probit approach. Labour Economics, 15(3), 459–481.
Pathak, S., Xavier-Oliveira, E., & Laplume, A. O. (2013). Influence of intellectual property, foreign investment, and technological adoption on technology entrepreneurship. Journal of Business Research, 66(10), 2090–2101.
Peng, M. W., & Luo, Y. (2000). Managerial ties and firm performance in a transition economy: The nature of a Micro-macro link. The Academy of Management Journal, 43(3), 486–501. https://doi.org/10.2307/1556406
Porter, D. (2013). Internet culture. Routledge.
Porter, M. E. (2001). Strategy and the Internet. Harvard Business Review.
Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. Simon and Schuster.
Ramoglou, S., & Zyglidopoulos, S. C. (2015). The constructivist view of entrepreneurial opportunities: A critical analysis. Small Business Economics, 44, 71–78.
Ren, S., Hao, Y., Xu, L., Wu, H., & Ba, N. (2021). Digitalization and energy: How does internet development affect China’s energy consumption? Energy Economics, 98, Article 105220. https://doi.org/10.1016/j.eneco.2021.105220
Shane, S. (2000). Prior knowledge and the discovery of entrepreneurial opportunities. Organization Science, 11(4), 448–469.
Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic guide to the network economy. Harvard Business Press.
Stel, A. V., Carree, M., & Thurik, R. (2005). The effect of entrepreneurial activity on national economic growth. Small Business Economics, 24, 311–321.
Tan, Y., & Li, X. (2022). The impact of internet on entrepreneurship. International Review of Economics & Finance, 77, 135–142.
Torero, M., & Von Braun, J. (2006). Information and communication technologies for development and poverty reduction: The potential of telecommunications. Intl Food Policy Res Inst.
Valliere, D., & Peterson, R. (2009). Entrepreneurship and economic growth: Evidence from emerging and developed countries. Entrepreneurship & Regional Development, 21(5–6), 459–480.
Verbeke, A., & Hutzschenreuter, T. (2021). The dark side of digital globalization. Academy of Management Perspectives, 35(4), 606–621.
Wang, J., Hu, Y., & Xiong, J. (2022). The internet use, social networks, and entrepreneurship: Evidence from China. Technology Analysis & Strategic Management, 1–15.
Webster, J. G., & Ksiazek, T. B. (2012). The dynamics of audience fragmentation: Public attention in an age of digital media. Journal of Communication, 62(1), 39–56.
Wei, X., Jiao, Y., & Growe, G. (2019). Language skills and migrant entrepreneurship: Evidence from China. Small Business Economics, 53, 981–999.
Wong, P. K., Ho, Y. P., & Autio, E. (2005). Entrepreneurship, innovation and economic growth: Evidence from GEM data. Small Business Economics, 24, 335–350.
Wu, H., Xue, Y., Hao, Y., & Ren, S. (2021). How does internet development affect energy-saving and emission reduction? Evidence from China. Energy Economics, 103, Article 105577.
Yang, X., Huang, Y., & Gao, M. (2022). Can digital financial inclusion promote female entrepreneurship? Evidence and mechanisms. The North American Journal of Economics and Finance, 63. https://doi.org/10.1016/j.najef.2022.101800
Yu, W., Zhu, F., Der Foo, M., & Wiklund, J. (2022). What does not kill you makes you stronger: Entrepreneurs’ childhood adversity, resilience, and career success. Journal of Business Research, 151, 40–55.
Zelekha, Y., Avnimelech, G., & Sharabi, E. (2014). Religious institutions and entrepreneurship. Small Business Economics, 42, 747–767.

Internet development and entrepreneurship

Internet development and entrepreneurship

Jia Guo a,b,* , Zhiming Cheng b,c , Ben Zhe Wang d

Step 2: Calculation of Proportional Values

Step 3: Entropy Measurement of Each Sub-indicator

Step 4: Calculation of Information Entropy Redundancy

Step 5: Weight Determination for Each Sub-indicator

Step 6: Compilation of the Internet Development Index

Leave a comment Cancel reply

Recent Posts

Choose Topic

Recent Comments

Social

Internet development and entrepreneurship

Jia Guo a,b,* , Zhiming Cheng b,c , Ben Zhe Wang d

Step 2: Calculation of Proportional Values

Step 3: Entropy Measurement of Each Sub-indicator

Step 4: Calculation of Information Entropy Redundancy

Step 5: Weight Determination for Each Sub-indicator

Step 6: Compilation of the Internet Development Index

Leave a comment Cancel reply

Recent Posts

Choose Topic

Recent Comments

“Namify Tech” Everything You Need to Know

“Corporate Online” Everything You Need to Know

“Corporate Entrepreneurship” Everything You Need to Know

“Sole Trader” Everything You Need to Know

Syllabus

Study Guides for every class

Short Notes

Mental Health