Digital Skills and Entrepreneurship in Mexico: Evidence from Probit Models and Implications for Digital Inclusion Policy

Abstract

This study examines the relationship between digital skills and entrepreneurial intention in Mexico, emphasizing demographic, educational, and technological determinants. Despite the recognized importance of digitalization, most Mexican entrepreneurs possess only basic competencies, which constrains productivity and growth. Using data from the Digital Skills Profiler (50,582 individuals), binary probit models were estimated to assess the effect of digital skills on both current and prospective entrepreneurs. Results reveal a paradox: individuals with advanced digital skills are less frequently engaged in entrepreneurship, often opting instead for better paid and more stable employment in the formal labor market. When engaging in entrepreneurship, individuals with advanced digital skills tend to concentrate in service sector or non-conventional activities, exhibiting weaker connections to trade in goods. Women and older generations face greater barriers to acquiring digital competencies, whereas younger cohorts show stronger skills that do not necessarily translate into opportunity-driven ventures. Necessity-driven entrepreneurship predominates, with only a small fraction of ventures evolving into opportunity-based projects. The findings highlight the need for differentiated policy approaches: fostering innovative, competitive entrepreneurship requires distinct instruments from those designed to support subsistence ventures. Aligning digital inclusion and entrepreneurship strategies with Sustainable Development Goals 8 and 9 will be crucial to narrowing digital divides and promoting sustainable, inclusive growth.

1. Introduction

In today’s technology-driven global economy, there appears to be a consensus regarding the benefits of digital technologies in terms of productivity, operational efficiency, and competitiveness, particularly for entrepreneurs starting new micro, small, and medium-sized enterprises (MSMEs), which generally face numerous resource constraints. For example, digital technologies can lower costs by streamlining processes, facilitating access to analytical data and social networks, and improving internal and external communication [1]. Social media platforms, email communication, and digital marketing strategies could potentially be effective ways to engage with customers, strengthening relationships with them [2].

Furthermore, digital technologies stimulate the acquisition of new skills, competencies, and knowledge, all of which contribute not only to improved decision-making but also to enhanced innovation and organizational adaptability in dynamic environments [3]. (At a slightly more advanced level, the use of digital technologies such as automated systems, data analytics, and cloud computing is recognized as vital for supporting business processes and enhancing operational efficiency [4]). Consequently, integrating digital technologies into business strategy is increasingly seen as essential for long-term success.

However, while basic digital technologies can be highly useful for improving promotional efforts, customer contact, or post-sale services, there remains an urgent need to adopt advanced technologies in order to achieve a genuine digital transformation that can render many microenterprises truly competitive. (In this context, Customer Relationship Management (CRM) systems play an essential role in enhancing these interactions).

Nevertheless, evidence consistently shows a pronounced technological divide: most microentrepreneurs operate with minimal competencies, whereas managers of medium-sized firms display more advanced digital skills. This divergence translates into more effective digitalization of internal processes, the clear definition of digital roles, and employee participation in transformation efforts [5].

At the regional level, it is acknowledged that the digital economy, innovation, and data-driven entrepreneurship are key components in transforming the development model of Latin American and Caribbean countries. (Equally important is promoting digital transformation across strategic sectors—such as agriculture, health, manufacturing, commerce, tourism, and mobility—to enhance innovation, productivity, and competitiveness, while contributing to the reduction in environmental impacts [6]). These findings highlight the importance of understanding how entrepreneurs adopt, integrate, and leverage digital technologies, thereby contributing—albeit indirectly—to the reduction in structural disparities in technological capacity. Nevertheless, it remains unclear whether higher levels of digital proficiency in individuals necessarily lead to greater interest in entrepreneurial activity, or whether entrepreneurs operate under enabling conditions and have the right incentives to strengthen their digital skills. This represents a significant challenge for the promotion of digitally native microenterprises and imposes additional costs on the modernization of those that are not.

Based on the foregoing analysis, we propose the following hypotheses for the case of Mexico: while access to multiple digital tools may be positively associated with the level of digital proficiency among prospective entrepreneurs, possessing only basic digital skills tends to increase the likelihood of necessity-driven entrepreneurship. However, the possession of advanced digital skills does not appear to foster the exploitation of opportunity-based entrepreneurship.

In addition to examining the determinants of entrepreneurial activity, this study contributes to a broader understanding of how digital innovation can support sustainable and inclusive economic development. The findings highlight how the distribution of digital skills among microentrepreneurs reflects structural inequalities that must be addressed through strategic digital inclusion policies. In line with the goals of integrating digital innovation into business strategy, this research provides insights into how digital proficiency—when fostered systematically—can enhance business resilience, improve long-term sustainability outcomes, and support progress toward the Sustainable Development Goals (SDGs), especially Goals 8 and 9.

2. Theoretical and Policy Background

2.1. Digital Literacy and Technologies: Concepts and Evolutionary Trends

To understand the rationale behind Mexico’s digitalization policies targeting MSMEs, it is essential to consider the evolution of digital literacy over the past two decades and to specify the main concepts involved. The concept of information and communication technologies (ICT) is broad and dynamic, with evolving boundaries and shifting emphases over time. This category includes what are now commonly referred to as digital technologies.

Many tools and systems once regarded as “new information technologies” have now reached technological maturity and become widely adopted across sectors. In this context, digital literacy has undergone a significant transformation, moving from basic operational competencies to more complex and strategic forms of engagement with digital environments.

In 2000, digital literacy was largely defined in terms of operational skills that are now considered basic: using a desktop computer, operating standard office software, browsing the internet, using email, typing on a QWERTY keyboard, and handling storage devices. These abilities were adequate for routine work-related and personal activities. For microentrepreneurs, this translated into tasks such as preparing spreadsheets, creating promotional materials, and maintaining email communication with suppliers or clients. Occasionally, a website could be developed—typically with external assistance. The above could have been considered sufficient to remain competitive a quarter of a century ago.

By 2025, the scope of digital skills has expanded considerably to include a broader set of competencies necessary for everyday digital engagement, such as the proficient use of smartphones and apps, assessment of online content, mobile banking, cybersecurity, and interaction with AI and e-government platforms. This broader skillset has important implications across sectors. Specifically, the microenterprise sector demands more sophisticated digital skills: managing social media for commercial purposes; using e-commerce platforms; operating point-of-sale systems; issuing digital invoices; utilizing cloud-based productivity tools; applying data analytics; search engine optimization; and digital marketing strategies. Modern digital literacy, therefore, encompasses not only operational proficiency but also cognitive, strategic, and evaluative capabilities [7].

Digital technologies can be classified into three categories (basic, advanced, and frontier), distinguished by the degree of organizational change their adoption entails. Basic technologies (e.g., email, social media, online banking, websites, digital interaction with government, communication-oriented e-commerce) are consolidated tools that require minimal organizational adjustment. At the opposite extreme, frontier technologies (e.g., big-data analytics, artificial intelligence, blockchain, programming tools and environments, Internet of Things) are highly dynamic; their adoption requires strategic planning and a reassessment of business models and competitive strategies. Advanced technologies (e.g., ERP and CRM systems, data-storage solutions, transactional e-commerce, cloud computing) occupy an intermediate position, demanding broader adjustments than basic tools but without the profound reorganizations of frontier innovations. This technological typology serves as a conceptual framework and, in this study, is translated into levels of digital skills [8], though its direct application to our dataset entails limitations (this approach is preferred over other classifications based on task performance [9], both because it is more consistent with our data collection instrument and because many of the tasks in alternative classifications have already become widely familiar among the population).

The empirical data reveal that 90.7% of both Mexican entrepreneurs and individuals interested in starting a micro or small business do not possess advanced or frontier digital skills, while 3.1% report lacking any digital skills at all. This suggests that the overall level of digital skills is both low and highly uneven. To mitigate a pronounced imbalance in the dataset, two modifications were introduced. First, the original basic category of technologies is now categorized in terms of skills and divided into foundational digital literacy—hereafter basic skills—(email, social media, online banking, communication-oriented e-commerce) and intermediate operational skills (websites, digital interaction with government, transactional e-commerce). Second, the advanced and frontier categories are now combined due to the extremely small number of observations in the latter, which makes separate statistical treatment infeasible.

2.2. Digital Transformation: Determinants and Entrepreneurial Implications

The push towards digital transformation is driven by the recognition that emerging technologies have substantial potential to foster innovation and provide a competitive edge for enterprises [10]. In recent years, this premise has generated an extensive body of scholarly literature aimed at conceptualizing digital transformation and identifying its key determinants.

Digital transformation refers to a profound and accelerating shift within organizations, characterized by the integration of advanced digital technologies and the convergence of physical and digital systems. This process entails changes in business models, processes, and organizational structures aimed at enhancing innovation and competitiveness [11,12,13].

Previous research has examined various aspects of digital transformation, including the integration of digital technologies into business processes, the reconfiguration of organizational structures, and the implications for overall business performance [14]. Findings in this field suggest that successful digital transformation is often contingent upon several critical factors, such as leadership commitment, organizational culture, and the alignment of digital strategies with business objectives. Moreover, other frameworks emphasize the importance for firms to navigate and leverage digital disruption [15]. In this respect, it cannot be ruled out that the degree of digitalization adopted may be influenced by the nature of the entrepreneurial venture.

Empirical evidence suggests that entrepreneurs’ adaptability to digital transformation is shaped not only by their initial digital capabilities and the characteristics of the ecosystem in which they operate, but also by the size and nature of the entrepreneurship or firm—especially in emerging economies.

Two main types of entrepreneurships are generally identified in the literature, particularly in the typology developed by the Global Entrepreneurship Monitor (GEM).

• Necessity-driven (push) entrepreneurship, which emerges in response to adverse conditions such as unemployment or low income, often with limited digital adoption due to resource constraints;
• Opportunity-driven (pull) entrepreneurship, in which ventures are created to exploit specific market opportunities and tend to display higher levels of strategic engagement with digital tools.

Understanding the relationship between these entrepreneurial typologies and the adoption of digital technologies is essential for designing policies and strategies that promote inclusive digital transformation, especially in contexts where disparities in digital readiness persist. Building on this typological distinction, it is important to explore the broader set of determinants that shape not only entrepreneurs’ ability, but also their willingness to engage with digital transformation.

Digital transformation is often assumed to expand entrepreneurial opportunities by reconfiguring costs, enabling scalability, and fostering platform-based business models [16,17]. Yet our findings challenge this canonical expectation: digital capabilities do not automatically translate into entrepreneurial entry. In contexts marked by low ecosystem maturity and scarce complementary assets, high-skill individuals may rationally redirect their digital expertise toward formal employment rather than early-stage venturing [18]. This contingent outcome refines the theoretical understanding of digital entrepreneurship, emphasizing that technological affordances and platform logics generate opportunities only when supported by adequate institutional, market, and policy conditions. Building on this contingent view, we next consider how personal, organizational, and systemic factors shape entrepreneurs’ digital engagement.

A closer look at these determinants reveals that entrepreneurs’ digital engagement is influenced by a combination of personal, organizational, and systemic factors —such as entrepreneurial training, awareness of digital technology as a strategic tool for competitiveness and a culture of lawfulness. Financial resources and credit availability are also critical enablers [19]. According to some authors, these factors collectively contribute to the development of managerial competencies and foster openness to innovation [20].

These personal, organizational, and systemic factors are embedded within broader systemic conditions, particularly the institutional and policy environments that either enable or hinder digital adoption. At the ecosystem level, entrepreneurship thrives where private stakeholders interact within a supportive public policy environment [21]. Policies that offer incentives, infrastructure, and education are key to promoting digital adoption, especially in emerging economies [22]. However, as previously noted, it remains unclear whether higher levels of digital proficiency in individuals necessarily translate into greater interest in entrepreneurial activity or whether other factors play a more decisive role. Thus, understanding both individual and systemic conditions is essential for advancing effective digital integration strategies, particularly for entrepreneurs and MSMEs.

Although this study does not directly assess environmental sustainability outcomes, it contributes to the literature on sustainable business development by emphasizing how the strategic integration of digital skills within entrepreneurship ecosystems can support more inclusive, resilient, and sustainable economic structures. This approach aligns with global sustainability frameworks—particularly SDGs 8 and 9—which underscore the role of inclusive economic growth, innovation, and digital infrastructure as catalysts for long-term transformation.

2.3. Key Elements of Mexico’s Digitalization Policy

The available evidence indicates pronounced cross-country disparities in the adoption of digital technologies by entrepreneurs and MSMEs, as measured by the Digital Maturity Index (DMI) [23]. According to the source, the regional average DMI for Latin America is 47%, although levels of digital maturity vary substantially among countries. Chile (98%) and Costa Rica (81%) exhibit considerably higher levels of digital maturity, whereas others, such as Nicaragua (8%) and Guatemala (17%), remain markedly behind.

Mexico, with a DMI score of 41%, falls below the regional average. This outcome may be attributable to structural characteristics of the Mexican MSME sector, including persistently low levels of financial inclusion, high entry costs for new enterprises, modest adoption of digital tools for commercial purposes, and limited public initiatives and resources specifically targeting MSMEs, which together constrain the country’s progress toward higher digital maturity.

At the national level, approximately 99.8% of businesses in Mexico are classified as MSMEs. These enterprises contribute around 52% to the national GDP and account for approximately 68% of total employment [24]. The sector is predominantly composed of microenterprises—firms employing fewer than 10 individuals—constituting 55.6%, with small firms representing 37.3%, and medium-sized enterprises making up only 7.1% [25]. In terms of economic activities, approximately 52% of MSMEs operate in retail trade, 40% in services, and just 2% in manufacturing. Moreover, recent estimates suggest that approximately 81% of early-stage entrepreneurs in Mexico start their business due to a lack of alternative job opportunities, indicating a prevalence of necessity-driven entrepreneurship [26].

Consequently, public strategies aimed at enhancing the digital capacities of microenterprises have been conceived not only as instruments of economic policy, but also as mechanisms for reducing structural inequalities [27].

Business-oriented digital skills development policies in Mexico, particularly those targeting entrepreneurs and MSMEs, have followed a relatively recent institutional trajectory. Arguably, the first significant effort to articulate a national vision for digital transformation was launched in 2011 through the Agenda Digital Nacional (National Digital Agenda) [28]. This strategy, developed collaboratively by the federal government and a coalition of industry, academic and civil society organizations known as Alianza ADN, was designed to identify long-term public policy proposals that would promote innovation and enhance Mexico’s competitiveness through the use of ICT, including the internet and broadband. Its key objectives included expanding internet coverage—particularly in rural and underserved areas—improving connectivity and network quality, fostering a favorable environment for technology startups, encouraging the digitalization of entrepreneurs and MSMEs, and promoting digital skills through training initiatives and formal education programs [6,29].

This initial effort was followed by the 2013–2018 Estrategia Nacional Digital (National Digital Strategy), which consolidated digital policy by promoting the development of the ICT industry, stimulating e-commerce, and fostering innovation among entrepreneurs and MSMEs through public procurement mechanisms and targeted programs.

The 2013 agenda proposed a stronger emphasis on leveraging digital technologies to drive economic growth and innovation, and it also focused more intensively on digital inclusion to ensure that marginalized communities had better access to technology [30]. This was the first-time specific programs and resources were explicitly designed to support digital inclusion. Most of them focused on facilitating the acquisition of basic technological equipment for entrepreneurs and MSMEs, including computers, tablets, and modems; providing training in basic digital skills and e-commerce. In addition, the Programa para el Desarrollo de Software y la Innovación (Program to Develop Software Industry and Innovation) assigned resources to support innovation projects for ICT enterprises.

In 2019, digitalization support for MSMEs was decoupled from the broader National Digital Strategy. A new initiative—the “Policy to Promote MSMEs”—was launched by the Ministry of Economy through the Productive Development Unit. Among other measures, this policy positioned digitalization as a cross-cutting axis to be embedded in all processes involving this segment of the business sector and also assumed a coordinating role, aligning initiatives from public institutions, non-governmental organizations, and the private sector. More recently, in 2025, the Agencia de Transformación Digital y Telecomunicaciones (Digital Transformation and Telecommunications Agency) was established as a ministerial authority within the Executive Branch. It plays a pivotal role in various strategic initiatives that extend beyond connectivity and the digital transformation of microenterprises, encompassing areas such as semiconductors and the aerospace industry.

Nevertheless, despite the growing policy efforts outlined above, a large proportion of businesses—particularly microenterprises—still operate with minimal technological resources, limited access to digital infrastructure, and low or none digital literacy. These structural limitations hinder their ability to integrate into more dynamic segments of the economy, adapt to innovation processes, and compete in increasingly digitalized markets. Given this context, the implementation of targeted digital policies has become essential in the pursuit of enhanced competitiveness, productivity, and inclusion.

3. Materials and Methods

3.1. Data Description

This study draws on data from the Digital Skills Profiler developed by Mexico’s Ministry of Economy. This instrument collected entrepreneurs’ self-perceptions regarding their digital competencies, including access to computers and internet connectivity, as well as the use of social media encompassing both personal and commercial activities. Its primary objective was to classify individuals according to their digital profiles in order to align them with relevant support programs—such as training and financial assistance—in 2020 and 2021.

In addition to assessing digital competencies, the dataset also contains demographic and educational information on Mexican entrepreneurs and prospective entrepreneurs. It comprises 50,582 respondents, providing a robust foundation for the empirical analysis. The sample consists of 57.9% women and 42.1% men. In terms of age, 33.4% of participants were 30 years old or younger, 64.1% were between 31 and 60, and the remaining individuals were over 60. Regarding educational attainment, 28.3% had completed at most secondary education, while 44.4% reported a high school diploma as their highest level. Appendix A 

Table A1. Benchmarking of the DSP sample against INEGI labor force statistics (2020–2021).

Dimension	Category	DSP (%)	INEGI (%)
Sex	Women	57.9	56.3
	Men	42.1	43.7
Age	≤30	33.4	35.1
	31–60	64.1	55.8
	>60	2.5	9.1
Educational attainment	Secondary	28.3	39.2
	High school	44.4	37.4

Notes: DSP: Digital Skills Profiler. INEGI: Official National Statistics in Mexico. Models are unweighted. Source: Authors’ own elaboration.

provides a benchmark comparison of the sample against national statistics from the National Institute of Statistics and Geography (INEGI).

It is particularly important to note that the distribution of digital competencies among entrepreneurs and prospective entrepreneurs reveals significant disparities. Approximately 3.1% of respondents report having no digital skills at all, 47.3% possess only basic skills, 43.4% demonstrate intermediate skills, and just 6.2% exhibit advanced or frontier skills. These figures point to a technological divide and underscore persistent inequalities in digital readiness across the entrepreneurial population in Mexico.

3.2. Variable Construction and Measurement

Using information from the dataset introduced above, we constructed the variables required for the econometric analysis.

Table 1. Description of variables.

Variable (Code)	Theoretical and Empirical Justification
Sex (sex)	The literature has documented gender gaps in the acquisition and use of digital skills. Women tend to face greater barriers in terms of access and technological self-confidence [31,32].
Age (age)	Digital competencies are generally lower among older generations, largely due to differences in technological exposure (i.e., digital natives vs. digital migrants) [33].
Educational attainment (edu)	It is common to assume a strong correlation between educational level and digital skills, as formal education facilitates technology appropriation and increases the likelihood of further training [34].
Type of economic activity (biz-type)	Certain sectors (e.g., e-commerce or digital services) demand a higher degree of digitalization, while others (such as informal trade or manual labor) operate with lower technological requirements.
Time in operation (firm-age)	This variable accounts for the life cycle of the enterprise, under the assumption that older businesses may have developed greater organizational structure and resilience, whereas newer firms may still be in early stages of adaptation [35,36].
Computer use (comp-use)	Frequent computer use implies continuous exposure to digital tools and learning-by-doing processes.
Internet use (net-use)	Regular access to and use of the internet is a necessary condition for acquiring more complex digital skills.
Accounting record-keeping (acct_r)	This variable reflects a formalized business practice. It is associated with more structured processes, the adoption of management technologies, and likely greater exposure to digital tools.
Entrepreneurial motivation (biz-mot)	This variable captures the structural drivers of business creation. Necessity-driven entrepreneurship is often linked to labor market exclusion, while opportunity-driven ventures reflect proactive strategies and innovation [37]. Including it enables a more nuanced interpretation of the relationship between digital skills, entrepreneurial intent, and individual agency.
Digital skills (dig-sk)	This variable refers to the ability to use digital technologies effectively for information processing, communication, problem-solving, and content creation [29,31].
Access to digital tools (dig-tl-ac-i)	Including a variable on access to key digital tools is crucial to understanding conditions enabling digital skills development. In unequal contexts, limited access is a barrier to digital inclusion [38]. Regular use also builds familiarity and cognitive competencies for advanced tasks [7,31], especially for micro-entrepreneurs in emerging economies.

Source: Authors’ own elaboration.

presents the explanatory variables considered in the analysis, along with their theoretical and empirical justification.

The operational coding of the variables is as follows: The variable sex was included to examine its influence on entrepreneurial activity [39,40], coded as 0 for men and 1 for women. The continuous age variable was rescaled by dividing it by 15 to improve interpretability and avoid reporting very small marginal effects “per year”; consequently, coefficients and marginal effects are interpreted per 15-year increase, which aligns with salient life-cycle bands (≤30, 31–45, 46–60) and facilitates comparison across models. The variable edu, representing the respondent’s highest educational level, was coded accordingly [41,42].

For the type of economic activity (biz-type), the response options were (1) trade of goods, (2) provision of services, (3) production, processing, or manufacturing of goods, and (4) other activities. The continuous firm-age variable was rescaled by dividing it by 5 to avoid reporting very small marginal effects “per year.” Consequently, coefficients and marginal effects are interpreted per 5-year increase [43].

Computer use (comp-use), internet use (net-use), and accounting record-keeping (acct_r) are binary indicators coded as 0 for negative responses and 1 for positive ones. The entrepreneurial motivation (biz-mot), in turn, captures the respondent’s primary reason for starting a business, based on the following response categories: (0) to secure a stable source of income for oneself or one’s family (proxying necessity), and (1) to take advantage of a business idea or opportunity previously identified (proxying opportunity) [44,45].

The extent to which the respondent possesses digital literacy is captured by the variable digital skills (dig-sk), coded as follows: (1) individuals without digital skills, (2) those with foundational literacy, (3) respondents with intermediate operational skills, and (4) those with advanced or frontier competencies. For estimation purposes, when digital skills are used as the dependent variable, categories (1) and (2) are coded as 0, and categories (3) and (4) as 1 [46]. This decision was based on two considerations. First, the extreme categories exhibited a highly asymmetric distribution, making the estimation of an ordered probit model statistically unfeasible. Second, from a substantive perspective, the study’s main interest lies in identifying the factors that explain the probability of reaching a minimum threshold of digital competence, which is a necessary condition for digital inclusion and entrepreneurial activity in contexts of inequality. This strategy is consistent with approaches adopted in the digital divide literature [33,34] and allows for a clearer interpretation of the results in terms of public policy.

Similarly, for access to digital tools (dig-tl-ac-i), was constructed from the following items: (1) smartphone, (2) internet connectivity, (3) social media platforms, (4) digital financial services, (5) computer, (6) digital payments, and (7) bank credit.

3.3. Econometric Specification: Binary Probit Models

This section examines the relationship between access to and use of specific technologies, demographic and educational characteristics, and entrepreneurial potential in Mexico. The analysis relies on a binary probit specification to assess the likelihood that individuals possess certain digital skills and entrepreneurial motivations, thereby capturing the factors and digital capacities that influence entrepreneurial development. The model seeks to identify structural and behavioral patterns that explain the uneven distribution of digital competencies among entrepreneurs in Mexico.

Given the dichotomous nature of the dependent variables—digital skills and entrepreneurial motivation—the probability of each outcome is modeled using maximum likelihood estimation, which evaluates how individual characteristics shape the likelihood of specific entrepreneurial motivations and digital skills. This approach highlights whether digital competencies and access to enabling technologies increase the probability of actual business ownership beyond entrepreneurial intention. In binary response models, the central objective is to analyze how explanatory variables affect the probability of the outcome [47]. The general specification is

$$P\left({Y_i=1|X}_i\right)=\Phi \left(X_i\beta \right)$$

(1)

where $Y_i$ is the binary variable taking the value 1 when the event occurs and 0 otherwise. The vector of independent variables $X_i$ includes a set of covariates that varies depending on the specific model estimated. The parameter vector $\beta$ represents the estimated coefficients of the model, which determine how each covariate affects the latent index $X_i\beta$. Finally, $\Phi \left(·\right)$ denotes the cumulative distribution function of the standard normal distribution. The marginal effects on the probability of the outcome are obtained as ${\displaystyle \frac{\partial P\left(Y_i=1|X_i\right)}{\partial x_{ik}}}$ = $\varphi \left(X_i\beta \right){\beta }_k$, where $\varphi (\cdot )$ denotes the standard normal density function. Reported marginal effects correspond to the sample averages across all observations.

To test the hypotheses of this study, three models were estimated, each addressing a specific analytical dimension: Model 1 examines microenterprise owners with the dependent variable defined as the possession of digital skills; Model 2 also focuses on microenterprise owners, but with the dependent variable capturing entrepreneurial motivation; and Model 3 shifts to prospective micro-entrepreneurs, where the dependent variable once again reflects the possession of digital skills. The statistical software used for these analyses was Gretl version 2025a (GNU General Public License), with probit models specified through standard commands. Scripts are available from the author upon request.

4. Results and Discussion

4.1. Preliminary Findings

This section presents the preliminary findings regarding the entrepreneurial status and digital conditions of the surveyed population. In terms of entrepreneurial status, 56.7% of the sample consisted of individuals who owned a business, while the remainder were prospective micro-entrepreneurs (these groups will be analyzed separately in the econometric exercise). Among these businesses, 61.1% had been operating for five years or less, and 86.1% had been established for fewer than ten years. Moreover, 93.8% of the businesses were engaged in one of the following three activities: trade of goods, provision of services, or the production, elaboration, or manufacturing of goods.

It is worth noting that 53.7% of businesses did not maintain accounting records. Additionally, 50.7% did not use a computer for any task, and 30.8% lacked internet access for business-related activities. Only 41.9% of those who are not entrepreneurs have five or more of the following digital tools: smartphone, computer, Wi-Fi, WhatsApp, social media, financial services, or payment options. Moreover, 10.5% of those who are not entrepreneurs, but are interested in becoming one, report not having access to any of these digital tools.

In terms of ordinal associations among the variables (Kendall’s tau-b), age and length of business operation exhibit a positive relationship (${\tau }_b=0.18$), suggesting that businesses with a longer operational history are generally managed by older individuals—an expected pattern in family-run microenterprises. However, older individuals also tend to display lower levels of digital skills $\left({\tau }_b=-0.18\right)$.

Furthermore, education emerges as a key factor, showing positive associations with the level of digital skills $\left({\tau }_b=0.39\right)$, and most notably with access to computers $\left({\tau }_b=0.37\right)$, internet connectivity $({\tau }_b=0.23)$ and financial services $\left({\tau }_b=0.22\right)$. A positive relationship also exists between the number of available digital tools and the level of education $({\tau }_b=0.27)$.

With respect to business practices, internet access appears to facilitate the maintenance of accounting records $({\tau }_b=0.20)$. Notably the variable the sex shows no association with any other variable.

Complementing the above findings, possessing digital skills is clearly positively related to access to all digital tools, including internet access $({\tau }_b=0.29)$. Finally, access to each digital tool is positively correlated with access to the others, suggesting that these tools are complementary rather than substitutive in nature. This pattern aligns with the conceptualization of digital tools as interconnected components within a broader digital ecosystem, which facilitates new forms of production, commercialization, and enterprise management—especially among micro-entrepreneurs and informal workers [16].

4.2. Econometric Results

This section presents the estimation results of models that examine different aspects of the relationship between entrepreneurial intention and digital skill levels in Mexico, using a standard set of demographics, education, and technology indicators. The choice of explanatory variables is guided by theoretical and empirical literature on digital inclusion and human capital formation. The results are summarized in

Table 2. Probit Model 1: Estimated Coefficients and Marginal Effects.

Variable	$\mathit{\beta }$ Full (SE)			$\mathit{\beta }$ Parsimonious (SE)			Average Marginal Effect (SE)
const	−1.8594	(0.051)	***	−1.8073	(0.043)	***
sex	−0.1306	(0.015)	***	−0.1279	(0.015)	***	−0.0397	(0.005)
age	−0.2469	(0.010)	***	−0.2473	(0.010)	***	−0.0768	(0.003)
edu	0.4410	(0.009)	***	0.4419	(0.009)	***	0.1362	(0.003)
firm-age	−0.0309	(0.006)	***	−0.0302	(0.006)	***	−0.0094	(0.002)
acct_r	0.0452	(0.017)	***	-	-		-
biz-type-1	0.0515	(0.030)	*	-	-		-
biz-type-2	0.2210	(0.032)	***	0.1787	(0.017)	***	0.0555	(0.005)
biz-type-3	0.0343	(0.047)		-	-		-
biz-type-4	NA			-	-		-
comp-use	0.6437	(0.017)	***	0.6494	(0.017)	***	0.2015	(0.005)
net-use	0.4105	(0.019)	***	0.4127	(0.019)	***	0.1281	(0.006)
biz-mot	−0.2062	(0.022)	***	−0.1829	(0.020)	***	−0.0568	(0.006)

Notes. Significance levels are indicated next to standard errors: * $p<0.10$, *** $p<0.01$. McFadden $pseudo-R^2=0.2098$. Correct classification rate of 72.7%. Likelihood-ratio test: χ²(8) = 10,485.3, $p\text{-}\mathrm{value}<0.001$. NA indicates that the corresponding coefficient is not available due to data constraints in that category. Source: Authors’ own elaboration.

,

Table 3. Probit Model 2: Estimated Coefficients and Marginal Effects.

Variable	$\mathit{\beta }$ Full (SE)			$\mathit{\beta }$ Parsimonious (SE)			Average Marginal Effect (SE)
const	−6.5348	(17.25)		−1.5829	(0.038)	***
sex	−0.0735	(0.021)	***	−0.1320	(0.017)	***	−0.0284	(0.004)
age	0.0161	(0.014)		0.0320	(0.012)	***	0.0068	(0.003)
edu	0.0260	(0.024)		0.1153	(0.021)	***	0.0262	(0.005)
firm-age	0.0663	(0.008)	***	0.0861	(0.006)	***	0.0189	(0.001)
acct_r	5.5851	(17.25)		-	-		-
biz-type-2	0.0347	(0.025)		0.0578	(0.020)	***	0.0129	(0.004)
biz-type-3	0.0853	(0.032)	***	-	-		-
biz-type-4	0.2501	(0.038)	***	0.3068	(0.031)	***	0.0673	(0.007)
dig-sk	−0.2395	(0.023)	***	−0.1327	(0.014)	***	−0.0439	(0.004)
comp-use	0.6704	(0.027)	***	0.7916	(0.022)	***	0.1762	(0.005)
net-use	0.1732	(0.031)	***	0.2614	(0.025)	***	0.0577	(0.005)

Notes. Significance levels are indicated next to standard errors: *** $p<0.01$. McFadden $pseudo-R^2=0.1052$. Correct classification rate of 83.7%. Likelihood-ratio test: χ²(9) = 3378.0, $p\text{-}\mathrm{value}<0.001$. Source: Authors’ own elaboration.

and

Table 4. Probit Model 3: Estimated Coefficients and Marginal Effects.

Variable	$\mathit{\beta }$ Full/Parsimonious (SE)			Average Marginal Effect (SE)
const	−1.6994	(0.078)	***
sex	−0.0794	(0.028)	***	−0.0226	(0.008)
age	−0.3136	(0.019)	***	−0.0892	(0.005)
edu	0.4764	(0.017)	***	0.1356	(0.005)
dig-tl-ac-1	0.2549	(0.036)	***	0.0725	(0.010)
dig-tl-ac-2	0.1120	(0.032)	***	0.0319	(0.009)
dig-tl-ac-3	−0.1412	(0.039)	***	−0.0402	(0.011)
dig-tl-ac-4	0.4305	(0.035)	***	0.1225	(0.010)
dig-tl-ac-5	0.2167	(0.034)	***	0.0617	(0.009)
dig-tl-ac-6	0.5953	(0.032)	***	0.1694	(0.010)
dig-tl-ac-7	0.1853	(0.034)	***	0.0527	(0.008)

Notes. Significance levels are indicated next to standard errors: *** $p<0.01$. McFadden $pseudo-R^2=0.2594$. Correct classification rate of 76.2%. Likelihood-ratio test: χ²(10) = 4035.3, $p\text{-}\mathrm{value}<0.001$. Source: Authors’ own elaboration.

, which report the full specification (all covariates) and, where applicable, a parsimonious specification.

The binary probit Model 1 presented in Table 2, includes 36,152 valid observations corresponding to microenterprise owners. The dependent variable captures the possession of digital skills, coded as 0 for individuals with no or only foundational digital skills, and 1 for those with intermediate or superior digital skills.

Based on the marginal effects, the likelihood of possessing higher digital skills is negatively associated with being female (−4.0%) (All % for AME denote percentage points (absolute changes in probability)), older age (−7.7% per 15-year increment), longer business operation (−0.9% per 5-year increment), and opportunity-driven entrepreneurship (−5.7%). Conversely, higher educational attainment (13.6%), operating in the service sector (5.6%), regular use of computer (20.2%), and internet use (12.8%) significantly increase the predicted probability of digital skill possession.

All variables display the expected sign, with two notable exceptions. First, opportunity-driven entrepreneurship is negatively associated with digital skill possession, suggesting that individuals motivated by opportunity do not necessarily exhibit higher digital competencies. Alternatively, this result may reflect a lack of incentives for digitally skilled individuals to engage in entrepreneurial activities, even when pursuing market opportunities. Second, owners of firms engaged in the production, processing, or manufacturing of goods (biz-type-3) show no statistically significant differences from those involved in the trade of goods in terms of their digital skills, whereas the owners of service-oriented firms exhibit higher levels of such skills.

In addition, Table 3 presents the results of binary probit Model 2, which also includes 36,152 valid observations corresponding to microenterprise owners. The dependent variable captures the entrepreneurial motivation, coded as 0 for necessity-driven entrepreneurship, and 1 for those with opportunity-driven nature.

The likelihood of exhibiting opportunity-driven motivation is negatively associated with being female (−2.8%) and with the level of digital skills (−4.4%). This counterintuitive sign may be explained by the fact that individuals with advanced digital skills often prefer employment in the traditional labor market. In the Mexican context, entrepreneurship can be interpreted as a rational response to structural labor market incentives, even if the aggregate outcome appears paradoxical relative to canonical expectations. Specifically, high payroll taxes and contributory systems reduce the appeal of formality, while entry barriers and weak enforcement discourage formal entrepreneurial activity [48]. Accordingly, even skilled individuals may rationally choose informal or necessity-based entrepreneurship when expected returns are higher outside formality. Interpreting advanced skills through this rational choice lens underscores the need for policies that enhance the relative benefits of formal entrepreneurship [49,50].

In contrast, age (0.7% per 15-year increment), longer business operation (1.9% per 5-year increment), and engaging in non-conventional business activities (6.7%) significantly increase the predicted probability of being opportunity-driven. Likewise, each additional year of education (2.6%), operating in the service sector (1.3%), regular computer use (17.6%) and internet use (5.8%) are positively associated with opportunity-driven entrepreneurship.

Overall, the estimated coefficients display the expected signs, with the exceptions noted above. In this case, both owner and firm age are positively associated with the likelihood of being opportunity-driven. Necessity-driven ventures tend to be more unstable and display a higher degree of substitutability with salaried employment. Therefore, it is reasonable to infer that the businesses which persist over time tend to reflect genuine market opportunities rather than necessity-driven survival. It is worth noting that the findings from Model 2 suggest that greater digital proficiency does not necessarily translate into the creation of businesses with higher productivity and added value.

Table 4 reports the estimates from Model 3, fitted to 11,473 observations on prospective micro-entrepreneurs. The dependent variable measures possession of digital skills, coded as in Model 1.

According to the previous table, higher digital skills are inversely related to being female (−2.3%) and to older age (−8.9% per 15-year increment). In contrast, higher educational attainment (13.6%) and access to a variety of digital tools, including smartphones (7.3%), digital financial services (12.3%), computers (6.2%), and digital payments (16.9%), exhibit particularly strong positive marginal effects.

However, actual use of social media shows a counterintuitive negative association (−4.0%) with the dependent variable. This result may reflect two mechanisms. First, for micro-entrepreneurs operating in contexts of low formalization, social media use may not necessarily translate into the acquisition of digital skills. Second, the social media variable may be capturing an inverse relationship, whereby individuals with lower levels of digital literacy rely more heavily on social media because it represents the most accessible option in terms of cost, usability, and low entry barriers. This interpretation is consistent with the broader finding that the effective use of some digital tools is not guaranteed—particularly when technologies have matured and their application is confined to routine, non-strategic activities.

In any case, this third model, which focuses on individuals interested in becoming entrepreneurs, highlights the critical importance of having access to digital tools in order to enhance digital skill levels.

The estimated models demonstrate good overall fit, as reported in the corresponding table, thereby reinforcing the robustness of the results. As a diagnostic, we assessed multicollinearity across the three models—paying special attention to Model 3, which includes all digital-tool access indicators. No variable exhibited problematic variance inflation factors (VIF; maximum = 1.8), confirming that multicollinearity is not a concern for the reported estimates.

Taken together, these findings indicate that both current and prospective entrepreneurs face gender-based barriers to the acquisition of digital skills, with women being particularly disadvantaged, a pattern consistent with the prevalence of necessity-driven ventures among them. Moreover, younger individuals display higher levels of digital competencies, although it does not appear that they are the ones predominantly engaging in opportunity-driven entrepreneurship.

In fact, entrepreneurial activity most often originates out of necessity and, if ventures manage to survive, gradually evolves into the exploitation of market opportunities. This suggests that the boundary between necessity-driven and opportunity-driven entrepreneurship may be dynamic [26]. Sectoral patterns are consistent: digital skills are more directly associated with service and other non-conventional activities, with a weaker link to trade in goods.

Overall, the empirical results offer clear guidance for policy design. The positive marginal effects for computer and internet use (Table 2, Table 3 and Table 4) underscore the importance of facilitating entry-level digital adoption among micro and small entrepreneurs. In contrast, the negative and significant marginal effect for advanced digital skills suggests that highly skilled individuals may face opportunity costs that discourage entrepreneurship relative to formal employment. Taken together, these findings point to differentiated policy needs: measures that lower digital adoption barriers for subsistence-oriented ventures, and instruments that mitigate early-stage risk and strengthen market access for entrepreneurs with advanced digital competencies. The evidence thus provides a direct empirical foundation for tailoring digital inclusion policies to diverse entrepreneurial trajectories.

5. Conclusions

This paper has examined how digital skills and entrepreneurial activity are related in Mexico, based on demographic, educational, and technological characteristics of entrepreneurs and prospective entrepreneurs. The evidence shows that digital skills among Mexican entrepreneurs remain largely rudimentary, which constrains productivity and growth potential across many microenterprises.

Moreover, the econometric results indicate that older generations and, to a lesser extent, women, face greater barriers to acquiring digital competencies, contributing to persistent inequalities in digital readiness. However, these disparities do not fully explain entrepreneurial outcomes, which are primarily driven by necessity rather than opportunity. Ventures that survive over time may evolve toward opportunity-based models, but this transition depends strongly on contextual factors such as sector, access to digital tools, and the ability to apply digital competencies strategically.

In addition, the analysis reveals a paradoxical aggregate outcome: individuals with higher digital skills are less likely to become entrepreneurs. This finding reflects a systemic allocation failure within Mexico’s entrepreneurial ecosystem, where size-dependent distortions and weak incentives misallocate talent and depress entrepreneurial returns [51,52]. High-skill individuals rationally shift toward formal employment—particularly in innovative firms—where their digital capabilities are better rewarded, generating a “skills drain” from early-stage entrepreneurship. As a result, many high-skill workers secure more stable and better-paid jobs rather than engage in high-risk, resource-constrained ventures. These dynamics reduce incentives to initiate new enterprises and highlight the need for instruments that revalue the entrepreneurial option for digital talent. Strengthening early-stage returns to innovation and improving market entry conditions are essential to ensure that Mexico’s entrepreneurial sector can attract and retain human capital required for long-term productivity growth.

These findings underscore the presence of structural asymmetries that should guide the design of differentiated entrepreneurship and digital inclusion policies. Current initiatives have largely focused on the acquisition of basic skills, which, while improving visibility and customer engagement, do not translate into structural competitiveness. For necessity-driven entrepreneurs, policies should prioritize foundational digital and managerial capabilities that improve operational efficiency and higher income. On the other hand, a productivity growth-driven policy should focus on creating incentives and institutional conditions that make opportunity-driven entrepreneurship a viable alternative for people with advanced digital skills that nowadays prefer formal employment—through innovation support, access to financing, and reduced market entry barriers. Such differentiated approaches could ensure that digital inclusion strategies foster not only access but also structural competitiveness and equitable outcomes.

Our findings contribute to the ongoing debate on digital innovation and entrepreneurship by challenging the canonical view that digitalization uniformly expands entrepreneurial opportunities [16,17]. In contexts characterized by informality and weak institutions, digital capabilities may yield higher expected returns in wage employment rather than in entrepreneurship. This contingent outcome underscores a structural boundary condition within existing theory: without complementary assets, market access, and ecosystem maturity, digitalization alone cannot foster new ventures [18]. Hence, the Mexican case refines prevailing assumptions and broadens theoretical understanding at the intersection of technology, innovation, and entrepreneurship [53].

From a theoretical standpoint, the study contributes to the literature on digital inclusion by showing that access to digital tools and competencies is not automatically associated with higher-quality ventures. Instead, digital skills are unevenly distributed and shaped by social, educational, and sectoral factors, reinforcing the importance of distinguishing between access, use, and effective application. The evidence also supports the view that the boundary between necessity-driven and opportunity-driven entrepreneurship is dynamic and context-dependent.

Finally, the study aligns with the broader sustainability agenda by linking digital inclusion and entrepreneurship to the Sustainable Development Goals (SDGs), particularly SDG 8 (Decent Work and Economic Growth) and SDG 9 (Industry, Innovation, and Infrastructure). Future research should expand on these findings by adopting longitudinal perspectives, conducting cross-country comparisons, and examining how emerging technologies reshape the balance between necessity- and opportunity-driven entrepreneurship. While these results are based on cross-sectional associations and should therefore be interpreted as indicative rather than causal, they nonetheless underscore the pivotal role of digital inclusion in shaping entrepreneurial pathways in Mexico; addressing the heterogeneity of entrepreneurial profiles—by necessity versus opportunity, and across gender and age groups—should remain a central principle of Mexico’s digital inclusion and entrepreneurship policy framework.