Digital development and employment gender gaps during the COVID-19 pandemic: evidence from Latin America and the Caribbean

Abstract

Purpose

Despite its negative effects, the COVID-19 pandemic has accelerated digital development in Latin America and the Caribbean. The rapid increase in connectivity and digital services helped mitigate the pandemic's negative impact on the labor markets, especially for those with enough flexibility to continue working from home. The shock affected women due to their household responsibilities and labor market characteristics.

Design/methodology/approach

This paper examines how digital development may have affected gender gaps in employment and job loss in the context of the COVID-19 crisis. Using a sample of countries from Latin America and the Caribbean and various econometric techniques, we explore the digitalization gender gaps and job market outcomes during the pandemic.

Findings

Our findings suggest that the expansion on digital technologies are associated with increased female employment and reduced job losses for both men and women. These findings hold even after controlling for child care, household chores and the COVID-19 shock. Our results are also robust to various econometric techniques.

Originality/value

The paper leverages on unique dataset that was collected during the pandemic and the results are contrasted with existing macro data with robust results.

Citation

Yang, Y., Granados Ibarra, S., Ghazanchyan, M. and Canavire-Bacarreza, G. (2024), "Digital development and employment gender gaps during the COVID-19 pandemic: evidence from Latin America and the Caribbean", Journal of Internet and Digital Economics, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JIDE-02-2024-0005

1. Introduction

Digital development is critically important as it significantly impacts the economy and society. The expansion of digital technologies brings direct benefits to multiple areas. It stimulates economic growth through job creation and increased productivity across sectors. Moreover, it improves the quality of life by increasing access to essential services such as healthcare, education, and banking, particularly in remote and underserved areas. It also simplifies daily tasks through online shopping, digital payments, and remote work options. Educationally, digital platforms provide extensive resources and courses, making learning more accessible and flexible and helping individuals develop new skills essential for today's job market. Socially, digital development reduces inequalities by bridging the digital divide and empowering marginalized communities to participate more fully in economic and social activities.

The COVID-19 has resulted in the disruption of global supply chains. Associated delays, shortages, and increased costs have had a profound impact on both supply and demand, intensifying the economic effects of the pandemic (e.g. Zhang et al., 2022; Abidi et al., 2023). On the demand side, the COVID-19 crisis has triggered a structural shift toward a digitalized economy, a trend likely to persist in the coming years. This shift has significantly impacted global supply chain procurement, forcing firms to adapt swiftly. The situation underscores the need for firms to address the structure and transparency of their supply chains to mitigate risks. While digitalization offers potential improvements, many firms currently lack the necessary knowledge, data, and infrastructure to achieve these changes (Kersten et al., 2023). Increased demand for digital infrastructure, such as high-speed mobile internet and fixed broadband, has largely been met in Europe and Central Asia (IFC, 2021). However, without greater investment in network capacity, there is a risk of a widening gap in quality connectivity, (Lutz et al., 2024) particularly in emerging markets like Latin America and the Caribbean, where little research has been conducted.

The COVID-19 pandemic has further accelerated the adoption of digital technologies across different sectors, such as education, healthcare, and labor markets, marking a rapid and unexpected shift. Education underwent significant changes, relying on digital technologies to sustain classes despite the quality challenges associated with these strategies. Telemedicine experienced a surge, improving healthcare access for previously underserved patients amid overwhelmed health systems.

In the labor market, digital development created new job opportunities in fields like e-commerce, digital marketing, and software development, often allowing remote work possibilities. While these sectors have predominantly employed men, efforts to address gender biases and expand educational access have gradually increased female representation (OECD, 2021). This shift has the potential to narrow gender employment gaps, as women previously excluded from traditional job markets due to caregiving responsibilities can now access these emerging opportunities. However, evidence (OECD, 2021) indicates that the pandemic disproportionately affected women's employment. Women were more likely to work in sectors severely impacted by the pandemic, such as hospitality, retail, and healthcare. Additionally, women are overrepresented in informal job sectors offering fewer benefits and flexibility to balance work and caregiving roles. Unequal access to digital technologies further exacerbates these gender disparities, limiting women's ability to leverage digital tools for work fully.

Therefore, understanding the relationship between digital development and gender gaps is crucial to ensure equitable recovery and sustainable digital development that benefits all segments of society. Moreover, the COVID-19 pandemic expansion provides an adequate setting to examine this relationship.

Latin America and the Caribbean (LAC) is a good setting to examine the effects of digital development since the region has historically faced challenges in digital development, including significant gaps in connectivity infrastructure, high levels of informality in the economy, and a largely unbanked population. Yet, the region has made important improvements in terms of digital development with the COVID-19 pandemic, accelerating a shift towards digital technologies in the region in various areas. For instance, within less than two years, e-commerce in the region increased by 36.7% relative to pre-pandemic levels, with over 13 million people across LAC completing online transactions for the first time in 2020 (Statista, 2023). While the pandemic significantly accelerated digital expansion in the region, the expansion of the internet predates the pandemic. Moreover, the expansion of the Internet in Latin America has been remarkable, in 2018, 68% of the population used the Internet regularly – almost twice the share in 2010, although lagging behind the OECD average of 84% (Figure 1).

Gender gaps in labor market participation have reduced over the past twenty years in Latin America and the Caribbean (LAC). Factors driving these improvements include improved women's education access, declining fertility rates, and later marriage (Chioda and Verdú, 2016). Despite these positive trends, women's participation rates in the labor force in LAC remain notably lower than those of men and women holding lower-quality jobs. Men are more likely than women to be employed in digitally intensive sectors such as high-tech, construction, utilities, and transportation, which generally offer better pay and benefits (World Bank, 2018). Conversely, women are frequently concentrated in sectors like education and healthcare or employed as domestic workers. This occupational segregation is particularly pronounced among indigenous women, who are disproportionately represented in domestic work compared to non-indigenous women, as observed in Colombia, Costa Rica, Panama, and Mexico (ECLAC, 2014). Therefore, the region presents an apt setting to examine the effects of digital development on labor market gender gaps.

Digital development is a complex and ever-evolving concept where the expansion of the internet plays a crucial role. Reis et al. (2020), who comprehensively review digital development's evolution, highlight that this concept evolves alongside business processes. Our empirical approach, aligned with existing literature, focuses on using the number of internet users as a key metric for measuring digital development. Increased internet access directly enhances digital development by facilitating rapid information exchange, connecting individuals and businesses, and fostering economic exchanges. These factors contribute to improved efficiency, reduced operational costs, and overall enhancement of digital development. Thus, it is evident that expanding internet access, particularly in developing countries, is essential for enabling digital development.

Analyzing high-frequency surveys conducted by the World Bank across 22 countries in Latin America and the Caribbean to examine microeconomic impacts and using a cross-country panel to explore macroeconomic effects reveals a significant correlation: higher levels of digital development are linked to increased female employment and decreased job losses for both men and women.

The remainder of this paper proceeds as follows: Section 2 presents the context of digital development in Latin America and the Caribbean and a literature review. Section 3 presents the data and methodology for the microeconomic and macroeconomic approaches. Section 4 reports the key findings. Section 5 presents robustness tests, and Section 6 concludes.

2. Where do we stand? A brief review of the literature

2.1 A brief context of digital development in Latin America and the Caribbean

Digital infrastructure is a key area of development in the LAC region. While the region has made progress in this area, it still lags behind other regions, such as Western Europe and Asia. The LAC region has 46% fixed broadband access, compared to 57% in Eastern Europe, 87% in Western Europe, and 59% in Asia Pacific. This suggests that there is still a significant gap in digital infrastructure between the LAC region and other regions worldwide. Digital public platforms are another important area of development in the LAC region. These platforms can help improve government services and increase transparency and accountability. Argentina, Brazil, Chile, and Uruguay are among the top 50 performers in 2018, performing slightly below the OECD average. However, Belize, Cuba, Haiti, and Nicaragua were among the worst LAC performers. This highlights the need for continued investment in digital public platforms to ensure that all citizens benefit from these services.

In terms of digital financial transactions, the LAC region is still facing challenges. While mobile money accounts are becoming more popular, the growth rate in the LAC region is lower than in other regions, such as Western Africa. According to the data provided, in 2019 the LAC region experienced the lowest growth rate in the number of registered mobile money accounts (+2.5%, Western Africa 14.5%) and the lowest growth rate in transaction value (+1.4, Western Africa +34.9%). This suggests that there may be barriers to adopting digital financial services in the LAC region that must be addressed. On the other hand, venture capital investments in digital business are showing positive trends in the LAC region. For example, in 2019, Brazil and Mexico led the region in terms of the number of deals and volume transacted. However, the region still lags behind the Asia Pacific in attracting investment. According to KPMG's (2019) report, the LAC region attracts less investment than the Asia Pacific region. This highlights the need for policies and initiatives that can help attract more investment to the region.

The LAC region faces a significant challenge in education and training in digital skills. As mentioned earlier, the levels of education and training in digital skills in the LAC region are low compared to advanced countries. IDC's (2017) report estimates that there are more than 450 thousand unfilled jobs in the technology area due to the lack of trained professionals. This skills gap can make it difficult for the region to fully realize the potential of digital technologies, which can significantly impact economic growth. However, the potential benefits of digital transformation are significant. According to GSMA, a 10% increase in mobile internet penetration can increase GDP by 1.2%, while a 10% increase in a country's digitalization can generate 1.9% in GDP growth. This highlights the importance of investing in digital infrastructure and skills to fully realize the potential of digital technologies.

From another angle, most Latin American countries' competitiveness is largely based on abundant natural resources or low-skilled labor. This has resulted in a poorly diversified production structure, entailing low value-added and an export specialization concentrated in goods with low technological content. While this structure can provide periods of rapid growth, sustained productivity growth requires incorporating technology and production diversification towards dynamic sectors, both in technology and in terms of international demand (Bárcena Ibarra, 2022; Gottschalk and Weise, 2023).

To escape the productivity trap, the LAC region must take advantage of the digital transformation and promote production transformation. Some countries in the region are already incorporating policies to boost the development of emerging technologies, such as advanced robotics and artificial intelligence (AI), to improve productivity. Examples of such efforts include Brazil's National Internet of Things Plan, Colombia's Fourth Industrial Revolution Center operated by the Ruta N Corporation in Medellin, and Uruguay's digital manufacturing laboratory. Challenges remain, especially in the productive application of digital technologies, the development of digital entrepreneurship, and business heterogeneity. A large share of smaller businesses has difficulties adopting new technologies. Furthermore, despite the rapid pace of technological change and its potential to improve efficiency, aggregate productivity growth, including in LAC, has slowed over the past decade, giving rise to a productivity paradox (De Backer and Flaig, 2017).

2.2 Digital development and gender gaps

Technological progress related to digital development has the potential to impact labor market outcomes through various channels, as highlighted by recent studies (Loko and Yang, 2022; Bakker et al., 2023). On the one hand, digital development can enhance labor market efficiency by improving access to information about job openings, reducing recruiting expenses for employers, and improving job matching quality (Autor, 2001). On the other hand, it can lead to increasing flexibility and deregulation, characterized by a reorganization of work tasks and a liberalization of employment forms (Eichhorst and Tobsch, 2015).

Moreover, an increasing number of empirical studies suggest that internet access has a more significant and positive impact on women than men. For instance, Klonner and Nolen (2010) reveal significant impacts of network expansion on labor market outcomes, highlighting notable gender-specific differences. When a community gets network coverage, employment grows by 15% points. However, a gender-disaggregated analysis indicates that most of this impact is driven by increased female employment. This suggests that digital development can promote gender equality in the labor market by providing women better access to job opportunities and reducing gender-based employment gaps.

Similarly, recent studies have shed light on the potential impact of digital development on labor force participation and household consumption. For instance, Dettling (2016) uses instrumental variables approach that leverages state-level variations in the availability of residential broadband internet access to demonstrate that high-speed internet leads to a substantial 4.1% point increase in labor force participation for married women. However, no such effect is observed for single women or men. Further analysis indicates that using the Internet for telework and saving time in household activities can account for increased participation. These findings suggest that access to home internet can improve work-family balance and promote gender equality in the labor market.

Likewise, Bahia et al. (2020) exploit a unique dataset that merges information from a national longitudinal household survey on living standards with data from Nigerian mobile operators concerning the rollout of mobile broadband coverage from 2010 to 2016. They show that mobile broadband coverage had significant and positive impacts on household consumption levels, which grew over time, although at a declining rate. Furthermore, mobile broadband coverage helped to decrease the percentage of households living below the poverty line, mostly by boosting food and non-food consumption in rural households. These outcomes were primarily attributed to increased labor force participation and employment, especially among women. These findings suggest that digital development can promote economic development and reduce poverty, particularly in rural areas, by providing better access to job opportunities and improving household consumption levels.

Kumar et al. (2023) identify key drivers of digital adoption, estimate fiscal costs to provide internet subsidies to households, and calculate social dividends from digital adoption. Using cross-country panel regressions and machine learning, the authors find that digital infrastructure coverage, internet price, and usability are the most statistically robust predictors of internet use in the short run. Based on estimates from a model of demand for the internet, the authors find that demand is most price-responsive in low-income developing countries and almost unresponsive in advanced economies. The authors also claim substantial aggregate and distributional gains from digital adoption for education quality, time spent doing unpaid work, and labor force participation by gender.

2.3 COVID-19 and gender gaps

Evidence in a recent study by Abidi et al. (2022) suggests that the COVID-19 pandemic resulted in an unprecedented shock to firms, with adverse consequences for existing productive capacities. On the other hand, digital development played a vital role in mitigating economic losses from the pandemic. The authors claim that firms facing digital constraints are less resilient to supply shocks. The paper uses firm-level data to investigate whether digitally enabled firms were able to mitigate economic losses arising from the pandemic better than digitally constrained firms in the Middle East and Central Asia region using a difference-in-differences approach. Controlling for demand conditions, the authors find that digitally enabled firms faced a lower decline in sales by about 4% points during the pandemic compared to digitally constrained firms, suggesting that digital development acted as a hedge during the pandemic.

Previous economic recessions have shown that male employment is typically more affected than that of women. This is largely because male-dominated sectors such as construction and manufacturing are typically hit harder during recessions than female-dominated industries such as education and health care, as demonstrated by Coskun and Dalgic (2020) and Smith and Villa (2013).

However, the COVID-19 pandemic impacted women's employment more than men. The literature has identified two primary factors contributing to this gender disparity. Firstly, lockdown measures and concerns about contagion disproportionately affected industries and occupations with a higher proportion of female employees. For instance, the pandemic severely impacted the hospitality and retail sectors, which employ many women [1]. Secondly, the closure of schools and daycare facilities and the transition to remote learning increased the need for childcare, resulting in many parents, particularly mothers, choosing between their jobs and being responsible for caring for their household and all that comes with it. This significantly reduced women's labor force participation and employment, as highlighted by recent studies (e.g. Alon et al., 2020; Adams-Prassl et al., 2020).

Women, young workers, and those with low levels of education and limited internet connectivity were disproportionately affected by the economic downturn caused by the pandemic-related lockdowns in Latin America and the Caribbean. The Covid-19 crisis also brought historically high levels of absences from work, particularly for women and independent workers. Absences preceded job losses, with almost one in five absent workers losing their employment after two months (World Bank, 2021).

Extensive literature has demonstrated that women's employment in certain occupations and industries has led to higher unemployment rates than men. Recent studies, such as Adams-Prassl et al. (2020), using real-time survey data from the United Kingdom, the United States, and Germany during the pandemic, have found that workers who are engaged in alternative work arrangements or have jobs that do not allow remote work are more prone to experience a reduction in working hours, job loss, and decreased earnings. These findings suggest that the pandemic disproportionately affected women's employment, particularly those in non-standard work arrangements.

Similarly, Bluedorn et al. (2021) shows significant diversity across countries, with more than half to two-thirds experiencing greater decreases in women's than men's employment rates. These gender disparities caused by COVID-19's effects are usually temporary, lasting only for an average of one or two quarters. Furthermore, the study demonstrates that the gender-based recession is closely connected to COVID-19's influence on the distribution of employment shares between genders within sectors.

Women often assume the role of primary caregiver, which has implications for their employment during the COVID-19 crisis. According to survey data collected by the American Time Use Survey (ATUS, 2021), married women tend to provide more childcare than married men. The data reveals that, among all married couples with children, husbands provide an average of 7.4 h of childcare per week, whereas wives provide an average of 13.3 h.

Research by Alon et al. (2020) found that when childcare needs increase, as during the pandemic, women are more likely to leave their jobs to care for children. The closure of schools and daycare centers also resulted in a substantial increase in childcare needs, disproportionately affecting working mothers. The negative impact on working mothers is also likely to be long-lasting due to high returns to experience in the labor market.

These findings suggest that the COVID-19 pandemic had a unique and disproportionate impact on women's employment, highlighting the need for policies that address the specific challenges women face in the labor market. Such policies could include measures to support the childcare needs of working parents and targeted support for industries and occupations with a higher proportion of female employees.

3. Empirical strategy and data

We use two approaches to examine the relationship between digital development and employment gaps. On the one hand, we exploit data from household surveys developed by the World Bank and UNDP to capture variation across households and countries. This allows us to identify the effects of digital availability on employment and job losses. On the other hand, we employ cross-country panel data to understand the effects of expanding internet availability on employment rates by gender in LAC.

3.1 Household-level analysis

3.1.1 Data

The data used in this section comes from the second phase of the High-Frequency Phone Survey (HFPS), wave 2 in LAC. The data was collected between October and December of 2021 as part of a global-scale effort to understand better the pandemic's effects and their mechanisms, led by the World Bank Group and the UNDP. The survey covered 22 countries in Latin America and the Caribbean and included questions regarding individual and household demographics, income, employment, health, and education, among other topics. It is important to note that this survey is not a panel, and the latest information available is from wave 2, which we are using for this paper. As explained in the paper by Ambel et al. (2021), phone surveys are usually subject to coverage and non-response bias, and these biases can be alleviated by sample reweighting; we apply the suggested weighting so that the estimates are as unbiased as possible.

The analysis focuses on two main outcome variables: 1. “job loss,” which is a binary variable that takes a value of one for individuals who reported not currently working and having lost their pre-pandemic job; and 2. “employed,” which is also a binary variable that takes a value of one if the individual reported having worked in the last week.

The main independent variable of the analysis is called “Women with internet,” which is an interaction variable of the dummies for being a woman (1 if woman and 0 otherwise) and having an internet connection in the household (1 if yes and 0 otherwise). Hence, this interaction variable only takes a value of one for women with an internet connection in their household. The analysis also includes control variables such as country, state, education level, and marital status, which are included as fixed effects.

The last group of variables are complementary variables that allow us to further explore the main hypothesis by splitting the sample to check how the results vary between groups with certain characteristics. We use four binary variables for this purpose: 1. “income reduced,” assigned a value of one for individuals who reported a decrease in household income compared to the previous wave (around June 2021), and zero otherwise; 2. “High COVID-19 level,” a dummy variable that takes a value of one if the country had a rate of COVID-19 cases higher than the world median; 3. “Increased household chores,” which is a dummy variable that takes a value of one if the individual reported an increase in household chores since the beginning of the pandemic; and 4. “Household with children,” which is a binary variable that takes a value of one if there is at least one child in the household. Similarly, for each regression, we control for important covariates that might be influencing the behavior of the labor market, such as if the woman has a smartphone or if there are children in the household.

Table 1 presents descriptive statistics of key variables. Although this survey does not include income level information, we use the question, “Has your household income stayed the same, decreased or increased compared to the previous wave (around June 2021)?” to assign a value of one to individuals who reported a decrease in household income.

Table 1 presents the mean values of all variables included in the analysis for the entire sample (row “total”), as well as disaggregated by country. Regarding the variables of interest, the percentage of individuals who lost their pre-pandemic jobs and are not currently working is around 17%, with some variation across countries such as Haiti, Panama, and Colombia. The employment variable has an average value of 64% and shows less variation among countries than the job loss variable.

The percentage of women and households with children has a small variation, with an average of 52 and 60%, respectively. In contrast, internet access shows significant variation among countries, with an average access rate of 61%. However, this mean is driven by countries with high rates, such as Dominica (83%), St. Lucia (81%), and Jamaica (71%). In contrast, some countries have extremely low rates, such as Haiti (9%).

In addition, the HFPS for 2021 clearly shows the gender gap in employment outcomes and how it is related to having access to the Internet and other household variables. Figure 2 displays the inequality in the labor market between women and men. In Figure 1, we observe that the employment level for women is 22 percentage points smaller than for men, and the proportion of job loss during the pandemic is 14 points larger for women than for men. On the other hand, when we look at the gender gap and its relationship with internet access, we see that it is alleviated for men and women with internet access. Figure 3 presents two important observed patterns: first, digital development measured as internet access, improves labor outcomes; second, digital development makes a larger difference in employment (7 vs 3 points) and job loss (11 vs 6 points), for women than for men.

3.1.2 Empirical strategy

To examine the effects of internet connectivity on female employment, we employ a fixed effects model utilizing female and internet access dummies and their interaction. This constitutes the primary specification of our analysis and can be expressed through the following equation:

(1)Yi=α+β1Femalei*Interneti+β2Femalei+β3Interneti+δc+γ1Educ+γ2Marital_status+εi

Where Yi is either the outcome variables “employment” or “job loss”, and δi represents the fixed effects for each individual's country-state, marital status, and education level. In Equation (2), the coefficient of interest is β1, as it represents the effect of being female and having internet access on the labor market outcomes we are analyzing, compared to women without internet access. Hence, this coefficient can give us an idea of the effect of internet on female labor market outcomes, including employment. It is important to note that this specification is prone to suffer from bias as it is likely that some unobserved variables are related both to the outcome and to having internet access, so the coefficient cannot be understood as a causal effect but more as correlations between the variables we are analyzing. For example, we lack information regarding the individuals’ inner abilities, experience and family and friends network, that improve their chances of having a job, as well as the probability of having internet at home, hence, leading to an upper bias in our estimations. This positive bias is reinforced for the reverse causality problem that also can be observed in this setting.

In addition, to study how these correlations change for different groups, we apply Equation (2), to a sample split by the group variable. These analyses can be represented as:

(2)Yi=α+β1Femalei*Interneti+β2Femalei+β3Interneti+δc+γ1Educ+γ2Marital_status+εi if groupi=1

Yi=α+β1Femalei*Interneti+β2Femalei+β3Interneti+δc+γ1Educ+γ2Marital_status+εi if groupi=0

In the previous equation groupi represents the variables that allow identifying different groups “High Covid-19 level”, “Decreased income”, “Increased household chores”, and “Household with children”. In addition, for all the estimations we use clustered standard errors at the individual level and survey weights for country population.

Equations (1) through 3 are estimated using a OLS regression with the respective, country-region fixed effects, we preferred this estimation method rather than a non-linear probability model, as Angrist and Pischke (2009), explain that OLS is acceptable even when dealing with binary dependent variables, as it provides more understandable effects and we are dealing with a causality/correlational problem and not with a prediction problem.

3.2 Country-level analysis

3.2.1 Data

To measure a country’s level of digital development, we construct two main indicators. Following the existing literature (Habibi and Zabardast, 2020; Myovella et al., 2020; Mgadmi et al., 2021; Lechman and Anacka, 2022), our first indicator Digitalization is the percentage of internet users in total population obtained from the World Bank Development Indicators [2]. We also collect data on the natural logarithm of the number of individuals using the internet in a given country as an alternative measure. We define women’s labor force participation rate, Female LFPR, as the ratio of female population aged 15 and older that is economically active and supplies labor for the production of goods and services. We also use the female employment to population ratio as an alternative measure. Male LFPR and employment ratio are also obtained for comparison.

To establish the relationship between digital development and female employment, we control for an array of variables that reflect a country’s economic, social, and institutional status. GDP Per Capita is the natural logarithm of a country’s per capita GDP, indicating the level of economic development. Account is proportion of women owning an account at a financial institution. We also included Schooling variable, defined as average year of total school for population above 15, to account for a country’s human capital. Women law, a score ranging from 0 to 100, symbolizes the level of female protection in a country. The higher the score, the more effective the legal system is at safeguarding women's rights. All data were obtained from the World Bank.

To incorporate the economic mechanisms underlying the relationship between digital development and female employment, we explore how digital development exerts a differential effect in countries with varying degrees of household burden, COVID conditions, income levels and geographic locations. More specifically, we focus on the following four variables: Trained teachers measures the number of trained teachers over the total number of teachers in a country. The higher the ratio, the better the teaching quality. COVID−19 is a dummy variable that takes the value of one if the number of COVID-19 cases in a country is above the median, based on the Johns Hopkins COVID information updates. Income Level is the classification of countries based on gross national income per capita data in U.S. dollars, converted from local currency using the World Bank Atlas method to smooth exchange rate fluctuations. LAC is a dummy that equals one if the country is located in the Latin American and Caribbean region and zero otherwise [3]. Descriptive statistics of key variables are presented in Table 2.

3.2.2 Empirical strategy

To estimate the relationship between digital development and female employment at the country level, we construct the following baseline model:

(3)Female LFPRi,t=β0+β1Digitalizationi,t−1+β2Xi,t−1+θi+μt+εi,t

where Female LFPRi,t refers to women’s participation in the labor force of country i in year t, measured by the ratio of female workers in the labor force. Digitalizationi,t−1 captures the level of digital development of country i in year t−1, measured by the ratio of internet users in the population. Xi,t−1 is a vector of country-level controls, including the natural logarithm of per capita GDP, the proportion of female account ownership at a financial institution, the average year of total schooling, and the level of women, business, and law protection. All explanatory variables are lagged by one year to mitigate endogeneity concerns.

By including θi, we account for country fixed effect that absorbs systematic differences across countries regarding economic development, government policies, ICT-specific reforms, etc. μt denotes year fixed effect that picks up any variation in the outcome over time and is not attributable to other explanatory variables.

The coefficient of interest is β1, which is associated with digital development variables. If it is positive and significant, we can conclude that a higher level of digital development is associated with higher female representation in the workforce, hence a lower degree of gender inequality. We can infer a negative correlation between digital development and female employment if it is negative and significant.

3.2.3 Examining mechanisms

We proceed with fixed-effect regressions with interaction terms, hoping to shed light on the mechanisms through which digital development exerts an impact. By estimating various forms of the model below, we examine the differential effects of digital development on female employment.

(4)Female LFPRi,t=α+β(Digitalizationi,t−1×Countryi)+γDigitalizationi,t−1+δIi,t−1+θi+μt+εi,t

where Countryi refers to country-level characteristics, including overall household burden, income level, and geographical location, that capture economic mechanisms and help identification. All explanatory variables are lagged by one year to mitigate simultaneity concerns. The sign and magnitude of the coefficient are of interest to our study.

The COVID-19 pandemic swept across the world. As social distancing became compulsory and telework became the new trend; access to digital resources should give women an advantage in participating in the labor market. We created a dummy variable to test whether digital development has helped women get through the pandemic. COVID−19, which takes the value of one if a country has above-median COVID-19 confirmed cases and zero otherwise and reruns model (4).

4. Results

4.1 Individual-level results

Table 3 presents the results for both outcome variables: job loss in panel A and employment in panel B, as well as a secondary outcome in panel C for hours worked per week. In the first two columns, we run a simple regression to check the correlation between labor market outcomes and having an internet connection, with and without fixed effects. These results align with the expected direction in theory (for example, Loko and Yang, 2022; Bakker et al., 2023): having access to the internet increases the probability of being employed and decreases the probability of losing one's job and being unemployed. In both cases, the coefficients have the expected signs and are statistically significant.

In column (3), we present our main results from Equation (1). Here, we find that the coefficient of interest displays a positive correlation between being a woman with internet access and being employed and a negative correlation with job loss. Additionally, this relationship holds when we control for living in a household with children, having increased household chores, or having a smartphone; the first two variables can be proxies for other common reasons why a woman may decrease her labor supply, and the presence of a smartphone allows us to distinguish the effect of having a mobile device. To further explore whether women with different contexts, such as having children in the household, present a different relationship between internet and labor outcomes, we proceed with Equation (4) and split the sample into different groups. Tables 4 and 5 provide the associated results.

Considering that there is a reverse causality between internet access and employment, these results should be interpreted as correlational. However, we also check that when we estimate the effect of having internet access on the probability of working remotely (see Table A5 of the Appendix), there is not a significant relationship between them nor with the interaction with being a woman.

In Table 4, we find that the interaction coefficients are larger for individuals who experienced increased household chores when analyzing the variable “Job loss.” However, for the variable “Employed,” we do not find a significant effect with the split sample, although the coefficients remain positive. This suggests that digital development may be more effective in mitigating the negative impact of household chores on job loss than in promoting employment.

Similarly, when we split the sample between individuals living in households with and without children, we find that for the variable “Job loss,” the interaction coefficients are negative and larger for those living in households with children, as shown in columns (3) and (4) of Table 4. This could be due to the possibility that the internet plays a more important role for women with more responsibilities at home, as it allows them to work from home when possible and provides them with more flexibility. These findings highlight the potential of digital development to help alleviate the burden of childcare responsibilities for women and promote their participation in the labor market.

In Table 5, we observe that the differential effects on income are reduced, and we find that the coefficients of the interaction between females and the Internet are larger for individuals whose household income was reduced. This suggests that digital development can be particularly beneficial for women in households experiencing income shocks, as it provides opportunities for remote work and reduces the need for physical presence in the workplace. A similar pattern can be observed in columns (3) and (4) of Table 5 for the sample of individuals living in countries with higher COVID-19 cases, highlighting the potential of digital development to mitigate the negative relationship of the pandemic on female employment.

Our findings suggest that digital development can play an important role in promoting gender equality in the labor market, particularly in contexts where women face significant household and childcare responsibilities [4]. However, additional policies and investments may be necessary to realize the potential of digital development in these contexts fully.

4.2 Country-level results

Our results on country-level analysis are provided in Table 6. In Column (1), we estimate Equation (3) using the ratio of female workers as the dependent variable and the level of digital development as the independent variable. After dropping the missing values, we have a sample of 169 countries. The coefficient on digital development is positive and significant at the 1% level. The result is consistent with our hypothesis that digital development is associated with a significant increase in female labor force participation rate. More specifically, a 1% increase in digital development is associated with a 0.07% point increase in the ratio of female workers in our sampled countries. It is worth noting that the coefficient on digital development is also significant when the dependent variable is replaced by the male labor force participation rate, though with a smaller magnitude than that for females.

In order to analyze the varying impacts of digital development on female and male labor force participation, we introduce a variable known as the female-to-male labor force participation ratio. Consistent with Valberg (2019) [5], the coefficient derived from our analysis reveals a positive and statistically significant value (Column 6), indicating that women are more profoundly affected by digital development than men.

4.2.1 Impact of digital development across countries

In this section, we interact the digital development variable with various country characteristics, hoping to shed light on the economic mechanisms underlying the positive relation between digital development and female employment. The variable Trained Teachers reflects the availability of teachers’ services to their students and teachers’ workload (Snyder et al., 2016). The higher the trained-teacher ratio in primary school, the better the availability of teacher services to students, and the lower the childcare burden relegated to households and, in most cases, women. In Column (1) in Table 7, the interaction term of internet and trained teacher ratio—our chosen proxy for the availability of childcare—is negative and statistically significant, indicating that the effect of digital development is less important in countries with more childcare resources. In other words, digital development synergizes with the availability of childcare resources to alleviate the household responsibilities faced by women, thus facilitating their reintegration into the workforce.

Next, we use high-income countries as the control group and put both low and middle income into the same regression; the two interaction terms are both negative and significant, indicating that compared with high-income countries, lower-income countries lack the institutional capacity to fully tap the potential of digital development. However, the negative effect is insignificant regarding male labor force participation rate. (Table 8)

We then zoom in on the LAC region, where gender inequality in labor force participation is a major concern due to a disproportionately large informal sector. The sample has a rather extensive coverage of countries in the region, covering 33 countries. The complete list of Latin American and Caribbean countries is in the Appendix. Consistent with Asongu and Odhiambo (2023) which indicate a positive impact of digital development on female economic inclusion in Sub-Saharan Africa [6], our results, which are summarized in Table 9, suggest that digital development exerts a positive effect on countries in Latin America and the Caribbean, with a significantly larger magnitude compared to other regions when all control variables are included. Again, the additional positive effect is muted for males.

4.2.2 Sectorial impact of digital development

Could sectoral differences drive the positive effect of digital development on female employment? Digital development often enables remote work through telecommuting and flexible work arrangements in the service sector, particularly knowledge-intensive services. This can be especially beneficial for women with caregiving responsibilities at home. They can work from home or choose flexible hours, making balancing work and family life easier.

To test this hypothesis, we collect data on the size of the service sector in each country and interact it with our digital development indicator. The service sector size is a dummy variable that takes the value of one if the service sector share of employment is above the average share of all countries in a given year and zero otherwise. The results in Table 10 indicate that countries with a larger service sector see a notable rise in female labor force participation, as suggested by the significant and positive sign on the interaction term. The ratio of female to male labor force participation is also raised. However, the impact on men’s labor force participation ratio is not statistically significant.

4.2.3 Did digital development help women get through COVID-19?

When distinguishing between countries with better/worse COVID-19 impact, we find that the benefits of digital development are larger in the latter, as indicated by the positive and significant coefficient on the interaction term of digital development and COVID-19 dummy. The results in Table 11 are consistent with our hypothesis that digital development, to some extent, helps women participate in the labor force during the pandemic. We do not observe substantial digitalization benefits for male workers during COVID-19.

5. Robustness checks

In the section on household-level results, although our main specification involves an OLS estimation with fixed effects, we also conduct estimations using logistic regression with country-state and the same fixed effects variables used in the main results. We run the logistic regression as a robustness check to account for the binary nature of our dependent variables, “Lost job” and “Employment”, although Angrist and Pishke (2009) argue that OLS is still preferred in this case. As shown in Table 12, the results from the logistic regressions confirm our main findings: there is a positive correlation between internet access and its interaction with being a woman for employment, as indicated by positive coefficients and greater-than-one odds ratios. Additionally, there is a negative relationship between internet access and job loss. These results provide further support for the positive effect of digital development on female employment and the potential of digital development to mitigate the negative impact of a pandemic on female employment.

The results remain robust for country-level results, using alternative digital development measures and female employment (see Table 13).

To further alleviate endogeneity concerns, we follow Arellano and Bond (1991) to adopt GMM models (see Table 14), which control for endogeneity by internally transforming the data and including the dependent variable's lagged values. The results remain statistically significant and consistent with our hypothesis, suggesting a robust positive relation between digital development and female labor force participation.

6. Conclusions

This paper provides a comprehensive analysis of the effects of digital development on female employment, using both micro and macro-level data and accounting for shocks such as the COVID-19 pandemic and household chores and childcare responsibilities. Our findings suggest that higher levels of digital development are associated with higher employment and lower job loss in the LAC region, and this effect is accentuated for women.

Furthermore, our results indicate that digital development has the potential to significantly alleviate the impact of shocks on employment and job loss, especially for women, as evidenced by the COVID-19 crisis. We also find that the effect of digital development is stronger in countries with fewer childcare resources, highlighting the potential of digital development to help alleviate the burden of childcare responsibilities for women and promote their participation in the labor market.

However, our analysis also reveals that the positive effects of digital development can be limited by institutional weaknesses in countries with already aggravated structural vulnerabilities. Therefore, policies aimed at improving institutional capacity, such as investments in education and infrastructure, may be necessary to realize the benefits of digital development in these contexts fully.

Overall, our findings contribute to the growing literature on the relationship between digital development and female employment and provide important insights for policymakers seeking to promote gender equality and economic development. Although the results presented are merely correlational, there is a clear pattern of the importance of digital development on women’s labor outcomes.

Finally, while the COVID-19 pandemic provided a useful example of an economic shock and its differential gender impact, it is premature to generalize these findings. With the causal effects of digitalization on employment in general being highly uncertain, future research should try to disentangle the role of digitalization and currently AI from other factors affecting employment and female employment by applying perhaps a modeling approach or a scenario analysis. The other related limitation of our study is that its micro part is based on short surveys that may hinder the long-run effects of both digitalization and other factors on employment and job loss. It is undoubtedly true that digitalization will continue play to be one of the most important aspects of employment studies, however and due to its evolving nature, the precise quantitative and causal effect on employment may remain an open question for a long time. Future research should also include estimating all-encompassing expected vulnerabilities of women to digital development, which could be useful for social and economic policy discussions. As the gender gap in job participation in LAC continues to be highly elevated and data shows that women in LAC had been 44% more likely to lose their jobs than men, future work should also study the impact of broader aspects of digital transformations such as automation and AI (Artificial Intelligence) in the hope to boost female employment with new technologies (World Bank, 2022).

Figures

Figure 1

Percentage of internet users in selected Latin American and Caribbean countries 2010 and 2018 (or the latest available year)

Figure 2

Gender gap on employment outcomes

Figure 3

Gender gap on employment outcomes by internet access

Table 1

Summary statistics for household-level analysis

	No. Obs	Job loss (%)	Employed (%)	Women (%)	Internet access (%)	Women with internet access (%)	Increased household chores (%)	Household with children (%)
Belize	898	22.9	57.7	50.3	69.3	34.3	31.3	67.1
Guatemala	1.521	15.9	66.9	51.9	33.3	17.1	20.6	72.7
El Salvador	812	13.7	66.1	55.3	47.3	24.8	20.3	63.5
Honduras	1.004	21.2	57.8	52.5	42.1	20.1	20.3	76.2
Nicaragua	865	13.2	67.5	51.8	32.6	15.9	23.4	72.5
Costa Rica	905	18.7	60.8	49.9	62.2	31.1	28.2	50.2
Panama	986	25.6	53.1	50.3	58.3	29.8	26.4	60.3
Haiti	2.361	33.3	46.2	51.6	9.0	3.9	35.4	74.8
Peru	1.302	21.9	67.7	50.7	48.9	24.0	33.2	69.0
Mexico	2.511	14.7	66.0	52.0	68.6	34.0	30.4	58.1
Argentina	1.321	13.1	65.6	51.8	76.9	39.3	25.8	49.4
Chile	1.329	13.1	62.5	51.1	73.6	32.9	37.9	45.2
Colombia	1.376	26.5	58.9	52.1	58.5	30.1	27.9	64.5
Bolivia	1.183	13.7	74.0	50.3	59.9	30.6	22.9	67.8
Guyana	875	16.3	63.1	49.8	70.3	34.2	37.7	61.6
Ecuador	1.615	17.3	64.7	50.6	74.2	36.6	25.8	71.8
Paraguay	1.061	10.9	77.4	50.0	51.9	22.2	20.7	61.8
Uruguay	930	16.8	61.0	52.3	72.4	35.9	22.6	42.8
St Lucia	860	13.2	69.1	50.7	81.5	42.4	31.0	47.2
Dominica	879	15.1	66.0	49.6	83.2	43.8	37.7	51.6
Dominican Republic	1.197	20.8	61.4	50.5	57.0	29.1	28.8	63.5
Jamaica	871	14.4	65.7	51.0	71.0	36.2	35.4	57.2
Total	26.662	17.4	64.4	51.7	62.4	31.0	28.7	60.5

Source(s): Own estimates using HFPS data for LAC

Table 2

Summary statistics for country-level analysis

Variable	No. of Obs	Mean	St. dev	Min	Max
Female LFPR	5,600	50.6	16.4	6.0	90.6
Male LFPR	5,600	72.9	9.3	40.6	96.2
Percent of internet users	4,445	25.0	28.9	0	99.7
GDP per capita	5,600	8.4	1.5	5.1	11.7
Schooling	5,600	7.3	3.0	0.2	13.2
Women, business and law	5,600	75.7	29.1	0.0	100.0

Source(s): Own estimates

Table 3

Household level results

	(1)	(2)	(3)	(4)	(5)	(6)
Panel A. job loss
Women with internet			−0.0515*	−0.0523**	−0.0510*	−0.0524**
			(0.0263)	(0.0262)	(0.0264)	(0.0264)
Women			0.185***	0.187***	0.178***	0.186***
			(0.0228)	(0.0225)	(0.0227)	(0.0228)
Internet	−0.0798***	−0.0585***	−0.0290	−0.0281	−0.0288	−0.0250
	(0.0153)	(0.0152)	(0.0186)	(0.0185)	(0.0188)	(0.0189)
Household with children				−0.0164
				(0.0123)
Increased household chores					0.0589***
					(0.0137)
Has an smartphone						−0.0642
						(0.0442)
Constant	0.225***	0.212***	0.125***	0.134***	0.111***	0.184***
	(0.0130)	(0.0122)	(0.0162)	(0.0191)	(0.0173)	(0.0442)
Panel B. employment
Women with internet			0.0403*	0.0425*	0.0397	0.0427*
			(0.0245)	(0.0244)	(0.0245)	(0.0245)
Women			−0.241***	−0.248***	−0.236***	−0.244***
			(0.0189)	(0.0187)	(0.0189)	(0.0189)
Internet	0.0612***	0.0315**	0.00223	−0.000671	0.00331	−0.00793
	(0.0144)	(0.0141)	(0.0190)	(0.0189)	(0.0190)	(0.0190)
Household with children				0.0558***
				(0.0124)
Increased household chores					−0.0533***
					(0.0140)
Has an smartphone						0.143***
						(0.0383)
Constant	0.605***	0.623***	0.754***	0.725***	0.766***	0.623***
	(0.0105)	(0.0105)	(0.0150)	(0.0169)	(0.0157)	(0.0390)
Panel C. hours worked
Women with internet			0.0403*	0.0425*	0.0397	0.0427*
			(0.0245)	(0.0244)	(0.0245)	(0.0245)
Women			−0.241***	−0.248***	−0.236***	−0.244***
			(0.0189)	(0.0187)	(0.0189)	(0.0189)
Internet	0.0612***	0.0315**	0.00223	−0.000671	0.00331	−0.00793
	(0.0144)	(0.0141)	(0.0190)	(0.0189)	(0.0190)	(0.0190)
Household with children				0.0558***
				(0.0124)
Increased household chores					−0.0533***
					(0.0140)
Has an smartphone						0.143***
						(0.0383)
Constant	0.605***	0.623***	0.754***	0.725***	0.766***	0.623***
	(0.0105)	(0.0105)	(0.0150)	(0.0169)	(0.0157)	(0.0390)
Observations	26,662	26,507	26,507	26,507	26,507	14,217
R-squared	0.012	0.087	0.135	0.137	0.137	0.143
Country FE	YES
Country-state FE	NO	YES	YES	YES	YES	YES
Education FE	NO	YES	YES	YES	YES	YES
Marital status FE	NO	YES	YES	YES	YES	YES

Note(s): Clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates using HFPS data for LAC

Table 4

Household level results by household chores and children in the households

	(1)	(2)	(3)	(4)
	Household chores increased	Household chores NOT increased	Household with children	Household without children
Panel A. job loss
Women with internet	−0.0904*	−0.0527*	−0.0631**	−0.0440
	(0.0481)	(0.0298)	(0.0292)	(0.0424)
Women	0.217***	0.174***	0.225***	0.125***
	(0.0390)	(0.0254)	(0.0234)	(0.0373)
Internet	0.00182	−0.0269	−0.00649	−0.0384
	(0.0379)	(0.0201)	(0.0190)	(0.0312)
Constant	0.131***	0.115***	0.0944***	0.151***
	(0.0293)	(0.0173)	(0.0149)	(0.0267)
Observations	5,402	11,731	10,982	6,164
R-squared	0.191	0.163	0.173	0.181
Panel B. employment
Women with internet	0.0555	0.0352	−0.000933	0.0974**
	(0.0441)	(0.0277)	(0.0281)	(0.0389)
Women	−0.219***	−0.247***	−0.252***	−0.235***
	(0.0353)	(0.0214)	(0.0210)	(0.0317)
Internet	−0.0309	0.0117	0.0131	−0.0177
	(0.0369)	(0.0209)	(0.0216)	(0.0304)
Constant	0.727***	0.765***	0.784***	0.719***
	(0.0290)	(0.0168)	(0.0163)	(0.0246)
Observations	8,147	18,334	16,594	9,886
R-squared	0.187	0.159	0.169	0.179
Country-state FE	YES	YES	YES	YES
Education FE	YES	YES	YES	YES
Marital status FE	YES	YES	YES	YES

Note(s): Clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1. The difference between groups is only statistically significant for women with internet on employment when comparing by children in the household

Source(s): Own estimates using HFPS data for LAC

Table 5

Household level results by income status and COVID-19 levels

	(1)	(2)	(3)	(4)
	Income reduced	Income NOT reduced	High Covid-19	Low Covid-19
Panel A. job loss
Women with internet	−0.0902**	−0.0525*	−0.0527*	−0.0472
	(0.0428)	(0.0295)	(0.0275)	(0.0362)
Women	0.223***	0.184***	0.187***	0.138***
	(0.0332)	(0.0246)	(0.0240)	(0.0280)
Internet	−0.0517*	0.00330	−0.0274	−0.0629***
	(0.0302)	(0.0180)	(0.0193)	(0.0236)
Constant	0.178***	0.0817***	0.122***	0.193***
	(0.0234)	(0.0144)	(0.0170)	(0.0157)
Observations	5,685	11,273	13,205	3,995
R-squared	0.210	0.147	0.128	0.146
Panel B. employment
Women with internet	0.0401	0.0534*	0.0441*	0.0135
	(0.0395)	(0.0284)	(0.0260)	(0.0337)
Women	−0.253***	−0.251***	−0.247***	−0.176***
	(0.0284)	(0.0222)	(0.0205)	(0.0218)
Internet	0.0336	−0.0206	0.000520	0.0111
	(0.0308)	(0.0212)	(0.0200)	(0.0264)
Constant	0.715***	0.781***	0.759***	0.674***
	(0.0226)	(0.0164)	(0.0161)	(0.0154)
Observations	8,527	17,557	19,744	6,763
R-squared	0.187	0.154	0.135	0.126
Country FE	YES	YES	YES	YES
Country-state FE	YES	YES	YES	YES
Education FE	YES	YES	YES	YES
Marital status FE	YES	YES	YES	YES
Panel A. job loss
Women with internet	−0.0902**	−0.0525*	−0.0527*	−0.0472
	(0.0428)	(0.0295)	(0.0275)	(0.0362)
Women	0.223***	0.184***	0.187***	0.138***
	(0.0332)	(0.0246)	(0.0240)	(0.0280)
Internet	−0.0517*	0.00330	−0.0274	−0.0629***
	(0.0302)	(0.0180)	(0.0193)	(0.0236)
Constant	0.178***	0.0817***	0.122***	0.193***
	(0.0234)	(0.0144)	(0.0170)	(0.0157)
Observations	5,685	11,273	13,205	3,995
R-squared	0.210	0.147	0.128	0.146
Panel B. employment
Women with internet	0.0401	0.0534*	0.0441*	0.0135
	(0.0395)	(0.0284)	(0.0260)	(0.0337)
Women	−0.253***	−0.251***	−0.247***	−0.176***
	(0.0284)	(0.0222)	(0.0205)	(0.0218)
Internet	0.0336	−0.0206	0.000520	0.0111
	(0.0308)	(0.0212)	(0.0200)	(0.0264)
Constant	0.715***	0.781***	0.759***	0.674***
	(0.0226)	(0.0164)	(0.0161)	(0.0154)
Observations	8,527	17,557	19,744	6,763
R-squared	0.187	0.154	0.135	0.126
Country-state FE	YES	YES	YES	YES
Education FE	YES	YES	YES	YES
Marital status FE	YES	YES	YES	YES

Note(s): Clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1. The differences between groups are not statistically significant

Source(s): Own estimates using HFPS data for LAC

Table 6

Country level results

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Female LFPR	Female LFPR	Male LFPR	Male LFPR	Female to male LFPR	Female to male LFPR
Percent of internet users	0.073***	0.065***	0.028***	0.025**	0.085***	0.074***
	(0.013)	(0.013)	(0.009)	(0.011)	(0.017)	(0.015)
GDP per capita		−3.895***		−0.536		−3.630***
		(1.073)		(0.995)		(1.258)
Account		0.753**		0.120		1.102***
		(0.306)		(0.183)		(0.418)
Schooling		0.196		−0.258		0.249
		(0.344)		(0.273)		(0.459)
Women law		0.011		−0.011		0.022
		(0.022)		(0.011)		(0.025)
Constant	48.895***	77.702***	71.866***	79.188***	68.013***	91.787***
	(0.320)	(9.440)	(0.232)	(8.427)	(0.406)	(11.418)
Observations	4,216	4,036	4,216	4,036	4,162	4,036
R-squared	0.970	0.971	0.948	0.945	0.971	0.973
Country FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 7

Regressions with interaction between digital development and household burden

Variables	(1)	(2)	(3)	(4)
	Female LFPR	Female LFPR	Male LFPR	Male LFPR
Percent of internet users	0.098***	0.087***	0.037***	0.035***
	(0.015)	(0.014)	(0.009)	(0.009)
Percent of Internet Users*Trained Teachers	−0.030*	−0.027*	−0.011	−0.012
	(0.016)	(0.014)	(0.009)	(0.009)
Trained teachers	3.159***	3.025***	0.964***	1.031***
	(0.453)	(0.440)	(0.351)	(0.355)
GDP per capita		−3.876***		−0.455
		(0.353)		(0.295)
Account		0.719***		0.115*
		(0.095)		(0.063)
Schooling		0.140		−0.261***
		(0.111)		(0.086)
Women law		0.014**		−0.010***
		(0.006)		(0.004)
Constant	46.351***	75.412***	71.066***	77.660***
	(0.400)	(3.083)	(0.307)	(2.551)
Observations	4,176	4,006	4,176	4,006
R-squared	0.970	0.971	0.948	0.946
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 8

Regressions with interaction between digital development and income levels

Variables	(1)	(2)	(3)	(4)
	Female LFPR	Female LFPR	Male LFPR	Male LFPR
Percent of internet users	0.078***	0.069***	0.028***	0.026**
	(0.014)	(0.014)	(0.010)	(0.011)
Percent of Internet Users*Low Income	−0.048	−0.007	−0.032	−0.021
	(0.059)	(0.057)	(0.062)	(0.062)
Percent of Internet Users*Middle Income	−0.049***	−0.030*	−0.017	−0.011
	(0.017)	(0.016)	(0.013)	(0.012)
GDP per capita		−3.272***		−0.303
		(1.042)		(0.909)
Account		0.726**		0.100
		(0.303)		(0.174)
Schooling		0.203		−0.254
		(0.347)		(0.273)
Women law		0.012		−0.010
		(0.022)		(0.011)
Constant	49.252***	72.498***	72.016***	77.258***
	(0.342)	(9.408)	(0.269)	(7.695)
Observations	4,216	4,036	4,216	4,036
R-squared	0.970	0.971	0.948	0.946
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 9

Regressions with interaction between digital development and LAC dummy

Variables	(1)	(2)	(3)	(4)
	Female LFPR	Female LFPR	Male LFPR	Male LFPR
Percent of internet users	0.069***	0.060***	0.028***	0.026**
	(0.013)	(0.013)	(0.009)	(0.011)
Percent of Internet Users*LAC	0.052**	0.053**	−0.009	−0.008
	(0.021)	(0.022)	(0.012)	(0.013)
GDP per capita		−3.765***		−0.557
		(1.063)		(0.996)
Account		0.707**		0.127
		(0.300)		(0.182)
Schooling		0.260		−0.268
		(0.339)		(0.274)
Women law		0.004		−0.010
		(0.022)		(0.011)
Constant	48.801***	76.756***	71.883***	79.341***
	(0.325)	(9.316)	(0.231)	(8.433)
Observations	4,216	4,036	4,216	4,036
R-squared	0.970	0.971	0.948	0.946
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 10

Regressions with interaction between digital development and service sector size

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Female LFPR	Female LFPR	Male LFPR	Male LFPR	Female to male LFPR	Female to male LFPR
Percent of internet users	0.097***	0.080***	0.020	0.020	0.110***	0.091***
	(0.021)	(0.019)	(0.017)	(0.016)	(0.022)	(0.021)
Percent of Internet Users*Service Sector Size	−0.019	−0.010	0.009	0.007	−0.018	−0.010
	(0.025)	(0.024)	(0.020)	(0.020)	(0.025)	(0.025)
Service sector size	−3.785***	−3.100***	−0.696	−0.618	−4.264***	−3.716***
	(0.785)	(0.695)	(0.663)	(0.575)	(0.966)	(0.971)
GDP per capita		−2.103**		−0.106		−1.574
		(1.016)		(0.930)		(1.267)
Account		0.558**		0.062		0.886**
		(0.276)		(0.174)		(0.392)
Schooling		0.148		−0.271		0.196
		(0.338)		(0.274)		(0.447)
Women law		0.007		−0.012		0.017
		(0.022)		(0.010)		(0.025)
Constant	51.315***	65.761***	72.320***	76.304***	70.712***	78.116***
	(0.477)	(8.864)	(0.461)	(7.978)	(0.623)	(11.204)
Observations	4,216	4,036	4,216	4,036	4,162	4,036
R-squared	0.972	0.972	0.948	0.946	0.975	0.974
Country FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 11

Regressions with interaction between digital development and COVID dummy

Variables	(1)	(2)	(3)	(4)
	Female LFPR	Female LFPR	Male LFPR	Male LFPR
Percent of Internet Users	0.054***	0.041***	0.017	0.014
	(0.015)	(0.014)	(0.012)	(0.013)
Percent of Internet Users*COVID-19	0.024*	0.029**	0.013	0.014
	(0.014)	(0.013)	(0.010)	(0.010)
GDP per capita		−3.795***		−0.504
		(1.085)		(1.016)
Account		0.715**		0.103
		(0.299)		(0.183)
Schooling		0.340		−0.194
		(0.363)		(0.267)
Women law		0.010		−0.012
		(0.022)		(0.011)
Constant	48.871***	76.062***	71.840***	78.552***
	(0.312)	(9.657)	(0.234)	(8.638)
Observations	4,199	4,025	4,199	4,025
R-squared	0.970	0.971	0.947	0.945
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 12

Household level results using logistic regression

	(1)	(2)	(3)	(4)	(5)	(6)
	Coefficients	Odds ratio	Coefficients	Odds ratio	Coefficients	Odds ratio
Panel A. job loss
Women with internet					−0.0235	0.977
					(0.195)	(0.190)
Women					1.203***	3.331***
					(0.158)	(0.527)
Internet	−0.543***	0.581***	−0.416***	0.659***	−0.384**	0.681**
	(0.0993)	(0.0577)	(0.103)	(0.0677)	(0.173)	(0.118)
Constant	−0.849***	0.428***	−1.140	0.320	−2.037**	0.130**
	(0.132)	(0.0566)	(0.835)	(0.267)	(0.863)	(0.113)
Panel B. employment
Women with internet					0.134	1.144
					(0.124)	(0.142)
Women					−1.135***	0.321***
					(0.0954)	(0.0307)
Internet	0.268***	1.307***	0.148**	1.159**	0.0372	1.038
	(0.0630)	(0.0824)	(0.0643)	(0.0746)	(0.105)	(0.109)
Constant	0.124	1.132	0.263	1.301	1.010**	2.746**
	(0.0940)	(0.106)	(0.452)	(0.589)	(0.421)	(1.157)
Obs	26,662	26,662	26,448	26,448	26,448	26,448
State FE	NO	NO	YES	YES	YES	YES
Education FE	NO	NO	YES	YES	YES	YES
Marital status FE	NO	NO	YES	YES	YES	YES

Note(s): Clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates using data from the HFPS for LAC

Table 13

Country level results using alternative variables

Variables	(1)	(2)	(3)	(4)
	Female employment ratio	Male employment ratio	Number of female employees	Number of male employees
Percent of internet users	0.080***	0.032***
	(0.013)	(0.011)
Number of internet users			0.027***	0.037***
			(0.007)	(0.005)
GDP per capita	−1.615	0.802	−0.302***	−0.256***
	(1.220)	(1.209)	(0.082)	(0.061)
Account	0.744**	0.066	0.011	−0.009
	(0.300)	(0.226)	(0.014)	(0.007)
Schooling	0.295	−0.107	−0.006	−0.011
	(0.359)	(0.321)	(0.019)	(0.015)
Women law	0.008	−0.009	0.001	0.001
	(0.021)	(0.015)	(0.001)	(0.001)
Constant	52.430***	60.614***	21.817***	21.764***
	(10.724)	(9.972)	(0.646)	(0.504)
Observations	4,036	4,036	4,029	4,029
R-squared	0.965	0.940	0.996	0.997
Country FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Table 14

GMM results

Variables	(1)	(2)	(3)	(4)
	Female LFPR	Female LFPR	Female employment ratio	Female employment ratio
Percent of internet users	0.002*	0.004*	0.004***	0.005**
	(0.001)	(0.002)	(0.001)	(0.002)
GDP per capita		−0.358*		−0.222
		(0.194)		(0.212)
Account		0.029		0.097*
		(0.052)		(0.053)
Schooling		0.119*		0.119*
		(0.063)		(0.066)
Women Law		−0.008*		−0.010**
		(0.004)		(0.005)
Constant	6.986***	10.661***	8.921***	11.045***
	(0.522)	(1.805)	(0.599)	(1.974)
Observations	3,987	3,816	3,904	3,785

Note(s): Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Own estimates

Notes

1. Owing to the lack of consistent data on sectoral employment by gender in the respective household sectors for all countries, this study does not address the comparative and the telework ability and female employment sectoral intensity.

2. It should be noted that digitalization is a multifaceted concept, and the ITU offers additional indicators, such as mobile and fixed broadband subscriptions and the number of mobile subscriptions. Due to its multidimensional nature, it cannot be fully assessed solely through internet user statistics. For the coherency of the analysis, however, we used one definition of digitalization – the percentage of internet users in the total population. See also, Reis et al. (2020) for broader definitions of digitalization and digital transformation. Habibi and Zabardast (2020), the authors used the percentage of Individuals using the Internet as their measure of digitalization. They found that digitalization positively contributes to economic growth in both countries. In Myovella et al. (2020), the authors used the percentage of Individuals using the Internet to measure digitalization. They found that digitalization positively contributes to economic growth in both countries. In Mgadmi et al. (2021), the authors investigated the contribution of digitalization on economic growth in both developed and developing countries from 1990 to 2020. They showed that the digital technologies seem to significantly and positively affect economic growth in both groups of countries. And in Lechman and Anacka (2022), this paper analyzed the role of digitalization in economic growth, with the percentage of internet users being one of the primary measures of digitalization.

3. See Appendix II for a complete list of the countries in the analysis.

4. In Appendix 1 we present the expanded tables from this section that include the effect of internet use alone, and the effect of the interaction with gender.

5. Valberg (2019) used panel data analysis for 156 countries from the period 1991–2014, and argued that ICTs contribute positively to narrowing the gender gap in labor market participation, mainly due to increased female labor force participation.

6. The study by Asongu and Odhiambo (2023) complements the extant literature (e.g. Cherif et al., 2020). By assessing how increasing penetration levels of ICT affect female economic inclusion and by extension, thresholds necessary for the promotion of ICT to increase female economic inclusion and growth.

Supplementary material

The supplementary material for this article can be found online.