Please use this identifier to cite or link to this item:
Year of Publication:
ITS Online Event, 14-17 June 2020
As the reach of mobile technology grows, it is becoming an increasingly powerful tool for access to welfare-enhancing information and services in low- and middle-income countries (LMICs). However, digital inclusion remains far from universal. Across LMICs, 14 per cent of adults still do not own a mobile phone, 39 per cent do not use mobile internet, and 38 per cent do not own a smartphone. Among other characteristics, these digitally excluded individuals are predominantly female. This study seeks to better understand the key drivers of mobile ownership, mobile internet, and smartphone use, with a particular focus on gender. Discrete-choice models, including logit, probit and linear models, are used to estimate the probability of adoption of these three types of technology. By including a suite of control variables for observable drivers of mobile adoption (e.g. education levels, age, employment, rural-urban location), the coefficient for gender represents non-observable effects which could be a product of discrimination and cultural norms. Furthermore, importance is placed on the inclusion of interaction terms in the regressions (for example, gender interacted with rural location), in order to isolate different degrees of marginalisation across the female population. In addition to the focus on gender, the marginal effects of the dependent variables for other factors (such as geography, education, employment, and age) will aid understanding of the key predictors of mobile use more generally. This research also shows how these predictors might vary by country and region, how they relate to each other, and which are the most important. This will provide relevant and important information for policy-makers. The research makes use of multiple years (2017, 2018, and 2019) of data from face-to-face consumer surveys sourced from the GSMA, which includes nationally representative samples of at least 1,000 respondents for 31 low- and middle-income countries. The wide geographic scope, and multi-year nature of the survey data results in a unique contribution to the literature, and the substantial number of observations allows for novel analysis of intersections of the female population. In summary, the initial results find that: women are 5 percentage points (pp) less likely to own a mobile phone then men, 6pp less likely to use mobile internet, and 4pp less likely to own a smartphone, even when other relevant socioeconomic and demographic factors are controlled for. This unobservable gender effect is more pronounced in certain regions, especially South Asia, but with no significant link in Latin America and Caribbean. The marginal effects of the interaction variables indicate that the negative impact is enhanced for women that live in rural areas, have low levels of literacy, and are not working. In addition, this study finds that the probability of mobile technology adoption increases (with varying magnitudes by technology type and region) with income, education, urban location, literacy, and employment. Adoption of mobile technology largely declines with age, but the impact generally does not appear to start until age 45 and above for mobile ownership.
Appears in Collections:
Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.