Beyond Bank Accounts: The Phone Skills and Security Gaps Behind Digital Payment Use

Opening an account or owning a mobile phone does not tell us whether someone will pay a shopkeeper digitally, or what might make that transaction easier. Research across 74 economies links internet use, basic messaging skills and independent management of phone security with greater use of digital merchant payments, raising questions about how progress in financial inclusion is measured.

Published in the Journal of Risk and Financial Management, the study, Beyond Account Ownership: Digital Capability, Secure Device Control, and Merchant-Payment Use Across 74 Economies, examined 62,480 phone-owning adults using World Bank Global Findex survey data collected in 2024. The researchers also studied historical indicators, drawing a distinction between gaining access to financial services and using them for everyday purchases.

Account ownership tells only part of the story

Account ownership rose from 53% in 2011 to 80.6% in 2024 across 108 economies, and making or receiving digital payments increased from 35.6% in 2014 to 62.4% in 2024 across 74 economies. Mobile-money account ownership climbed from 4.1% to 29.9% over the same decade across 50 economies. Digital merchant payments recorded a smaller increase, from 36.3% in 2021 to 39.9% in 2024 across 74 economies.

Growth did not produce a uniform pattern of narrowing gender and income gaps, with conclusions changing according to the indicator, period and weighting of economies. The evidence could not establish that account-access gaps consistently closed faster than payment-use gaps.

In the separate 2024 core sample of phone owners, around 50% had a digitally enabled account, 54.8% reported any digital-payment use, and 26.1% reported merchant payments. These are sample-specific figures, and merchant payments include purchases using cards, phones or online channels during the preceding year.

Messaging skills and control over phone security stand out

The researchers accounted for age, gender, education, income, participation in the labour force, rural residence and characteristics shared within each economy. Their model associated smartphone ownership with a 2.8-percentage-point higher predicted probability of merchant-payment use, compared with 6.9 points for recent internet use, 7.7 points for basic messaging capability and 8.3 points for secure device control. Messaging capability meant being able to read and send text or app messages, a limited measure that does not establish broader financial literacy.

Secure device control meant protecting a phone through a PIN, password or biometric lock and being able to change credentials without help. A separate analysis distinguished unlocked phones, locked phones whose owners needed help changing credentials, and locked phones whose owners could manage credentials independently.

Adjusted payment probabilities were approximately 20.2%, 23.8% and 29.5%, respectively. Among people with locked phones, independent credential management was associated with a 5.71-percentage-point higher predicted probability of payment use, with a 95% uncertainty interval of 3.87 to 7.51 points.

Education and income remain powerful dividing lines

Education and income differences remained substantial after digital characteristics entered the models. After adjustment, predicted merchant-payment use was 19.6% among adults with primary education or less and 40.1% among those with tertiary education, leaving an adjusted gap of about 20.4 percentage points. Predicted use rose from 20.5% in the poorest household-income fifth to 31.5% in the richest fifth within economies. Adding digital characteristics reduced the education and income gaps in the models; it did not prove that those characteristics caused the reductions.

Rural residents retained an adjusted disadvantage of 3.5 percentage points, and women had a smaller remaining disadvantage of about 1.2 points. Exploratory tests suggested that the association between secure device control and payment use was stronger among women and in Latin America and the Caribbean, excluding high-income economies. Those patterns survived the researchers' correction for multiple comparisons, leaving a question for future research about which groups would benefit most from a security programme.

Useful financial inclusion requires skills and practical protection

Checks using four weighting approaches preserved the positive main associations, and the internet, messaging and security findings held up when regions were excluded in turn. Tests focused on digitally paid online purchases, using eligible samples of 40,678 and 42,529 adults, produced a similar pattern for internet use, messaging and secure device control. Smartphone ownership had a less certain association in these narrower online models, reinforcing the distinction between possessing a device and having the capability to use digital services.

The authors propose testing programmes that combine connectivity with practical instruction in messaging, authentication and fraud recognition. Accessible ways to recover account credentials, secure default settings and wider merchant acceptance could support more useful participation, particularly in underserved areas. Measuring actual transactions alongside account ownership would give policymakers a fuller picture of participation, and the study calls for future research to evaluate whether these measures improve payment use and financial well-being.

The findings describe associations from a single survey period, so they cannot show that teaching messaging skills or changing phone-security practices would increase payments. Existing payment users may become more attentive to security, and unmeasured experience could influence both behaviours. Detailed questions were available only for selected phone owners in 74 economies, with just one high-income economy in the core sample. The results cannot represent all adults worldwide, establish transaction safety or demonstrate improved financial resilience; longer-term and experimental studies are needed to test the proposed explanations.