ASSESSING PERCEIVED RISK IN MOBILE MONEY ADOPTION UNDER COVID-19: A COMBINED SEM-ARTIFICIAL NEURAL NETWORK TECHNIQUES

Abstract

The introduction of social distancing measures to curb the COVID-19 pandemic and support the stabilization of the social economy has motivated consumers to do contactless activities, including mobile money service (MMS). Although this service remains beneficial to consumers, the adoption rate is still at its formative stage in Togo. The socio-economic background and peoples' inclination are hesitant for such rising digital transactions, seemingly due to risk perception. Therefore, the study develops a model to capture multidimensional perceived risk regarding the adoption decision. A total of 275 respondents were tested using a hybrid structural equation modeling (SEM) and artificial neural network (ANN) approach through a multilayer perceptron (MLP) with feed-forward back-propagation (FFBP) algorithm. The ANN model is found to seize better performance and high prediction accuracy than SEM regarding nonlinearity and linearity. Our results suggest that perceived privacy risk (PRR) stands out as the most critical antecedent of the perceived overall risk (POR), in which the latter negatively affect the behavioral intention (BI) to use MMS. This research remains one of the first to test the acceptance of MMS empirically during the COVID-19 crisis and contributes both theoretically and practically toward understanding factors influencing its widespread adoption. To promote citizen's trust, service providers must provide instructions on using MMS safely and coping with privacy breaches and security problems if they arise. The SEM-ANN methodology will aid fulfill the current literature gap of MMS acceptance and provide practical guidance for evidence-based decision-making.