Predictors of financial sustainability for cryptocurrencies: An empirical study using a hybrid SEM-ANN approach

Abstract

Cryptocurrencies stand as a significant financial innovation, poised to transform traditional financial systems by facilitating faster, cost-efficient, and inclusive value transfers. Understanding their sustainability is essential in comprehending their potential to reshape established financial frameworks. Hence, this study aimed to identify the factors predicting the financial sustainability of cryptocurrencies. The study proposed a model based on the “Expectation Confirmation Theory” (ECT) and tested the model using a multi-analytical approach by combining “Structural Equation Modeling” (SEM) and “Artificial Neural Network” (ANN). The sample of the study included 1649 participants, ranging in age from 17 to 70. Results indicated that perceived risk, regulation, price volatility, innovativeness, and confirmation of expectations significantly predicted perceived usefulness, which in turn significantly predicted satisfaction with cryptocurrencies. The findings highlighted that users' degree of confirmation of expectations significantly influenced their satisfaction with and perceived usefulness of cryptocurrencies. The specified paths within the model accounted for 61 % and 74 % of the variance in perceived usefulness and financial sustainability, respectively. In comparison to the results obtained through SEM analysis, the deep ANN multi-layer perceptron displayed superior performance in predicting perceived usefulness. This was evident from its enhanced predictive accuracy, achieving averages of 87.34 % for training and 87.76 % for testing.