This study aims to find an appropriate theoretical framework to address the driving mechanisms of government digitalization on changes in government functions. Our empirical strategy mainly consists of two distinctive steps. The first step involves digital transformation and digital technologies related to public finance, while the second step describes the estimation procedure. To focus on the public finance side, we subdivide government digitalization into several dimensions (at least 4 dimensions) and its functions alterations into at least two related dimensions in tax collection performance. Variables are defined based on the process of digitalization. In the second stage, applying panel data estimating techniques, we used DEA and SFA derived efficiency data to regress the efficiency of the selected tax units on the independent and control variables. Including 16 selected emerging economies over the 2018–2022 period, our results indicate that digital technologies such as e-invoicing, artificial intelligence related to machine learning, cloud computing, data science, chatbot, and e-payment have a positive and significant effect on tax collection efficiency and productivity. Furthermore, digital transformation in the financial sector, education, online services, and other departments have an enhancing effect on the efficiency. Additionally, our robustness checks indicate that the endogeneity effect in the model cannot be rejected and previously obtained results are consistent. It is worth noting that some part of inefficiency can be justified by the theory of “rational inefficiency”.