METHODOLOGICAL PRINCIPLES OF A COMPREHENSIVE STUDY OF BANKING POTENTIAL USING NEURAL NETWORK FORECASTING

Abstract

Ukraine’s banking sector operates under persistent shock conditions created by war, global turbulence, and accelerated structural change. In this setting, banking potential is treated not as a snapshot of isolated indicators but as a system-level capacity to transform available resources into strategic change while preserving solvency, liquidity, and public trust. The article develops a methodology for comprehensive assessment and forward-looking projection of banking potential that integrates four interacting blocks-resource, internal process, external context, and result-measured on compatible, risk-adjusted scales and tracked through time. The framework departs from static rating logics by emphasizing temporal dynamics, cross-block feedbacks, and sustainability constraints relevant to post-war recovery. Methodologically, the study specifies a composite potential index built from normalized financial and non-financial factors (capital adequacy, funding structure, asset quality, operating cost architecture, digital and human capital). For forecasting, the article applies neural architectures tailored to sequences (e.g., gated recurrent units and long short-term memory networks) to learn nonlinear dependencies across the four blocks and project the joint trajectory of potential and efficiency. Model interpretability is increased through constraint design, out-of-sample stability checks, and contrasted partial-impact analyses on the most influential drivers. The proposed approach yields several practical outputs for banks and supervisors. First, it distinguishes sustainable capacity from performance spikes caused by one-off deals or transient macro tailwinds, reducing the risk of pro-cyclical decision-making. Second, it provides an early-warning profile by linking deterioration in specific resource or process factors to expected declines in the result block, enabling timely capital planning, liquidity buffers, and product repricing. Third, it embeds sustainability by design: social, environmental, and governance attributes are incorporated as amplifiers or dampeners of potential under stress, which aligns strategic choices with long-term resilience targets for the economy. Fourth, the method scales across institutions and over time, supporting benchmarking, policy calibration, and evaluation of recovery strategies. The article’s results demonstrate that robust assessment of banking potential requires moving beyond static profitability, liquidity, or capital ratios toward a dynamic, system-aware view that captures how resources, processes, external conditions, and outcomes coevolve. Such a perspective is essential for Ukraine’s post-war rebuilding, where institutional resilience and efficient allocation of scarce resources will determine the speed and stability of recovery.