Estimation of Bank Profitability Using Vector Error Correction Model and Support Vector Regression

Abstract

The research problem defined in this article is the identification and the metrification of the macro-economic drivers of bank profitability of the Bulgarian banks. The aim is to establish the relationship between the macroeconomic factors and the indicators measuring bank profitability using an iterative testing methodology of the Vector Error Correction Model (VECM) and the Support Vector Regression (SVR). The relationships between key indicators of bank profitability - ROE and ROA, are analytically modeled as factor-dependent on nine key macroeconomic indicators. The sample size includes quarterly and monthly data for the recovery period of the banking system after its restructuring in 2014. The empirical results of the VECM form two similar models, which are distinguished based on dependent variable ROA and dependent variable ROE. They show the significance of the established results at high degrees of determination and minimal degrees of standard errors of the cointegration vectors. The analysis of bank profitability has been refined using the SVR methodology, but the optimization of the hyperparameters does not confirm high significance. It is concluded that with increasing regulatory requirements for the formation of buffers against macroprudential risks, banks report a relatively neutral dependence in ROA and ROE on the impact of macroeconomic factors, which is logically offset by effective management decisions and innovations in interbank competition.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 19 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.