The Profit Rate-Interest Rate Nexus: Evidence from Machine Learning Algorithms


The main purpose of this study is to examine potential predictors of profit rates and dee posit rates and to examine whether these rates are affected by identical factors. This paper empirically addresses tree-based machine learning algorithms (e.g., boosting, bagging, random forest). The empirical findings of the study demonstrate participation banks’ profit rates to be more influenced by industrial production due to these banks being in contact more with real economic activity. As expected, however, domestic and global interest rates appear to have great significance in how deposit banks set their rates. This study contributes to the literature in two ways. First, it determines the potential predictors of profit rates and deposit rates in a data-rich environment. Second, the study uses random forest, bagging, and boosting algorithms as methodological tools and benefits from the apparent advantages these algorithms have empirically.