ENTROPY BASED CORRELATION MATRIX APPROACH IN GENERALIZED ESTIMATING EQUATION: TURKISH BANKING SECTOR APPLICATION

Banking sector which is the headstone of the financial sector in Turkey should profit. There are some indicators available to measure profitability of a bank and monitor the sustainability of the efficiency. These indicators are called financial ratios. In this study, 3-month dataset including ratios and income statements of deposit bank, development bank and investment bank in Turkey in 2010 was created by using dataset of Bank Association of Turkey (BAT) and Banking Regulation and Supervision Agency (BRSA). The main purpose of this study is to determine the structure of relation between repeated measures and explore the suitable correlation structure of the dataset via generalized estimation equation (GEE). As a result of the analysis it was revealed that entropy based correlation coefficient which is not available in statistical package program and suitable for nominal variables could be an alternative for other known correlation coefficients. According to the results of this study, it was stated that capital adequacy and profitability ratios are the most important factors for estimating the change of total assets.