The Effect of CAR, NPL, and LDR on ROA of SOE Banks in Indonesia (Case Study at a State-Owned Bank Listed on the IDX)

Herman Budi Sasono ,Mawarto

doi:10.7176/rjfa/11-10-14

Abstract

Bank health is an important thing in the banking world, because an unhealthy bank will be a threatmenr for the bank's internal element that affected the customers. This research is to find out and analyze the effect of CAR, NPL and LDR partially or simultaneously on the ROA of BUMN banks, for examples, Bank Mandiri, Bank Negara Indonesia (BNI 46), Bank Rakyat Indonesia (BRI) and Bank Tabungan Negara (BTN) since 2009-2017. Data collection techniques was by accessing www.idx.com. This type of research was quantitative research. The analytical tool used in this study was multiple linear regression analysis, using E-Views software. Chow Test and Hausman Test was used to choose the best multiple linear regression analysis between Common Effect, Fixed Effect, and Random Effect.The partial test results show that the NPL variable has a significant effect on ROA of state-owned banks in Indonesia. Simultaneous test results show that CAR, NPL and LDR significantly affected ROA Keywords: bank health, IDX DOI: 10.7176/RJFA/11-10-14 Publication date: May 31 st 2020

Highlights

From the results of testing the research variables simultaneously showed a significant relationship between the variables Capital Adequacy Ratio (CAR), Non-Performing Loans (NPL) and Loan to Deposite Ratio (LDR) on Return On Assets (ROA)
LDR simultaneously have a significant effect on ROA in state-owned banks in Indonesia, accepted
If the three variables are in a healthy level, it can be said that bank profits are proxied by ROA, so the child increases

Summary

Hausman Tes

Seeing from the probability value of F and Chi-square with the assumption: a. a. Seeing from the probability value of F and Chi-square with the assumption: a. A. If the probability value F and Chi-square >α = 5%, the panel data regression test uses the Random. If the probability value of F and Chi-square

No LM test

The best model is FE with a regression equation

Findings

DISCUSSION