PERBANDINGAN ALAT ANALISIS (DISKRIMINAN & REGRESI LOGISTIK) TERHADAP PERINGKAT OBLIGASI (PT PEFINDO)

Abstract

Bond rating is a guide to investor who intends fo buy bond because that rating gives informa¬tive statement and giving signal about company's debt failure possibility. The purposes of this re¬search are getting more empirical evidendonce about the ability of financial ratios in forming model that used to predict bond rating. this research also test the difference of two statistical models (descriminant analysis and logistic regression) in predicting bond rating. The results show that descriminant analysis and logistic regression are not so different although descriminant analysis re¬sults more significant ratios. It used 38 manufactures companies with 160 bond issuances from 1999 to 2004. The bond predicting models have accuracy 96,9 % for descriminant analisis and 91,3 % for logistic regression. It means both of two models can be used as tool in predicting bond rating for Indonesian manufacturer company in Indonesia. Keywords : bond rating, financial ratios (leverage, likuidity, solvability, profitability and productivity)