KEMAMPUAN RASIO KEUANGAN DALAM MEMPREDIKSI LABA (Penetapan Rasio Keuangan sebagai Discriminator)

Abstract

Financial statements users need financial information of companies to analyze their financial condition and performance. Financial ratios are useful measures for predicting the future earning changes. The study focuses on the usefulness of financial ratios in predicting future earnings.
The objective of the study is to empirically examine whether financial statement-based financial ratios have ability for predicting future earnings and which ratios are significant discriminator. Predictive accuracy of financial ratios are measured with high hit ratio from classification matrix list on 2 group are sample analysis and holdout sample.
Data in this study were manufacturing firms listed on the Jakarta Stock Exchange. Discriminant analysis were used in testing the ability financial ratios for predicting earnings changes. The multicollinearity test shows that there is no association between independent variables, indicating multicolinearity is not a serious problem. The heteroscedasticity test shows that variances of disturbances are constant for all observation in independent variables. Therefore, heteroscedasticity is not a problem.
Because the objective of these analysis was to determine which variables are the most efficient in discriminating earning changes, a stepwise procedure was used. There are five significant ratios (DIV/NI, S/TA, LTD/TA, NI/S, and INPPE/TU). DIV/NI ratio is ratio which most discriminating, while INPPE/TU ratio which least discriminating.