Date of Award

2010

Degree Type

Thesis

Degree Name

Bachelor of Business Honours

School

School of Business

Faculty

Faculty of Business and Law

First Advisor

Malcom Smith

Abstract

This thesis examines the predictive ability of corporate financial disclosures. Factor Analysis and Discriminant Analysis are used to differentiate between good companies and poor companies in Australian manufacturing industry. Good and poor companies are identified based on their 2009 financial data and the prediction models are constructed based on their 2008 data. 64 companies are selected finally, with 29 good companies and 35 poor companies. Financial ratios, company size, corporate governance and conservatism are employed in this study to examine whether they can predict corporate perforn1ance. Because only 3 7 companies disclosed research and development (R&D) expenses, which are used to measure conservatism among 64 companies, two models are derived in this study, one without the conservatism variable and one with the conservatism variable. All four categories of variables are found to have predictive value. The model without the conservatism variable has quick assets ratio, company size and 'percentage of shareholdings held by executive directors' representing corporate governance as its independent variables and the other model has conservatism, total debt/total assets, company size, and 'percentage of shareholdings held by non-executive directors' representing corporate governance as its independent variables. The classification accuracy of the two models is 72.6% and 80.6% respectively. Due to the small sample size, the predictive ability of the two models is also evaluated with 2009 financial data. The accuracy of the model without the conservatism variable based on 2009 financial data is 64.1% and that of the model with the conservatism variable is 62.2%.

Included in

Accounting Commons

Share

 
COinS