A large number of authors have developed statistical models, which are based solely on conventional financial ratios constructed from published accounting data, with the aim of predicting corporate failure as evidenced by the event of ‘bankruptcy’. The purpose of this paper is to report some empirical results for a study of the UK corporate sector in which corporate failure is predicted employing a statistical model which incorporates both conventional accounting ratios and a number of new variables which are not derived from profit and loss accounts and balance sheet items, but which are computed from annual company reports and accounts. The empirical results suggesting that our new variables enhance the predictive power of models which employ conventional financial ratios only.