A note on the time series behavior of earnings per share data of Taiwanese firms

The time series properties of earnings per share data and other accounting numbers has been extensively examined and the motivations for such studies include the forecasting of future corporate profits (Firth, 1975) and the derivation of earnings expectation models for use in research and public policy formulation. While this research has been conducted in a variety of 'western-style' capitalist economies, there have been few, if any, studies conducted in the fast growing Asian markets. The purpose of this note is to briefly report the results of a study that investigated the time series properties of the annual earnings per share (e.p.s.) numbers of Taiwanese companies. The results from other studies cannot be automatically imputed to Taiwan as the very rapid growth of their economy and the close Government involvement with, and control exercised over, the corporate sector created conditions very different from those seen in 'western-style' capitalist systems. The rapid growth of the economy and the interventionist approach of the Government to promote certain companies and industrial sectors suggest conditions may exist which make corporate profits follow different time series patterns than those seen in Australasia, Europe, and North America. In particular, the random walk type models that seemingly represent annual e.p.s. numbers in other studies may not be applicable to the Taiwanese environment. Given that time series models are used for benchmarking corporate profits in a variety of research settings it seems appropriate for the basic statistical properties of Taiwanese firms' earnings per share data to be analysed and reported.