ABSTRACT
Recent years have witnessed an increasing interest in the time-series properties of historical cost (HC) accounting data (e.g., Foster, 1977; Brown and Rozeff, 1978; Hopwood and McKeown, 1981; Fries, 1985), but research on the statistical characteristics of accounting signals generated by inflation adjusted income measures have been lacking (e.g., Hillison et al., 1983; Shalchi and Smith, 1985). Indeed, virtually nothing is known about the degree to which actual accounting time-series differ under various alternatives, presumably due to the unavailability of such data. This void has usually been filled by making a priori assumptions about the processes generating such data. For instance, studies by Beaver et al. (1982) and Beaver and Landsman (1983) on the information content of constant dollar accounting (CDA) earnings make implicit assumptions, with no prior empirical evidence, about the earnings-expectation models appropriate for such data. 1 Given the fact that the results of such studies are earningsexpectation dependent, possible misspecification of the earnings-expectation models is likely to induce erroneous inferences on their findings. The present study provides explicit empirical evidence regarding the time-series characteristics of various deflated and nondeflated annual CDA-based income measures and the effects of monetary gains or losses on such series.
