[In a recent issue of The Accounting Review, Stone and Bublitz (S&B) [1984] report on the frequency of data collection errors for the 1981 FASB computer tape containing FAS 33 disclosures for 1979 and 1980. This paper extends their work by examining the problems encountered when using a subsequent version of that data base in conjunction with the Compustat and Value Line historical cost data bases. Failure of the data bank user to reconstruct the difference between historical cost and current cost amounts in the supplementary disclosures of a firm's annual report is designated as an error. Errors in income and depreciation expense for a filtered sample of 171 firms are reported for the years 1979 through 1982. An error rate for income in 1981 of 12.9 percent and error rates for depreciation ranging from 23.4 to 39.8 over the four years seem to suggest that manual checking of annual reports is necessary when using these amounts.]

[The accuracy of forecasts of one-year-ahead current cost earnings using four component models is examined. ASR 190 data for 1976 to 1978 and FAS 33 data for 1979 to 1981 are used in calculating current cost earnings for a sample of 129 firms. The component forecasting models also incorporate forecasts of certain components of historical cost net income developed by Value Line financial analysts. The forecast accuracy of the component models is compared to four benchmark models. A component model that incorporates analysts' forecasts of sales and changes in (historical cost) margins provides the most accurate forecasts for the aggregate sample and for two subsamples. It also provides similar or more accurate forecasts than the other models in eight of the nine industries examined. Thus, researchers may consider using this model when examining the incremental information content of current cost earnings vis-a-vis that of historical cost earnings.]

In a recent issue of The Accounting Review, Stone and Bublitz (SERB) [1984] report on the frequency of data collection errors for the 1981 FASB computer tape containing FAS 33 disclosures for 1979 and 1980. This paper extends their work by examining the problems encountered when using a subsequent version of that data base in conjunction with the Compustat and Value Line historical cost data bases. Failure of the data bank user to reconstruct the difference between historical cost and current cost amounts in the supplementary disclosures of a firm's annual report is designated as an error. Errors in income and depreciation expense for a filtered sample of 171 firms are reported for the years 1979 through 1982. An error rate for income in 1981 of 12.9 percent and error rates for depreciation ranging from 23.4 to 39.8 over the four years seem to suggest that manual checking of annual reports is necessary when using these amounts.

The accuracy of forecasts of one-year-ahead current cost earnings using four component models is examined. ASR 190 data for 1976 to 1978 and FAS 33 data for 1979 to 1981 are used in calculating current cost earnings for a sample of 129 firms. The component forecasting models also incorporate forecasts of certain components of historical cost net income developed by Value Line financial analysts. The forecast accuracy of the component models is compared to four benchmark models. A component model that incorporates analysts' forecasts of sales and changes in (historical cost) margins provides the most accurate forecasts for the aggregate sample and for two subsamples. It also provides similar or more accurate forecasts than the other models in eight of the nine industries examined. Thus, researchers may consider using this model when examining the incremental information content of current cost earnings vis-a-vis that of historical cost earnings.