Choosing an Appropriate Model for Predicting Earnings Based on Comparing the Relevant Models in the Tehran Stock Exchange

The present study investigates the prediction earning models and compares the models based on absolute forecast errors. For analyzing the data 232 listed firms Tehran stock exchange (TSE) are used during 1380-1387. Comparing the models is done at the level of the all selected firms and at the level of the industries. Using panel data analyses, the results show that disaggregation earning has more ability to predict future earnings than aggregation earning. The results also show that there is a significant relationship between operating cash flows components - cash receipts from customers, cash paid to suppliers and cash paid for operating expenses -and future earnings. It means that these items have ability for predicting future earnings. Also the results indicate that there is a significant relationship between accruals components -accounts receivables, prepayments, inventory, depreciation and amortization, accounts payable advanced receipts, other accounts receivables, other accounts payable and other accruals and future earnings. As a result, these items have ability for predicting future earnings.