عنوان مقاله [English]
نویسندگان [English]چکیده [English]
This Study is aimed at examining the usefulness of financial statements information. In particular, usefulness criteria of financial statements depend upon relevance and predictive ability.
So far many models and hypotheses have been developed in order to evaluate and predict stock return through different viewpoints. A large number of empirical accounting researches have been done in order to achieve this goal. Different groups of investors and decision makers are interested in evaluation and prediction of Stock return.
This guided research procedure is different from the mere statistical searches because, we have chosen the fundamental variables through theory and famous models. We developed Samuel Stewart’s model and ended up selecting 42 independent Variables (fundamentals). Eighty two listed companies were selected from Tehran Stock Exchange. The selection was done from various industries.
Regression cross-sectional models were developed for the firms within the years 1374 to 1380; 1378 to 1380 and, for each single year of the research period.
We constructed the following hypotheses:
1- The financial Statements information has ability to predict stock return.
2- Using Accounting models for selected industries, industries, increase the predictive ability of financial statements information.
3- Using models based upon sign of variables increase the predictive ability of financial statements information.
The concluding results of this study, confirms the predictive ability of accounting information. Variables such as rate of return on assets (ROA), rate of return on investments (ROI), growth of sales to total assets ratio (GSTTA), growth of net income to sales ratio (GNITS) and financial expenses to sales (FEXTS), have had the most influence among 9 regressed models, in prediction of stock return.
The performance of models in short run was better than long run. Using models for special selected industries (drug and chemical; mineral and cement) improved the predictive ability of our selected variables. But the models based on sign of variables did not increase the predictive ability of the models.