A Multivariate Time-Series Prediction Model For Cash-Flow Data
开发了一个多元时间序列预测模型，利用历史盈余、短期应计和现金流数据预测未来现金流，在1989-1991年样本中优于传统ARIMA模型和截面回归模型。
This paper provides evidence on the time-series properties and predictive ability of cash-flow data. It employs a sample of firms on which the accuracy of one-step-ahead cash-flow predictions is assessed during the 1989- 1991 holdout period. We develop a new multivariate, time-series prediction model that employs past values of earnings, short-term accruals and cash-flows as independent variables in a time-series regression. Our predictive results indicate that this model clearly outperforms firm-specific and common-structure ARIMA models as well as a multivariate, cross-sectional regression model popularized in the literature. These findings are robust across alternative cash-flow metrics (e.g., levels, per-share, and deflated by total assets) and are consistent with the viewpoint espoused by the FASB that cash-flow prediction is enhanced by consideration of earnings and accrual accounting data.