University of Isfahan Financial Accounting Research 2322-3405 16 3 2024 11 21 A Comparative Study of Cash Flow Forecasting with Cash and Accrual Items: Evidence from Cross-Sectional Heterogeneity and Transitory Items A Comparative Study of Cash Flow Forecasting with Cash and Accrual Items: Evidence from Cross-Sectional Heterogeneity and Transitory Items 31 60 29526 10.22108/far.2025.144226.2101 FA Milad Darvishi Ph.D., Accounting Department, Faculty of Economics and Administrative Sciences, Ferdowsi University of Mashhad, Mashhad, Iran. Roghayeh Mahmoudi Yekebaghi Ph.D., Accounting Department, Faculty of Economics and Administrative Sciences, Ferdowsi University of Mashhad, Mashhad, Iran. Mohammad Javad Saei Assistant Professor, Accounting Department, Faculty of Economics and Administrative Sciences, Ferdowsi University of Mashhad, Mashhad, Iran. Journal Article 2025 02 02 The information needs of investors for making informed decisions, as well as companies' need to effectively manage their resources, have consistently drawn researchers' attention to the prediction of future operating cash flows. However, prior empirical evidence on cash flow prediction models has been mixed. This study examines the importance and impact of considering transitory items in accrual accounting, as well as cross-sectional heterogeneity among firms, on the conflicting prior evidence regarding the comparison between accrual-based earnings and operating cash flows in predicting future operating cash flows and related contexts. To test the hypotheses, a sample of 62 companies listed on the Tehran Stock Exchange from 2013 to 2022 was used. Four different earnings components, along with operating cash flows, were analyzed under several forecasting approaches, including cross-sectional, pooled, industry-level, and firm-level estimations, in both univariate and bivariate comparisons. The results indicate that accounting earnings that exclude transitory items have better predictive power for future operating cash flows than operating cash flows themselves. Furthermore, accounting for cross-sectional differences in the relationship between firms' earnings and future cash flows enhances the predictive ability of accruals and increases the reliability of accounting earnings in forecasting future operating cash flows. Introduction A company’s success significantly depends on the effective management of its resources. Firms must maximize their resources to overcome various challenges that impact business development. In the face of increasing competition in the modern era, entrepreneurs must adopt strategies to ensure the sustainability of their businesses (Noury et al., 2020). In this age of globalization, a company may either perform as planned or encounter unforeseen challenges leading to unfavorable outcomes, including bankruptcy. One of the most critical factors in assessing a company’s success or failure is the prediction of future cash flows, as it helps evaluate the business’s ability to generate cash, its need to utilize these flows, and enables more informed decision-making. Furthermore, cash flow information serves as an indicator of a company’s actual success or performance, making performance evaluation more meaningful (Sharawi, 2021). A fundamental assumption in accounting is that financial statements prepared under accrual methods contain more informational content than cash-based accounting. Consequently, the informational value of accrual-based earnings exceeds that of cash flows (Ball & Nikolaev, 2022). Prior research indicates that realized cash flows suffer from timing and matching issues, making them a noisy performance metric. Accruals mitigate this problem (Dechow, 1994). The adoption of accrual accounting is essential for financial reporting, playing a key role in providing a comprehensive and accurate picture of financial performance and reporting entities’ positions. Empirically, researchers have sought to assess the value-added of accrual accounting by comparing accrual-based performance metrics (e.g., earnings) with cash-based metrics (Columbano, 2023). Extensive evidence suggests that accrual measures (particularly operating earnings) possess desirable properties, such as smoother earnings, enabling better prediction of future cash needs (Ball & Nikolaev, 2022; Dechow & Dichev, 2002; Kim & Kross, 2005). According to Nallareddy et al. (2020), prior studies primarily relied on bottom-line earnings to predict future cash flows. Meanwhile, most earlier research examined cross-sectional regression models, though some studies emphasized the need for firm-level analysis (Finger, 1994). To clarify, this study moves beyond the prior literature’s focus on bottom-line earnings and examines the limitations of cross-sectional regression in inferences about accruals. Methods & Material To test the hypotheses, a sample of 62 companies listed on the Tehran Stock Exchange from 2013 to 2022 was used, employing four distinct earnings components plus operating cash flows across several prediction approaches—including cross-sectional, pooled, industry-level, and firm-level estimations—in both univariate and bivariate comparisons. This study evaluates the predictive abilities of several earnings variables. First, earnings before discontinued operations (IBC), as reported in financial statements, serves as a proxy for bottom-line earnings. IBC excludes extraordinary gains/losses (which are rare) but does not exclude non-operating and transitory (unstable) earnings components, making it the weakest predictor of future operating cash flows. Non-operating accruals typically contain less informational content or introduce noise when explaining future cash flows (Barth et al., 2001). Next, adjusted earnings before discontinued operations (IBCA) is calculated by removing items with no equivalent in future operating cash flows, including extraordinary items, discontinued operations, and gains/losses from the sale of fixed assets and investments. IBCA is expected to outperform IBC in predicting future operating cash flows. As an accrual-based earnings measure, operating earnings (OP) is used. To compare earnings with operating cash flows, working capital accruals (i.e., operating accruals) are added to operating cash flows, following Dechow and Dichev (2002) and Barth et al. (2001), referred to as OE (Operating Earnings). By definition, working capital assets and liabilities have a cycle of one year or less. Thus, working capital accruals adjust current earnings for cash flows generated by current-year operating activities but realized in cash either in the prior year or expected in the next year. Since operating cash flows arise from transactions with a one-year cycle (while longer-term transactions are classified as investing/financing cash flows), working capital accruals are incorporated into OE. Conceptually, OE is an accrual-based earnings measure expected to align best with operating cash flows. OE excludes all long-term (non-operating) accruals and cash flows embedded in bottom-line earnings (e.g., gains from asset sales). These earnings components do not naturally map to next-period operating cash flows, thereby adding noise in predictive regressions. Since OE removes the most non-operating noise in predicting operating cash flows, it is expected to exhibit the highest predictive power. Findings Cross-sectional analyses show that operating cash flows have greater predictive power than earnings variables, with only operating earnings improving model predictability. Pooled estimations reveal that when cross-sectional heterogeneity in the earnings–future cash flow relationship is controlled via a simple firm fixed-effects model, earnings’ predictive power surpasses that of operating cash flows, supporting the hypothesis that accrual-based earnings outperform cash flows. Industry-level analysis indicates that operating earnings dominate operating cash flows in predictive ability while also incorporating cash flow information. Firm-level analysis shows that firm-level heterogeneity is a confounding factor in cross-sectional regressions, with accrual-based earnings outperforming cash flows in prediction. Conclusion & Results The findings demonstrate that operating earnings outperform operating cash flows in predicting future operating cash flows. This earnings measure includes accruals that adjust for the timing limitations of operating cash flows, highlighting the incremental information provided by accrual accounting. Our evidence aligns with the argument that operating cash flows are a noisy measure of operating earnings, mitigated by accruals (Dechow, 1994). Additionally, results indicate that the relationship between current earnings and future operating cash flows varies across firms due to differences in industries, business models, operating cycles, growth rates, accounting methods, and other factors. Accounting for this heterogeneity significantly alters results, improving the predictive power of earnings over operating cash flows. These findings support Ball and Nikolaev’s (2022) study. Overall, accrual-based earnings metrics provide a superior basis for predicting future cash flows, consistent with both academic (Dechow, 1994) and professional literature (AICPA, 1973; FASB, 1978). The results also underscore the importance of addressing heterogeneity in cross-sectional models. The information needs of investors for making informed decisions, as well as companies' need to effectively manage their resources, have consistently drawn researchers' attention to the prediction of future operating cash flows. However, prior empirical evidence on cash flow prediction models has been mixed. This study examines the importance and impact of considering transitory items in accrual accounting, as well as cross-sectional heterogeneity among firms, on the conflicting prior evidence regarding the comparison between accrual-based earnings and operating cash flows in predicting future operating cash flows and related contexts. To test the hypotheses, a sample of 62 companies listed on the Tehran Stock Exchange from 2013 to 2022 was used. Four different earnings components, along with operating cash flows, were analyzed under several forecasting approaches, including cross-sectional, pooled, industry-level, and firm-level estimations, in both univariate and bivariate comparisons. The results indicate that accounting earnings that exclude transitory items have better predictive power for future operating cash flows than operating cash flows themselves. Furthermore, accounting for cross-sectional differences in the relationship between firms' earnings and future cash flows enhances the predictive ability of accruals and increases the reliability of accounting earnings in forecasting future operating cash flows. Introduction A company’s success significantly depends on the effective management of its resources. Firms must maximize their resources to overcome various challenges that impact business development. In the face of increasing competition in the modern era, entrepreneurs must adopt strategies to ensure the sustainability of their businesses (Noury et al., 2020). In this age of globalization, a company may either perform as planned or encounter unforeseen challenges leading to unfavorable outcomes, including bankruptcy. One of the most critical factors in assessing a company’s success or failure is the prediction of future cash flows, as it helps evaluate the business’s ability to generate cash, its need to utilize these flows, and enables more informed decision-making. Furthermore, cash flow information serves as an indicator of a company’s actual success or performance, making performance evaluation more meaningful (Sharawi, 2021). A fundamental assumption in accounting is that financial statements prepared under accrual methods contain more informational content than cash-based accounting. Consequently, the informational value of accrual-based earnings exceeds that of cash flows (Ball & Nikolaev, 2022). Prior research indicates that realized cash flows suffer from timing and matching issues, making them a noisy performance metric. Accruals mitigate this problem (Dechow, 1994). The adoption of accrual accounting is essential for financial reporting, playing a key role in providing a comprehensive and accurate picture of financial performance and reporting entities’ positions. Empirically, researchers have sought to assess the value-added of accrual accounting by comparing accrual-based performance metrics (e.g., earnings) with cash-based metrics (Columbano, 2023). Extensive evidence suggests that accrual measures (particularly operating earnings) possess desirable properties, such as smoother earnings, enabling better prediction of future cash needs (Ball & Nikolaev, 2022; Dechow & Dichev, 2002; Kim & Kross, 2005). According to Nallareddy et al. (2020), prior studies primarily relied on bottom-line earnings to predict future cash flows. Meanwhile, most earlier research examined cross-sectional regression models, though some studies emphasized the need for firm-level analysis (Finger, 1994). To clarify, this study moves beyond the prior literature’s focus on bottom-line earnings and examines the limitations of cross-sectional regression in inferences about accruals. Methods & Material To test the hypotheses, a sample of 62 companies listed on the Tehran Stock Exchange from 2013 to 2022 was used, employing four distinct earnings components plus operating cash flows across several prediction approaches—including cross-sectional, pooled, industry-level, and firm-level estimations—in both univariate and bivariate comparisons. This study evaluates the predictive abilities of several earnings variables. First, earnings before discontinued operations (IBC), as reported in financial statements, serves as a proxy for bottom-line earnings. IBC excludes extraordinary gains/losses (which are rare) but does not exclude non-operating and transitory (unstable) earnings components, making it the weakest predictor of future operating cash flows. Non-operating accruals typically contain less informational content or introduce noise when explaining future cash flows (Barth et al., 2001). Next, adjusted earnings before discontinued operations (IBCA) is calculated by removing items with no equivalent in future operating cash flows, including extraordinary items, discontinued operations, and gains/losses from the sale of fixed assets and investments. IBCA is expected to outperform IBC in predicting future operating cash flows. As an accrual-based earnings measure, operating earnings (OP) is used. To compare earnings with operating cash flows, working capital accruals (i.e., operating accruals) are added to operating cash flows, following Dechow and Dichev (2002) and Barth et al. (2001), referred to as OE (Operating Earnings). By definition, working capital assets and liabilities have a cycle of one year or less. Thus, working capital accruals adjust current earnings for cash flows generated by current-year operating activities but realized in cash either in the prior year or expected in the next year. Since operating cash flows arise from transactions with a one-year cycle (while longer-term transactions are classified as investing/financing cash flows), working capital accruals are incorporated into OE. Conceptually, OE is an accrual-based earnings measure expected to align best with operating cash flows. OE excludes all long-term (non-operating) accruals and cash flows embedded in bottom-line earnings (e.g., gains from asset sales). These earnings components do not naturally map to next-period operating cash flows, thereby adding noise in predictive regressions. Since OE removes the most non-operating noise in predicting operating cash flows, it is expected to exhibit the highest predictive power. Findings Cross-sectional analyses show that operating cash flows have greater predictive power than earnings variables, with only operating earnings improving model predictability. Pooled estimations reveal that when cross-sectional heterogeneity in the earnings–future cash flow relationship is controlled via a simple firm fixed-effects model, earnings’ predictive power surpasses that of operating cash flows, supporting the hypothesis that accrual-based earnings outperform cash flows. Industry-level analysis indicates that operating earnings dominate operating cash flows in predictive ability while also incorporating cash flow information. Firm-level analysis shows that firm-level heterogeneity is a confounding factor in cross-sectional regressions, with accrual-based earnings outperforming cash flows in prediction. Conclusion & Results The findings demonstrate that operating earnings outperform operating cash flows in predicting future operating cash flows. This earnings measure includes accruals that adjust for the timing limitations of operating cash flows, highlighting the incremental information provided by accrual accounting. Our evidence aligns with the argument that operating cash flows are a noisy measure of operating earnings, mitigated by accruals (Dechow, 1994). Additionally, results indicate that the relationship between current earnings and future operating cash flows varies across firms due to differences in industries, business models, operating cycles, growth rates, accounting methods, and other factors. Accounting for this heterogeneity significantly alters results, improving the predictive power of earnings over operating cash flows. These findings support Ball and Nikolaev’s (2022) study. Overall, accrual-based earnings metrics provide a superior basis for predicting future cash flows, consistent with both academic (Dechow, 1994) and professional literature (AICPA, 1973; FASB, 1978). The results also underscore the importance of addressing heterogeneity in cross-sectional models.

https://far.ui.ac.ir/article_29526_c1dfccbc3cda34daff1e9ebc51350a92.pdf