Cash flows, Accruals and Corporate returns


Published: Mar 7, 2025
Augustinos Dimitras
George Peppas
Konstantinos Seremetis
Efstathios Magerakis
Abstract

This paper investigates the comparative predictive value of accrual-based earnings and operating cash flows for future corporate performance, focusing on listed non-financial companies in the Athens Stock Exchange over the 2013–2022 period. Guided by three research objectives, we first test whether accounting earnings outperform operating cash flows in forecasting next-year return on assets (ROA). Our findings strongly support the notion that accrual-based earnings carry significant informational advantages over raw operating cash flows, explaining a substantially higher percentage of variation in subsequent ROA. Second, we assess the incremental benefit of disaggregating earnings into its accrual and cash flow components. Results indicate that while both components positively predict next-year ROA, the cash flow portion exhibits higher persistence, yet accruals also provide distinct, value-relevant signals. A combined model that distinguishes cash and accrual elements offers modest but statistically significant improvements in predictive accuracy. Third, we analyze whether the interaction of high accruals and low cash flows—often viewed as low-quality earnings—undermines future performance relative to firms exhibiting low accruals and strong cash flows. Our findings confirm that companies with heavily accrual-based earnings and weak cash generation underperform in subsequent years. By contrast, firms presenting robust cash flows and moderate or even high accruals often maintain strong future returns, suggesting that high accruals are detrimental only when not supported by sufficient cash inflows. Overall, our study underscores the value of accrual accounting in performance prediction, while highlighting the necessity of scrutinizing the interplay between cash flows and accruals for a more comprehensive assessment of earnings quality.

Article Details
  • Section
  • Articles
Downloads
Download data is not yet available.