«Leif Atle Beisland University of Agder Dissertation submitted to the Department of Accounting, Auditing and Law at the Norwegian School of Economics ...»
Teets, W. R., & Wasley, C. E. (1996). Estimating earnings response coefficients: Pooled versus firm-specific models. Journal of Accounting & Economics, 21(3), 279-295.
Thomas, W. B. (1999). A test of the market's mispricing of domestic and foreign earnings.
Journal of Accounting & Economics, 28(3), 243-267.
Tsuji, C. (2006). Does EVA beat earnings and cash flow in Japan? Applied Financial Economics, 16(16), 1199-1216.
Vincent, L. (1997). Equity valuation implications of purchase versus pooling accounting.
Journal of Financial Statement Analysis, 2(4), 5.
Williams, J. (1938). The Theory of Investment Value. Harvard University Press, Cambridge.
accounting measures. Step 1 investigates the ability of cash flow and aggregate accruals, i.e., earnings, to predict short term firm performance. Future earnings and cash flow from operations are applied as measures of firm performance. One to three year-ahead periods are examined. I find that cash flow is significantly associated with both future cash flow and future earnings. There is strong evidence that cash flow is related to future short term performance, as it does not matter whether next year or the mean of the next three years is investigated. Accruals, on the other hand, appear to be unrelated to future short term cash flow. They are, however, significantly associated with future earnings, both next year’s earnings and the mean of the three next earnings.
Step 2 of the study investigates cash flow and accruals’ relation with stock return, i.e., their value relevance. As the market value of equity is equal to the discounted value of all future cash flows/earnings, step 2 measures the association of cash flow and accruals with long term (infinite) firm performance. The value relevance study presents evidence that both cash flow and accruals are highly related to contemporaneous stock return. The analysis shows that stock return is positively related to cash flow, while it is negatively related to accruals. In other words, for a given earnings level, the investors react more favourably when the cash flow components of earnings is larger (compare Sloan, 1996; Wilson, 1986). The results of steps 1 and 2 are, however, heavily dependent on the sign of earnings. When split into cash flow and accruals, earnings appear to have almost no predictive ability or value relevance when they are negative (compare Hayn, 1995).
ability for future cash flow and earnings, and their value relevance, i.e., possible relations between step 1 and step 2. In general, value relevance research studies the statistical relationship between market values of stocks and accounting information. However, since stock value is the present value of future cash flow/earnings, value relevance can also be indirectly assessed by studying accounting information’s ability to forecast future cash flow/earnings (Barth, Cram, & Nelson, 2001; Dechow, Kothari, & Watts, 1998; Finger, 1994;
M. Kim & Kross, 2005). For instance, Finger (1994) states, “This paper examines the value relevance of earnings by testing their ability to predict two future benefits of equity investment: earnings and cash flow from operations” (Finger, 1994, p. 210). Francis and Schipper (1999) deepen Finger’s statement: “…financial information is value relevant if it contains the variables used in a valuation model or assists in predicting those variables” (Francis & Schipper, 1999, p. 325). As such, it can be claimed that prediction studies in the accounting literature implicitly and sometimes explicitly (e.g., Finger, 1994) assume that there is a close relation between those prediction studies and the value relevance research, i.e., between indirect and direct analysis of value relevance. Still, while stock value is the present value of all future cash flow/earnings, studies of cash flows’ or earnings’ predictive ability for future firm performance often consider rather short time horizons. In fact, sometimes only next year’s cash flow or next year’s earnings are assessed (e.g., M. Kim & Kross, 2005;
Sloan, 1996). Step 3 of the study investigates whether or not analyses of accounting variables’ short term predictive abilities of firm performance can act as substitutes for analyses of the same variables’ association with stock value, i.e., their value relevance. I predict that if cash flow and accruals are related to short term firm performance, it is reasonable to expect that the variables are also value relevant. I find empirical support for this prediction. However, it is neither a necessary nor a sufficient condition that accounting variables are related to short
accounting variables’ association with future short term cash flow and/or earnings may provide indications with respect to the variables’ value relevance, the two types of studies are not equivalent.
Several papers explore the predictive ability of current cash flow and accruals with respect to future cash flows. Most of them conclude that earnings are a better cash flow predictor than current cash flow (see for instance Barth, Cram, & Nelson, 2001; Dechow, Kothari, & Watts, 1998). As for earnings predictions, several studies find that current earnings are a significant predictor of future earnings (Finger, 1994; Francis & Smith, 2005; Sloan, 1996). Sloan (1996) reports that the accruals component of earnings is less persistent than the cash flow component. The value relevance of accruals is also heavily investigated in prior research.
Most empirical studies find that earnings, relatively speaking, are a more value relevant measure than cash flow (Biddle, Seow, & Siegel, 1995; Dechow, 1994; Rayburn, 1986;
Subramanyam & Venkatachalam, 2007). In general, prior capital market-based accounting research has tended to focus on either the accounting variables’ relation to stock values or their relation to future firm performance (cash flow and/or earnings), i.e., either direct or indirect investigation of value relevance. The study by Kim and Kross (2005) is one of very few studies that discuss possible relations between the two lines of research. Kim and Kross investigate current earnings’ ability to predict future cash flow, and they anticipate that this ability has decreased over time. They cite several studies that show that the value relevance of earnings has decreased over the prior decades: “If the relationship between current earnings and prices is decreasing, the relationship between current earnings and future cash flows should also be decreasing” (M. Kim & Kross, 2005, p. 759). However, Kim and Kross document that the relationship between current earnings and future operating cash flow has
studies of value relevance may produce contradictory results. Kim and Kross suggest several reasons for their findings. For instance, their findings are consistent with accounting information generally bearing little value relevance. They also note that the findings may be because they are investigating the ability of earnings to predict next year’s cash flow, while stock prices are the discounted value of all future cash flows. An additional explanation, which Kim and Kross do not note, might be that while stock valuation models discount free cash flows, Kim and Kross only study cash flow from operations. Note also that Kim et al.
(2007) provide evidence that cash flow predictions are contaminated by noise in the cash flows and a spurious (value unrelated) correlation between one-year-ahead cash flows and current earnings. Consistent with Kim and Kross’ (2005) results on intertemporal variations, I also find that earnings may be more value relevant in periods where they appear less able to explain future firm performance.
This paper contributes to the existing research by further investigating the relationship between the time-series properties of accounting variables and their value relevance, i.e., the accounting variables’ indirect and direct value relevance. I extend the analysis of Kim et al.
(2007) by also studying earnings’ predictive ability for longer term cash flows, i.e., the three next annual cash flows, as well as for future earnings. Valuation theory, or more generally finance theory, has traditionally had a “cash is king” perspective. However, equity valuation today does not solely focus on cash flow discounting. In fact, when using the residual income model, the value driver of firm equity is accounting earnings, not cash flow. Earnings forecasts have become just as important as cash flow forecasts1. Special attention is drawn to In practice, financial analysts tend to focus on earnings forecasts and compute cash flows from pro-forma income statements and balance sheets (compare for instance Penman, 2001). As such, one of the objectives of accounting accruals should be to help investors and other users of accounting information assess the amount and timing of both earnings and cash flow. The cash flow model and the residual income model are equivalent in
judgment from managers and accountants. Earnings are, through the accruals, also a function of the prevailing accounting framework and regulations. Cash flow is claimed to be an objective measure of firm performance, a figure that in principle is unaffected by accounting laws and standards. The Financial Accounting Standards Board (FASB) has, on several occasions, stated that accruals make earnings a better predictor of future cash flows than current cash flow2. A primary objective of financial reporting is to provide information to help investors, creditors, and others assess the amount and timing of prospective cash flows. This is accomplished through the accrual adjustment process. However, the FASB does not define future cash flows further. If the FASB by future cash flow refers to rather short term future cash flow, an assumption often applied in empirical research (see for instance Finger, 1994), this paper provides evidence against the FASB assertion. On the other hand, if the FASB is referring to the relation to all future cash flows, i.e., the market value of equity (compare discussion in Kim et al, 2007), this paper provides evidence consistent with the FASB assertion. The analysis is performed on Norwegian data, but the conclusions can likely be generalised to other countries as well. Specifically, during the last decades, Norwegian accounting legislation has changed from a tax-based and relatively conservative model to an Anglo-American investor-oriented accounting model (Gjerde, Knivsflå, & Sættem, 2007b).
This paper is organised as follows: Section 2 summarises relevant conclusions from prior research. Section 3 describes the research design of the paper and develops the prediction to be tested. Section 4 presents the data employed in the analyses, while section 5 summarises theory, but it is sometimes claimed that inconsistent implementation may cause the two models to differ (Lundholm, O'Keefe, & Feltham, 2001; Lundholm & O'Keefe, 2001; Penman, 2001b; Penman & Sougiannis, 1998).
See, for instance, FASB’s Objective of Financial Reporting by Business Enterprises (1978).
2 Prior Research The time series properties of accounting numbers have been heavily investigated empirically.
For instance, the FASB-statement that current earnings provide better forecasts of future cash flows than do current cash flow is a frequently studied issue. One such study is conducted by Bowen et al. (1986). Bowen et al. find no support for the FASB assertion. Using five different cash flow measures, they conclude that earnings numbers do not provide better forecasts of future cash flows than do cash flow numbers. Their findings are supported by Finger (1994).
She finds that cash flow is a better short term predictor of cash flows than are earnings, but that the two are approximately equivalent in the long term. On the other hand, Dechow et al.
(1998) state that current earnings are a better forecast of future cash flows than current cash flow. In a regression of cash flow on lagged values of both earnings and cash flow, they find that earnings are consistently incrementally useful in forecasting future cash flows, while cash flows themselves exhibit only modest incremental forecasting power3.
Barth et al. (2001) disaggregate earnings into cash flow and accruals in order to predict future cash flows. They conclude that this disaggregation significantly improves the explanatory power of the specification (see also Barth, Beaver, Hand, & Landsman, 2005). They also run regressions in which accruals are further disaggregated into major components. Each accrual component is significant, and the explanatory power increases further. The cash flow and accrual components of earnings actually have substantially more predictive ability for future Chen et al. (2006) maintain that fair value accounting reduces the predictive ability of earnings for future cash flows.
US data. Nikkinen and Sahlstrøm (2004) show that equivalent conclusions are valid in Canada, France, Japan, and the UK as well. Francis et al. (2004) state that accrual quality – the degree to which current accruals map into future cash flows – is associated with lower cost of equity.