Prediction Development and Variable Definitions

The purpose of this three-step analysis is to investigate cash flow and accruals’ predictive ability and value relevance, and then to examine possible relations between steps 1 and 2. Specifically, step 1 analyses the predictive ability of cash flow and accruals for short term future cash flows and earnings. Step 2 examines the value relevance of the same explanatory variables. Step 1 can be regarded as an indirect investigation of value relevance (compare Finger’s statement from section 1 that the value relevance of earnings is studied by testing their ability to predict future earnings and cash flow from operations). This section starts out by discussing possible relations between the prediction analysis of step 1 and the association studies of step 2. The discussion is founded on traditional valuation theory.

The ultimate return of every investment is the cash flow generated by the investment. Financial theory says that an asset’s value is the present value of its future cash flows. The current value of firm equity is the present value of all future dividends (the dividend model is often attributed to Williams, 1938). The dividend model can be restated in several equivalent forms. For instance, Feltham and Ohlson (1995) show that, under some fairly reasonable

assumptions11, equity value (EV) is today’s value of net financial assets plus the present value of all future free cash flows from operating activities:

(

)

(

)

∑

∞ = + + = 1 t t t t 0 0 r 1 CFO E NFA EV where

NFA = net financial assets (negative if debts exceed gross financial assets) CFO = free cash flow from operating activities

r = discount rate

Calculation of cash flow predictions is a vital task in asset valuation. One objective (compare Holthausen & Watts, 2001) of financial reporting is to assist investors, creditors, and others in predicting cash flow. This is accomplished through the accruals, and the FASB asserts that information about earnings and its components (i.e., cash flow and accruals) is generally more predictive of future cash flows than current cash flow. However, modern valuation theory has proven that different versions of cash flow discounting are not the only methods for computing intrinsic equity values. Ohlson (1995) shows that if the clean surplus relation holds, then the dividend model can also be expressed as the book value of equity plus the present value of all future residual income:

(

)

(

)

∑

∞ = + + = 1 t t t t 0 0 r 1 RI E BV EV 11

Specifically, the Financial Asset Relation (FAR) and the Financial Asset Marked-to-Market Relation (FAM) must hold. FAR says that all transfers to the common equity holders are made through the financial assets, and these assets are further influenced by financial income and the free cash flows from operations. FAM says that the risk-adjusted expected financial income equals the riskless spot interest rate times the opening book value of the financial assets (Christensen & Feltham, 2003).

where

BV = book value of equity

RI = residual income, defined as EARN_t−BV_t−1*r_t. EARN is accounting earnings. r = discount rate

Hence, equity value can be computed both as a function of cash flows and as a function of accounting earnings. Both of the presented models are deduced from the dividend model and will therefore give the exact same equity estimate when applied consistently. Today’s accounting earnings should help investors and others assess the level of future cash flow and

the level of future earnings. In theory, equity value is a function of the infinite cash flows or earnings. However, one may expect that shorter term firm performance, as measured by cash flow and earnings, gives an indication of company value. For instance, numerous studies document that both cash flow and earnings are relatively persistent (Barth et al., 2001; Dechow et al., 1998; Finger, 1994; M. Kim & Kross, 2005; Sloan, 1996). Thus, the current accounting measures that predict short term future firm performance should also be associated with current stock returns. This assumption is implicit in most papers studying the accounting variables’ ability to forecast future cash flows or earnings; compare the vast amount of what I earlier referred to as indirect value relevance research. I expect that the accounting measures that are significant predictors of future firm performance are also significantly related to current stock return. Specifically, if accruals and cash flows are related to short term future firm performance as measured by accounting earnings and cash flow, it is reasonable to expect that they are also significantly associated with current stock return.

The prediction is one-directional. It is easy to construct examples in which cash flow and accruals are totally unrelated to short term firm performance but are still value relevant. Transitory cash flow and/or earnings items may be associated with stock return even if they, by definition, are not related to future cash flows or earnings. However, if current earnings are highly related to earnings and cash flows one to three years ahead, I suggest that current earnings on average will also be related to current stock return. Prior research strongly proposes that permanent cash flow or earnings are more value relevant than transitory cash flows or earnings (Elliot & Hanna, 1996; Ramakrishnan & Thomas, 1998). Overall, my prediction suggests that there is an association between accounting variables’ predictive ability and their value relevance, but that the association may not be as one-to-one as implicitly assumed in the indirect value relevance studies (compare O. Kim et al., 2007). It should be noted that if there is no association between accounting variables’ relation with future firm performance and their value relevance, then previous studies of accounting variables’ predictive ability would give no indication of value relevance whatsoever.

The primary focus of many value relevance and prediction analyses (e.g., Ball & Shivakumar, 2006; Barth et al., 2001; Francis, LaFond, Olsson, & Schipper, 2005; Rayburn, 1986) is the role of accruals. Accruals, and consequently earnings, are a function of the prevailing accounting regime. Accountants cannot, in principle12, influence the size of the cash flow. In contrast, the size of the accruals follows from subjective judgements of legislators, standard setters and accountants. Thus, particular attention is given to the predictive ability and value relevance of accruals. As for cash flows, there are several versions of the cash flow valuation

12

The net cash flow, the change in cash, is a one hundred percent objective measure. No accounting law or standard can change this measure. However, as soon as one starts to use other measures of cash flow, for instance, cash from operations, the prevailing accounting regime may perfectly well influence the measure. For example, the treatment of intangibles will directly influence the cash flow measure applied in this study; see later in this section.

model. It is not obvious which cash flow measure should be analysed13. The appendix examines this issue more thoroughly. However, since this study focuses on the possible connection between cash flow/earnings predictions and value relevance, it can be regarded as a follow-up paper of prior research. I leave, therefore, the choice of cash flow concept to future research and choose to employ the same cash flow definitions as the other papers within this research tradition (compare Subramanyam & Venkatachalam, 2007). Hence, cash flow from operations (CF) is defined as (Biddle et al., 1995; Finger, 1994; Klein & Marquardt, 2006; Rayburn, 1986)14:

CF = Net income before extraordinary items15 (EARN) - Accruals (ACC) where

Accruals = Change16 in total current assets - Change in cash

- Change in total current liabilities

+ Change in interest bearing short term debt17 - Change in deferred taxes

- Depreciation and impairment

I focus on EARN before extraordinary items, since extraordinary items are expected to have little persistence (Barth et al., 2005; Dechow & Ge, 2006) and bear little value relevance

13

Although not specifically related to this discussion, Bowen et al. (1986) look into the relationship between various measures of cash flow.

14

According to Biddle et al. (1995), this cash flow measure is designed to approximate the definition of cash flow from operations specified by FASB’s Statement No. 95.

15

The ratio of earnings to cash flow from operations is affected by leverage. Since leverage varies over time and across companies, one might argue that earnings exclusive of interest payments should have been applied in the analysis.

16

The change variables are defined as follows: 1 t 1 t t MVE X X − −

− . X is the accounting variables in question, while MVE is market value of equity, compare also section 4.

17

(Bradshaw & Sloan, 2002; Landsman et al., 2007; Ramakrishnan & Thomas, 1998). All variables – CF, EARN and ACC – are scaled by market value of equity at the beginning of each year; this is the preferred scaling factor according to Easton and Sommers (2003). In this study, observations are pooled cross-sectionally and over time. Christie (1987) states that scaling avoids spurious correlation due to size and reduces problems with heteroskedasticity. In comparable studies, there seems to be no standard for which variable is preferable for scaling. Some researchers choose to deflate by average total assets (for instance Barth et al., 2001; Francis & Smith, 2005; Sloan, 1996), while others scale by market value of equity (for instance Biddle et al., 1995; Dechow, 1994; Francis et al., 2003; Rayburn, 1986). Barth and Kallapur (1996) claim that scale should be handled by including a scale proxy as an independent variable in cases where the true scale factor is not known. Easton and Sommers (2003), however, state that the scale factor is known in market-based accounting research. They maintain that market capitalisation is more than just a possible scale proxy - it is scale. In this study, the choice of scale factor follows from my econometric models (see the next section). Step 2 of the analysis uses a return specification. In the return specification, market value of equity is the scale factor. To assure a consistent analysis of my proposed prediction, it is essential that the same scale factors are used in both steps of the study.

In document Essays on the value relevance of accounting information (Page 107-112)