«Leif Atle Beisland University of Agder Dissertation submitted to the Department of Accounting, Auditing and Law at the Norwegian School of Economics ...»
Prediciton of earnings is another subject frequently discussed within this line of research. In fact, earnings persistence is considered as an important attribute of earnings quality (see for instance Francis et al., 2004). Finger (1994) reports that current earnings are a significant predictor of future earnings in 88 % of her sample. She tests the association between current earnings and a varying number of lags of historical earnings and finds that the ability of earnings to predict themselves increases as more earnings lags are included in the specification. In fact, when only two earnings lags are used, the firm-specific model is outperformed by the random walk (as measured by root mean squared errors). However, the firm-specific model is superior when four or eight lags are used. Sloan (1996) estimates his models both in pooled form and on an industry-level and reports that current earnings are a highly significant predictor of next year’s earnings. He rejects the hypothesis that earnings follow a random walk. His findings suggest that accounting earnings are slowly mean reverting4 (an excellent study of the mean reverting properties of earnings is provided by Fama & French, 2000). Hope (2004) reports that forecast accuracy is positively correlated with greater use of accrual accounting, but that availability of choice among accounting methods is negatively associated with forecast accuracy.
Ben-Hsien and Da-Hsien (2004) present empirical results that earnings smoothers may have higher earnings – to-price multiples than non-smoothers.
components of earnings exhibit lower persistence than the cash flow components of earnings.
This conclusion is confirmed in Collins and Hribar (2000). Francis and Smith (2005) state that the accounting-based measures of cash flow and accruals do not align with current-period income. Using alternative measures of accruals and cash flows, they find that the differential persistence of cash flows over accruals is more than 70 % smaller than when using the traditional definitions5. Still, Chan et al. (2004) find that the aggregate future earnings will decrease by $0.046 and $0.096, respectively, in the next one and three years for a $1 increase in current accruals. They also state that the empirical results are consistent with the notion that earnings management causes the negative relationship between current accruals and future earnings.
Lev et al. (2005) study the predictive ability of both cash flow and accruals on future cash flow and earnings. They find that accruals and their embedded estimates do not improve the prediction of cash flow beyond that achieved by current cash flows. Accruals do marginally improve the prediction of earnings, but according to Lev et al., the improvement is economically insignificant. Lev et al. claim that the poor predictive ability is due to accrual estimates of low quality, specifically: “…the objective difficulties of generating reliable estimates and projections in a volatile economy, and their frequent misuse by managers appear to offset the positive role of estimates in conveying forward looking information to investors” (B. Lev et al., 2005, p. 1). Dechow and Ge (2006) report that accruals improve the persistence of earnings relative to cash flows in high accrual firms, but reduce earnings persistence in low accrual firms.
Traditional measures of accruals are functions of current- and non-current-period transactions. Francis and Smith (2005) show that the inclusion of non-current-period transactions leads to a downward bias on the persistence of accruals. For instance, deferred expenses increase current-period accruals and decrease nextperiod income, whereas deferred revenues decrease current-period accruals and increase next-period income.
current earnings are significantly related to future values of themselves; i.e., both metrics appear to be auto correlated. There is clear evidence that the accrual component of earnings is relevant for earnings predictions. However, with respect to accruals’ predictive ability for future cash flows, the empirical evidence is mixed. For instance, while Finger (1994) and Lev et al. (2005) report that current earnings are not more highly associated with future cash flows than is current cash flow, Dechow et al. (1998) and Barth et al. (2001) reach the opposite conclusion. Several studies suggest that accrual components of earnings exhibit lower persistence than the cash flow components of earnings (Sloan, 1996; Collins and Hribar, 2000; Lev et al., 2005).
The studies that focus strictly on time-series properties of earnings, cash flows and accruals rarely look into the value relevance of the measures that they analyse. However, since Ball and Brown’s seminal article from 1968, numerous studies on the relationship between accounting earnings and stock returns have been performed. Over the last decades, several researchers have also disaggregated accounting earnings and measured the value relevance of earnings components. Rayburn (1986) analyses operating cash flow and accruals’ association with security returns. She finds that both cash flow and accruals have a significant association with abnormal returns. The conclusion holds both for aggregate accruals and for most accrual items when split into major components. Dechow (1994) concludes that there is a stronger contemporaneous association between stock returns and earnings than between stock returns and realised cash flows. The association of stock returns with cash flow does, however, improve relative to the association of stock returns with earnings as the measurement interval is increased. The association between stock returns and earnings is relatively high when there
Subramanyan and Venkatchalam (2007) report that the value relevance of accrual-based earnings also dominates the relevance of operating cash flow when ex post intrinsic values of equity are considered. Francis et al. (2003) document that the value relevance of earnings exceeds that of cash flow from operations and EBITDA6. Liu et al. (2007) analyse valuation multiples (on an industry level) and find that earnings dominates both cash flows from operations and dividends with regard to valuation performance. In a variance decomposition analysis, Callen and Segal (2004) present evidence that accrual earnings news is a more important factor than cash flow earnings news in driving current stock return.
Consistent with the finding that accrual components of earnings exhibit lower persistence than the cash flow components of earnings (Sloan, 1996; Collins and Hribar, 2000; Lev et al., 2005), Wilson (1986) states that for a given amount of earnings, the stock market reacts more favourably the larger the cash flow component. Bernard and Stober (1989) on the other hand, find no evidence of this when performing a study similar to that of Wilson. Equivalently, Sloan (1996) reports that even though the cash flow component of earnings is more persistent than the accrual component, stock prices do not reflect this difference. Investors “fixate” on earnings and do not make use of information contained in the accrual and cash flow components of current earnings until that information impacts future earnings. Lev and Nissim (2006) refer to this phenomenon as the “accrual anomaly”. They show that the accrual anomaly persists, and that it has not declined over time7.
Francis et al. (2003) also show that earnings dominate their ex-ante preferred non-GAAP performance metrics, for instance, revenue per passenger mile (airline industry), value of new orders (homebuilding industry), and same-store sales (retail restaurants).
One may ask why sophisticated investors do not arbitrage away the anomaly. Lev and Nissim offer the following explanation (Baruch Lev & Nissim, 2006, p. 193): “By and large, institutions shy away from extremeaccruals firms because their attributes, such as small size, low profitability, and high risk stand in stark contrast to those preferred by most institutions. Individual investors are also by and large, unable to profit from trading on accruals information due to the high information and transaction costs associated with implementing a consistently profitable accruals strategy. Consequently, the accruals anomaly persists and will probably endure.”
relative information content. They analyse net income, cash flow and net sales and find that, in pairwise comparisons, each measure provides incremental information content beyond each of the other. In pairwise comparisons of relative information content, net income provides significantly greater information content than net sales and cash flow, and net sales provide significantly greater information content than cash flow. However, there is evidence that sales revenue outperforms earnings for high-tech “New Economy” stocks (Davis, 2002; Lianzan & Cai, 2005). Bradshaw and Sloan (2002) suggest that the market response to so-called “Street” earnings8 (modified definitions of GAAP earnings, or pro-forma earnings) has displaced GAAP earnings as a primary determinant of stock prices.
Briefly summarised, prior research suggests that both cash flow and earnings are value relevant. However, there is evidence that earnings are more value relevant than cash flow (Biddle et al, 1995; Dechow, 1994; Francis et al, 2003; Subramanyan and Venkatchalam, 2007). When earnings are split into cash flow and accruals, both components appear to be equally well associated with stock return (Bernard & Stober, 1989, Sloan, 1996). This conclusion is not consistent with the finding that the accrual component of earnings is less persistent than the cash flow components (the accrual anomaly).
Several studies analyse either the predictive ability or the value relevance of accounting variables. Kim and Kross (2005) and Kim et al. (2007) combine these two lines of research.
Kim and Kross (2005) investigate how earnings’ ability to forecast future cash flows has Cornell and Landsman (2003) report that none of the pro-forma earnings measures released by companies are specifically defined. However, the pro-forma earnings/Street earnings typically exclude special items and noncash items. Landsman et al. (2007) present evidence that these exclusions are actually value relevant, but that they are mispriced by the market.
has been decreasing over time (Collins, Maydew, & Weiss, 1997; Francis & Schipper, 1999;
Gu, 2007; Baruch Lev & Zarowin, 1999), they expect that earnings’ ability to predict future cash flow has decreased as well. Their expectation is founded on the fact that stock prices are the present value of future cash flow. Kim and Kross are surprised to find that the ability of earnings to forecast future (operating) cash flow has actually been increasing over the last decades: “If stock price is the present value of future cash flows, the deterioration in the association between accounting earnings and stock prices implies a growing inability of accounting numbers to forecast future cash flows, but that is not what we find” (Kim & Kross, 2005, p. 754). Although they claim that their finding is consistent with market inefficiency, they nevertheless conclude that they are unable to reconcile the increasing ability of current earnings to predict future cash flows with the decreasing ability of current earnings and cash flows to explain prices. This avenue is pursued by Kim et al. (2007). Kim et al. show theoretically that cash flow prediction regressions are contaminated by the presence of noise in the cash flows and value-unrelated (spurious) correlation between one-year-ahead cash flows and current earnings9. They find empirical evidence that both factors contributed to the findings of Kim and Kross (2005).
Note that the research quoted above is conducted on U.S. data. Relatively few studies analyse the predictive ability and value relevance of Norwegian accounting data. King and Langli (1998) find that book value of equity per share and earnings per share are significantly related to stock prices when analysing Norwegian accounting data for the period 1982-1996.
Earnings per share do, however, have a low incremental explanatory power when book value per share is already included in the regressions. In an analysis of value relevance over time, Specifically, one-year-ahead cash flows may be a very noisy proxy for all prospective cash flows because they contain significant value irrelevant noise, which is correlated with current earnings (Kim et al., 2007).
on the Oslo Stock Exchange has increased significantly over the past four decades. Gjerde et al. confirm King and Langli’s results that book value per share and earnings per share are significant explanatory variables for stock prices. They also test the explanatory ability of earnings and change in earnings for stock returns. Both earnings and the first difference of earnings are significantly associated with stock returns. Similar results are reported by Hope (1999)10. Gjerde et al. (2007a) report that the value relevance of Norwegian accounting numbers was little influenced by the introduction of IFRS for consolidated statements of quoted companies in 2005.
As for value relevance research in the other Nordic countries, Hellstrøm (2006) and Hassel et al. (2005) find that both book value of equity and accounting earnings are value relevant for their Swedish samples. However, Hellstrøm reports that earnings changes are generally not related to stock return. In their study of the Finnish stock market, Juntilla et al. (2005) present evidence that both earnings and the change in earnings are significantly related to stock return. In Denmark, Banghøj and Plenborg (2006) find that neither earnings nor the change in earnings are significantly related to stock return. None of these Nordic studies compare earnings with cash flow to assess the influence of accruals on value relevance.