«The relative value relevance of earnings and book value across industries. Mary Hilston Keener University of Tampa ABSTRACT Many studies evaluate the ...»
Journal of Finance and Accountancy
The relative value relevance of earnings and book value across
Mary Hilston Keener
University of Tampa
Many studies evaluate the impact of accounting information on price. This study, based
on Collins, Maydew, and Weiss (1997), examines the disparity in the impact of book value and
earnings on price over a twenty-year time period and for many industries. This paper extends
Collins et al. (1997) by looking at changes in value relevance over a more recent time period.
Also, this paper examines the differences in the value relevance of earnings and book values across industries. The results of this study suggest that the joint value relevance of earnings and book values has not decreased over the sample period. This study also demonstrates that the incremental value relevance of earnings (book value) has increased (stayed constant) for the sample period. Finally, this study demonstrates that there is no significant variation in the incremental value relevance of earnings and book values across industries.
Keywords: value relevance, earnings, book value, industries, time The relative value relevance, Page 1 Journal of Finance and Accountancy
Several important findings should be noted. First, this study confirms the Collins et al.
(1997) finding that the joint value relevance of earnings and book values has not decreased over the twenty year period examined. Second, the results demonstrate that, over the sample period, the incremental value relevance of earnings has been enhanced and the incremental value relevance of book value has not changed significantly. Third, the results suggest that there is not a significant difference in the incremental value relevance of earnings and book values or in the joint explanatory power of earnings and book values across industries.
LITERATURE REVIEWChanges in Value Relevance over Time As previously mentioned, this study will extend the Collins et al. (1997) paper related to the variation in the value relevance of earnings and book value over an extended period of time.
The Collins et al. (1997) study has three primary findings. First, the authors determine that “the combined value-relevance of earnings and book values has not declined over the past 40 years and, in fact, appears to have increased slightly” (p. 41). Second, Collins et al. (1997) find that the incremental value relevance of earnings has diminished and been replaced by an increase in the value relevance of book values over a forty year window. Third, Collins et al. suggest that “much of the shift in value-relevance from earnings to book values can be explained by the increasing significance of one-time items, the increased frequency of negative earnings, and changes in average firm size and intangible intensity across time” (p. 65).
Although the Collins et al. (1997) study is the foundation for this paper, many other studies have examined important issues related to changes in the value relevance of earnings and book value. Holthausen and Watts (2001) and Negakis (2005) identify and summarize some of the issues addressed in the extant value relevance literature. Lev and Zarowin (1999) find a decrease in the value relevance of earnings over the period from 1978-1996. The authors of several other studies have established that negative earnings and special items have had a negative impact on the value relevance of earnings over time (Hayn, 1995; Elliot and Hanna, 1996; Basu, 1997; Kang, 2003). These studies have also determined that companies have become more likely to report losses over time, which may further reduce the value relevance of earnings.
Although some studies have documented a decline in the value relevance of earnings over time, other studies including Barth, Beaver, and Landsman (1998) have found that book values are more value relevant than earnings when losses are present or when earnings include The relative value relevance, Page 2 Journal of Finance and Accountancy special items. This may be because book value serves as a surrogate for a firm’s abandonment value or because book values better predict future earnings if current earnings have many shortterm components. Burgstahler and Dichev (1997) determine that “equity value is a convex function of both earnings and book value, where the function depends on the relative values of earnings and book value” (p. 187).
Chandra and Ro (2008) find that the combined value relevance of earnings and revenues has stayed constant and that the value relevance of earnings has declined while the impact on price of revenues has not decreased. Jenkins, Kane, and Velury (2009) find that the value relevance of earnings is higher during economic contractions if an estimate for future earnings expectations is included in the model, and they show that the value relevance of expected future earnings is greater during expansions. In summary, these studies have primarily found that earnings and book values move in opposite directions.
Reasons for Documented Changes in Value Relevance over Time
Several findings documented in prior studies may have caused the value relevance of both earnings and book values to change over time. First, the increased number of firms in technological industries over time may have affected the value relevance of earnings and book values because of the importance of intangibles to these firms. Both Amir and Lev (1996) and Lev and Zarowin (1999) establish that financial accounting information is less important to investors if they are focusing on firms from service and technological companies because accounting standards require the immediate expense of accounting intangibles in many cases.
Second, the large number of special items that companies are reporting may impact the value relevance of earnings and book values over time. Elliot and Hanna (1996) indicate that the market does not put as much credence in special items as in earnings before special items and that there has been an increase in the number of special items that companies have reported over time. Also, Ohlson (1995) indicates that the decrease in the persistence of earnings connected with the increase in the number of special items may cause less weight to be placed on earnings than on book values.
Third, the number of losses reported by companies has increased over time, and this increase is expected to impact the value relevance of earnings and book values. Basu (1997) examines the function of conservatism in accounting and suggests that firms incorporate bad news more quickly into earnings than good news, which implies that losses are more short-lived than increases in earnings. The increase in the number of losses over time may also reduce the ability of earnings to predict returns. Hayn (1995) finds that firms with losses have smaller ERCs than firms that report positive earnings and confirms that firms have indeed reported more losses over time. Taken together, the findings of Hayn (1995) and Basu (1997) indicate that the increased number of losses over time may be one cause for the decline in the incremental value relevance of earnings over time.
Fourth, the increased number of small firms on the COMPUSTAT database over time may be one source for the incremental importance of book value over earnings in explaining market values. Finally, Dontoh, Radhakrishnan, and Ronen (2004) suggest that the decline in the value relevance of accounting information over time has been “driven by an increase in noninformation-based trading” (p. 30).
Industry Differences in Value Relevance A review of the extant value relevance literature indicates a gap in the research related to a general examination of industry-specific effects on the value relevance of earnings and book values. The study will attempt to fill in this gap. Many prior value relevance studies have examined specific industries, but most prior research does not examine a broad range of different industry classifications. For example, many studies examine the value relevance of various financial and nonfinancial performance measures in specific high-tech industry sectors (Amir and Lev, 1996; Hirschey et al., 2001; Aaker and Jacobson, 2001; Graham et al., 2002; Al-Harbi, 2003; Xu, 2003; Liang and Yao, 2005; Junttila et al., 2005; Tan and Lim, 2007; ).
Barth et al. (1998) examine the value relevance of earnings and book values across three different industry classifications chosen based on how likely unrecognized intangible assets are in these industries. The authors determine that for pharmaceutical firms, the value relevance of earnings is greater than that of book value and that for financial service firms, the impact on price of earnings is significantly lower than that of book value. The authors also find that the incremental value relevance of earnings and book value are equivalent for firms in manufacturing industries.
Hughes (2000) examines the electric utility industry and determines that industry-specific nonfinancial information including measures of air pollution is value relevant. Boone (2002) determines that oil and gas asset present values are more value relevant than oil and gas assets measured at historical cost. Riley, Pearson, and Trompeter (2003) examine the value relevance of nonfinancial performance measures and traditional accounting information for the airline industry. Stunda and Typpo (2004) and Kang and Zhao (2010) examine the real estate industry to determine the value relevance of several industry-specific financial measures. Although the papers discussed in this section examine the value relevance of many financial and nonfinancial items for specific industries, none of these studies examine the relative value relevance of earnings and book value across different industries over time; therefore, this paper will address these issues.
Developing Value Relevance Topics
Other important value relevance paradigms have also been examined over the last decade. Many studies examine the value relevance of various financial and nonfinancial measures in countries around the world (Alsalman, 2003; Martinez, 2003; Habib, 2004, Junttila et al., 2005; Goodwin & Ahmed, 2006; Wulf, 2007; Ibrahim et al., 2009, Bo, 2009). Marquardt and Wiedman (2004), Habib (2004) and Lapointe-Antunes et al. (2006) examine the effects of earnings management on the value relevance of financial performance measures. Callen, Livnat, and Segal (2006) and Caylor, Lopez, and Rees (2007) examine issues related to whether the value relevance of earnings is dependent on the timing with which earnings information is released. Ou and Sepe (2002) and Tan and Lim (2007) determine how analyst forecasts impact value relevance. As the economy becomes more global and high-tech over time, the number of issues for future research related to value relevance have increased.
METHODOLOGYThe initial regression model for this study demonstrates that price can be modeled as a
function of earnings and book value, as in Ohlson (1995) and Collins et al. (1997):
Pit = α0 + α1 Eit + α2 BVit + εit, (1) where P is the price per share, E is earnings per share, BV is book value per share, and ε is other value relevant information. The explanatory power of earnings and book value can be
disaggregated by breaking the total explanatory power into two parts as follows:
Pit = β0 + β1 Eit + εit and (2) Pit = γ0 + γ1 BVit + εit. (3) Next, as in Collins et al. (1997), this paper attempts to determine whether the value relevance of earnings and book values has changed over time by regressing the R-squared values from
equations (1), (2), and (3) on a time dummy variable over time as follows:
Rt2 = φ0 + φ1 TIME + εit, (4) where TIME = 1, …, 20, which corresponds to years 1982-2001. The incremental explanatory power is said to have declined if φ1 is significantly negative.
Because the purpose of this paper is to examine whether the changes in the incremental explanatory power of earnings and book value is the same across industries, an industry dummy variable will be substituted into equation (4) as follows:
Rt2 = φ0 + φ1 SIC + εit, (5) where SIC = 1, …, 10 where 1 represents the agriculture, forestry, and fishing industry (codes 01-09), 2 represents the mining industry (codes 10-14), 3 represents the construction industry (codes 15-17), 4 represents the manufacturing industry (codes 20-39), 5 represents transportation and public utilities (codes 40-49), 6 represents whole trade firms (codes 50 and 51), 7 represents retail trade firms (codes 52-59), 8 represents the finance, insurance, and real estate industries (codes 60-69), 9 represents the service industry (codes 70-89), and 10 represents nonclassifiable establishments (code 99).