Value Relevance of Conditional Conservatism and the Role of Disclosure : Empirical Evidence from Iran

This study examines the effect of conditional conservatism on value relevance of accounting variables. This study shows that conservatism decreases value relevance of earnings. More importantly, the evidence, indicate that the strength of the relation is conditional upon the firm’s disclosure, being the strongest for firms with high disclosure level and the weakest for firms with low disclosure level. In order to measure the conditional conservatism, C-Score from Khan and Watts have been used. Also measurement of disclosure quality has been done by TSE corporate disclosure scores. Based on a sample of Iran listed entities, between 2004 and 2011, this study posits conservative procedure mapping into reduced value relevance and this manner being the strongest for the firms with high conservatism and low disclosure.


Introduction
Conservatism has had a historic role in choosing and applying principles and accounting procedures.Many researchers believed that financial reporting goes to conservatism and speed of this movement has fluctuated in response to financial condition.With this trend, many studies are considered conservative bias in reporting and influence on value relevance.The extant literature on value relevance has consistently found that the value relevance of earnings has been decreasing over time (Francis & Schipper, 1999;Lev & Zarowin, 1999).Appealing to the empirical literatures, there are different views about association between conservatism and value relevance.Many studies have posited conservative as a reason for the decrease in value relevance.This study stated that conservatism has inverse relation with earnings value relevance and believed that increasing accounting conservatism is one reason for the decline in the value relevance of accounting earnings over time (Lev & Zarowin, 1999).In opposed approach, have stressed the role of conservatism in accounting for the purposes of valuation and contracting and believed that conservatism provides more reliable contracting and valuation by improving reliability and verifiability of information.This suggests that increasing conservatism should enhance rather than value relevance (Watts, 2003).According to signaling role of conservatism and disclosure, there are two information policies that chosen by management to present relevant information to market.Conservative is one of this strategies that defined by accounting practices.Furthermore, investigating the effect of disclosures on the relation between accounting conservatism and the value relevance of accounting information is essential.The study of this interplay is important because voluntary disclosures and accounting conservatism can provide value relevant information that has not yet been captured.

Conceptual Framework and Research Background
In examining value relevance of conservative and non conservative accounting and their relation whether disclosure policies and conservatism have an association, different level of disclosure and information environment can be the main reason of this relation.Importantly, strength of this relation is conditional upon the firm's information environment and being the strongest for firms with low information asymmetry and the weakest for firms with high information asymmetry.Underlying this prediction is the supposition that conservatism and disclosure represent jointly-determined strategies which form a part of the overall reporting strategy of the firm (Artiach & Clarkson, 2011).In developing our theoretical perspective, we draw on insights following from the models of Gietzmann and Trombetta (2003) and Bagnoli and Watts (2005), which offer a signaling role for conservative accounting policy choice.Gietzmann and Trombetta (2003) envisage a role for both the adoption of conservative accounting policies and voluntary disclosure within their model.They concluded that the economic implications of either accounting policy choice or voluntary disclosure cannot be meaningfully studied in isolation but rather each has the potential to influence and additionally, that an investigation of the economic benefits of accounting policy choice must control for the disclosure strategy of firms.

Conservatism and Value
Relevance in Different Disclosure Level Basu (1997) defined conservatism as the tendency to require a higher verification standard to recognize good news as gains than to recognize bad news as losses.There are two types of conservatism in accounting literature.Conditional, or ex post, conservatism refers to a system of reporting whereby the understatement of accounting value occurs in response to economic events.Unconditional, or ex ante conservatism, refers to a system of reporting whereby the understatement of accounting value occurs independent of economic events (e.g., Beaver & Ryan, 2000).Value relevance is the degree to which accounting earnings summarize information impounded in market prices (Mary, Barth, William, Beaver & Landsman, 2005).Relevant Literature shows that direction of relation between value relevance and conservatism is not clear.It is generally agreed upon that the greater the value relevance of earnings, the more useful it is for market participants when making investment decisions (Mary et al., 1996).Many researchers argue that there is a tradeoff between conservatism and value relevance (Balachandran & Mohanram, 2006;Brown, He, & Teitel, 2006).Givoly and Hayn (2000) found that the level of earnings conservatism has increased from 1956 to1998.Moreover, Lev and Zarowin (1999) have recorded a decrease in the value relevance of reported income over the same period (1977)(1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996), indicating that more conservatism is associated with less value relevance.Presuming a tradeoff exists, researcher has argued that increasing conservatism is a possible reason for decreasing value relevance (Lev & Zarowin, 1999).
In contrast, Watts (2003) argues that conservatism was a reaction for many years despite criticism.He believes that while the demand for conservative accounting is often attributed to contracting, conservatism may also lead to more relevant information.Balachndran and Mohanram (2004) stated that one reason for conservative accounting is that less conservative accounting approaches have the potential for errors in measurement and correspondingly have reduced reliability.Since conservatism favors more objective measures and less estimation, conservative accounting policies may lead to accounting numbers that are likely to be more reliable.If information becomes sufficiently unreliable, the usefulness of that information for decisions (i.e.relevance) could be limited and hence relevancy could actually increase with conservatism.They find that the value relevance of accounting has declined only for firms with the least conservative accounting and changed insignificantly for firms with the most conservative accounting.Further tests indicate that the decline in value relevance cannot be attributed to alternate definitions of conservatism based on the asymmetric timeliness of earnings.Recently, Balachandran and Mohanram (2010) examined the association between conservatism and the value relevance of accounting information over the 1975 through 2004 period.They find no evidence that firms with increasing conservatism exhibit greater declines in value relevance.Based on these results, it is implausible that increasing conservatism drives the decline in value relevance.
If only studies posited a relation between conservatism and value relevance, but also there is no consensus about the direction and strength of this relation.If there is an association between conditional conservatism and increasing level of disclosure, such differences can be justified.The use of conditional conservatism and increasing disclosure level, as two relevance signaling devices, can improve relevant information of accounting figures.In order to take advantage of reduced information asymmetry and internal information about the company, management uses several information strategies.Managers are aware of the information about the company's future earnings and she can choose conservative and / or disclosure as a means of signaling policy.Voluntary disclosure involved in reducing the risk of any conservative accounting system, the benefits of voluntary disclosure will not increase (Gigler & Hummer, 2001).Healy and Palepu (2001) examined Manager's motivation to expand voluntary disclosures and the effects of disclosures on equity market value.However, they did not address the question of whether the market responds to disclosures in different accounting regime.Investigating the effect of accounting conservatism on the relation between disclosure level and the value relevance of accounting information is worthy.More directly, our reasoning is base on Gietzmann and Trombetta (2003) that investigate trade-off between voluntary disclosure and signaling through the firm's choice of conservative accounting policy to disclosure private information about firm value.They concluded that the value relevance of either accounting policy adoption or voluntary disclosure cannot be meaningfully studied separately.Any study of the value relevance of accounting needs to control for the voluntary disclosure strategy of the firm.Therefore any study of the link between value relevance and conservatism should control for differences in disclosure policy.Unlike other study, Gietzmann and Trombetta (2003), presented a model that conservatism and voluntary disclosure are used as an effective mechanism for transfer of confidential data.They argued that the selection of any of these policies will impose costs (contractors and agency costs) affecting companies, the accounting policy is conservatively advantage signaling.If the cost is large enough, only companies with profitable economic future, select conservative accounting policies.In fact, they found that companies select conservative accounting policy under cost effective manner and also contain a message about profitable future.In contrast, it is possible that some companies realize conservative accounting policies with high costs and through the voluntary disclosure should inform their future (Richardson, Sloan, Soliman & Tuna, 2005).
In another approach, conditional conservatism strength the financial performance of contracts and it is a tool to increase the quality of accounting information (Ball & Shivakumar, 2005).Firms with high disclosure quality are expected to exposure more reliable information about losses and other adverse events.Information with low credibility has ability to manipulate and has less utility.In other words, conservatism reduces information asymmetry and can be effectively managed by selective disclosure policy.(Lafond & Watts, 2008).Cano-Rodriguez (2010) argued that unconditional conservatism stems from the pressures of tax, legal and may reduce the value of the company's reported financial information.It is expected that firms with lower disclosure quality, exercise greater unconditional conservatism.Management Reputation, stock price and future financial statements are affected by financial statements or contracts based on accounting items (Lafond & Watts, 2008).Therefore, companies have different motives for conditional and unconditional conservatism.As a result, given management incentives and constraints on earnings management, unconditional conservatism can have different effects on the quality of reporting.Management may have an opportunistic incentive to overstate assets and gain against of understating value of the debt and costs and therefore increases their wealth or shows better financial conditions regarding company.Watts (2003) stated that Conservative is known as a management tool for preventing opportunistic behavior.Several studies argue that voluntary disclosure affect firm value based on signaling theory.Voluntary disclosures in the annual report send signals to the market and these signals are expected to increase a firm's net present value and stock market value (Gordon, Loeb & Sohail, 2010).Timely financial reporting can be a reliable source to present relevance information and explaining management's strategies and internal information and reduce information asymmetry between management and investors.This disclosure has many consequences in form of decrease: information asymmetry, uncertainty regarding future cash flow, loss of major clients and increase in creditors' interests due to high default risk (Wruck, 1990;Whitaker, 1999).

Hypothesis Development
Due to differences in corporate disclosure level and financial reporting among firms, this research aims to investigate the role of disclosure level in value relevance of conservatism field.The study of disclosure is important from information policy perspective.Expanded disclosure can be important in making changes in information asymmetry and therefore affect on choosing other information policy such as conservative procedure.Since managers have discretion in how businesses deal with the event and have a choice of accounting procedures, conservatism and disclosure are mechanisms that enable managers to communicate their personal information to investors.Purpose of this study is answer to this problem how managers use accounting conservatism or/and disclosure as a factor to decrease information asymmetry and what the role of increasing disclosure level in variation of conservative accounting relevancy.Given the theoretical framework and research background, the following hypotheses are proposed: H 1 : conditional conservatism has a significant effect on accounting disclosure.
H 2: value relevance of high and low conservative firms varies and this difference, increases after considering different level of disclosure level.

Research Design
The first step of this study is to examine the relation between accounting disclosure and conditional conservatism and explore impact of conservatism on disclosure level.After determining how conservative affect on disclosure levels, the sample firms are divided based on this interchange.In the other words, to examine the H 2 and explore variation in value relevance of conservative earning after controlling for disclosure level, at the second step, we investigate the variation in value relevance of earnings in different levels of conservatism.Thus, companies are divided into two groups with high and low conservatism firms.For this classification, sample firms sort by conservatism score and three top and bottom deciles are selected.In the third step, cause of relation between disclosure and conservatism, high and low conservative firms are examined in environments with different disclosure level (i.e.high/low conservatism and low/high disclosure).Also for this classification, sample firms sort by conservatism/disclosure interchange and three top and bottom deciles were selected and value relevance of this two groups are examined.Therefore, the value relevance of high and low conservative accounting variables are calculated after controlling for changing in level of disclosure levels that was provided by companies.In this study, the sample collection starts with the entire population all companies from TSE during the 2004 to 2010.After deleting firm observation with missing data and negative book value, there are 903 firmyear observations available from the TSE Database.
Model1 was presented to examine H1 which is the effect of conditional conservatism on disclosure levels.This model included some controlling variables that influence on disclosure levels.This controlling variables are return on assets (ROA) as a measure of company performance, market value to book value (MVBV) as a measure of firm growth, size (SIZE) and leverage (LEV) as follows ( Lang & Lundholm, 1993, 1996): In this model: DisQ it : Score of corporate accounting disclosure.
CONS it : Measure of conditional conservatism (C-SCORE).
ROA i,t : Earnings before extraordinary items Adjusted by total assets.
MVBV i,t : The market value of corporate equity divided by book value of equity.
LEV i,t : Total debt divided by total equity market value of the company.SIZE i,t : Natural logarithm of total market value at the end of the year.
e i,t : Model error.

Conditional Conservatism Measurement
Khan and Watts ( 2009) developed a measure of conditional conservatism by extending the model of Basu (1997).
They introduced a measure of conservatism with cross Sectional regression coefficients were extracted from Basu's model, as a function of the specific characteristics of the company and replaced these coefficients in the Basu's regression.In result, they able to resolve some of the weak points of the Busu's model and provided a firm specific measure of conditional conservatism.Their annual cross-sectional model was presented as follow: (2) In this model, X equal to earnings, R equal to return (measuring news), SIZE is Market value, M / B is Company's market value to book value, LEV is Market value of debt and D is a dummy variable equal to 1 when R<0 and equal to 0 otherwise, and e is the residual.In order to calculate the coefficients, the first model was calculated for each year and then extracts the coefficients of this model.Conservatism score (c-score) measure incremental timeliness for bad news over good news or conservatism.(Khan & Watss, 2009)

Disclosure Quality Measurement
Two basic approaches for measuring disclosure quality have evolved within the academic literature.The first involves direct disclosure scoring by the researcher based on a self-developed disclosure index (e.g.Wiseman, 1982;Botosan, 1997).The second involves the use of external rating of disclosure such as those published in the Association for Investment Management and Research (AIMR) reports.Examples of studies using the AIMR ratings include Lang andLundholm (1993, 1996) and Botosan and Plumlee (2002).Tehran stock exchange (TSE) publishes disclosure score like those published in AIMR.Base on reliability and timeliness of disclosed information, TSE issues disclosure quality score (DISQ) of listed firms.This score are published annually and summarized as a score and are assigned to each firm.Advantages of these criteria included: application of objective indicators to measure the timeliness and reliability criteria relied to prevent strains from personal judgments, subjective interpretations and researcher able to compare the verifiability in the future study.This study, similar to Karami, Hajiazimi and Attaran, ( 2012) is based on the second approach.Subjective judgment by researcher's disclosure index could be named as a disadvantage of the first approach.As such, the index may not fully reflect the views of relevant user groups such as investors, creditors or financial analysts (Artiach & Clarkson, 2011).In DISQ, one of the fundamental attributes of relevance is timeliness, an influencing factor on firm's score to wit forecast of return on stock, quarterly financial statements, portfolio condition, dividend payout schedule and etc. reliability points to standard deviation of financial forecasts and the difference between actual results and budget statements sent to TSE.The Public releasing, availability of disclosure reports for all kind of investors and using objective factors to measure this score are main incentives for doing this study in Iran.

Value Relevance Measurement
Typically, value relevance measured as the adjusted R 2 of regressions with stock price as the dependent variable and book value and earnings as independent variables (Francis & Schipper, 1999;Lev & Zarowin, 1999).Alternatively, value relevance has also been measured as the adjusted R 2 of regressions with returns as the dependent variable and level of earnings and change in earnings as independent variable.Finally, value relevance has also been measured as the stock returns that could be earned from perfect foresight of accounting information (Francis & Schipper, 1999).To investigate H2, similar to Easton and Harris (1991), we use the second approach and examine adjusted R 2 of regression with annual stock return on scaled earnings and changes in earnings whit fallowing regression: Where RET is annual stock return, EARN is annual earnings before extraordinary items divided beginning market value of equity, ΔEARN is change in EARN (annual earnings before extraordinary items divided beginning market value of equity), i and t are firm and year, respectively.In order to investigate the changes in the value relevance of earnings conservatism in different levels of disclosure and information environment, companies were divided into two groups with high and low conservatism firms.For this classification, sample firms sort by conservatism score and three top and bottom deciles were selected.After that value relevance (adjusted R 2 ) of accounting information, is measured.Prior studies provided intuitive R 2 comparisons, the lack of test statistics places limitations on assessing the strength of the findings.Therefore, we use the Cramer Z-test to test for the difference in the adjusted R square (e.g.Francis & Schipper, 1999).

Relation between Conditional Conservatism and Disclosure Score
Table 2 shows the relation between conditional conservatism and disclosure score (models 3).This table shows the a 2 coefficient from CONS, is significant at 99% confidence.These results show, as expected, conditional conservatism (measuring by C-Score), has a negative influence on disclosure score.On the other words, conservative accounting and disclosure are substitutive information policies.Meanwhile, Conservative practices in financial statements strains disclosure quality and management are less willing to provide high disclosure level.The key results emerge from Model 3, consistent with prior research (Gigler & Hummer, 2001;Artiach & Clarkson, 2010), is substitutability of conservatism and disclosure.According to this relation, sample firms were divided to high conservative/low disclosure and low conservative/high disclosure firms and then value relevance of each group was examined.

Value Relevance Test
Table 3 presents results from value relevance tests.In this study value relevance test is base on Easton and Harris (1991) framework, in which returns is a function of both earnings levels and earnings changes.Also, table 3 presents the slope coefficients, related t-statistics in parentheses and the Cramer Z-test for adjusted R 2 comparing between HCONS with LCONS groups and HCONS/LDISQ with LCONS/HDISQ groups.According to pooled GLS results, the F-tests are highly significant in all models.The coefficient of two independent variables (EARN, ΔEARN) have different significant levels in the HCONS, LCONS and LCONS/HDISQ groups.Meanwhile, EARN is significant only in HCONS/LDISQ firms.The results of regressions and Cramer Z-test show that adjusted R 2 significantly (p<0/05) higher in LCONS firms (24%) than HCONS firms (14%).This means that the accounting information is more associated with the prices in firms with low conservatism procedure.In addition, the explanatory power manifested in adjusted R 2 in firms with conservative accounting procedure and weak disclosure (11%) significantly (p<0/01) lower than low conservative and high disclosure firms (28%).These suggests that conservative procedure followed by low disclosure quality, decrease explanatory power of stock returns by accounting figures reflected in the financial statements.

Conclusion
Given the economic consequences of reducing information asymmetry in the capital market, in form of selection and presentation of accounting data, this study examines the effect of conditional conservative accounting on value relevance and the role of disclosure levels.The sample used consists of a sample of Iranian firms listed on the TSE over the period 2004-2011.The models of Khan and Watts (2009) and Easton and Harris (1991) are used to assess the level of conservatism in financial reporting and the value relevance of earnings respectively.
Meanwhile TSE disclosure quality score is used for measuring the disclosure levels of firms.After exploring effect of conservatism on disclosure level, sample firms base on their level of Conservatism and disclosure score are classified to three portfolios.The results of GLS regression in the first model shows a significant negative relation between conditional conservatism and disclosure.In other words, the conservative strains in the financial statements, reducing the incentives to provide high level of disclosure to the market participant.These findings suggest that conservative accounting systems reduce benefits of providing timely information (in the form of risk sharing) and thus reduce the tendency of managers to provide timely information.Importantly, as predicted, value relevance of accounting conservatism in should examine by the influence of conservative accounting on disclosure levels.Based on the conceptual framework and prior literature, this study states that the strength of this relation is conditional upon the firm's information environment, and disclosure level.In the first step, we explore that high conservative procedure mapping into reduced value relevance of accounting variables.However, in the second step, this study posits that this manner being the strongest for firms with high conservatism and low disclosure (high information asymmetry).These findings posit that conservative procedures cause to decline in disclosing high level of information (publishing timely and reliable information) which that decrease value relevance of accounting variables.There are some limitations of the study.The sample includes only Tehran listed firms.Therefore, the findings might not be valid for other firms.Meanwhile, since using TSE disclosure score as a measurement of disclosure level revealed by companies, this survey has some constraints in practice.Not only disclosure level can be measured by different approach but also using other methods can be obtained different results in other countries.

Table 1 ,
shows the descriptive statistics of study's variables.In this table, Panel 1-A, shows full sample statistics.As previously mentioned, full sample was divided to four groups: high conservative (HCONSE), low conservative (LCONSE), high conservative and low disclosure (HCONS/LDISQ), low conservative and high disclosure (LCONSE/ HDISQ).Table1-B presents the descriptive statistics and difference of means analysis for these groups.From table 1-A, mean disclosure score is 55 for full sample and 50 (59) for the HCOSE (LCONS) sample.Meanwhile, mean deflated earnings are 0.19 for the full sample and 0.18 (0.19) for the HCOSE (LCONS) sample.As table 1-B shows, stock returns, deflated earnings and disclosure score are significantly greater for the LCONS group than HCOSE group.Meanwhile, mean deflated earnings and stock returns are significantly greater for the LCONS group than LCONSE, HDISQ group.

Table 2 .
Relation between conservatism and disclosureDisQ it = a 1 + a 2 CONS i,t + a 3 ROA i,t + a 4 MVBV i,t + a 5 SIZE i,t + a 6 LEV i,t + e i , t