Accounting conservatism and earnings management constraints

The main purpose of this research is to review the relation between accounting conservatism andthe limit to earnings management in case of Tehran Stock Exchange (TSE). There are numerous studies about the noted relation, in countries which their market capital is the main source of firms financing. However, as long as rapid development of TSE, necessity of such studies become more obvious in Iran. So, by analyzing data gathered from 86 listed firms at TSE for years 1380 to 1391, the relation between accounting conservatism and earnings management restrictions was tested using panel data. Givoly and Hayn (2000) approach for measuring accounting conservatism and adjusted jones model for measuring accruals based earnings management were used. The result shows that there is significant relation between accounting conservatism and earnings management restrictions.


Introduction
Investors are the main suppliers of companies' sources, asking for complete and true information to make investment decisions. In efficient capital markets it's supposed that all the present data are absorbed by the individuals and the effects reflected in securities. The goal of financial statements is also presenting anddata Summarizationand classified data about financial status, financial performance and financial flexibility of the business, being beneficial for a vast variety of the users. The most important information is the earning and loss statement. Earning as the final result of accounting procedure which is being so much under attention and emphasis of the information users, is affected by the accounting procedures chosen by the management, has been calculated. This possibility to choose accounting procedures, gives this opportunity to the management to make decisions about costs and revenues recognition time and measurement.Management is motivated to use non-conservative procedures (aggressive) in order to show theearnings growth of the company (watts,2003). Despite the reported earning amount is important to the investors and is effective in their decisions, quality features of earning is also noticed by the investors , as one of the earning information (Francis et al, 2002).
Based on the theoretical concepts of financial reporting,beneficial information is the one having quality features. According to the theoretical concepts of financial reporting, the main quality features related to the information contentis the relatedness and reliability, andcaution or conservatism is one of the main features of reliability. In fact, conservatism is of the financial reporting features, as a limitingprinciple of accounting in the concepts and framework of accounting, it plays an important role in limiting optimistic behaviors of the managers as information producers, and estimating the minimum revenues of the investors and creditorsas the most important users from the other hand(Karami and Ahmad Abadi, 1387). In this study, we are to investigate the role of conservatism in limiting the earning management.

Literature Review
Most of the stock market activist's sources, are the financial reporting published by the companies cyclically being in public's hand and the decision basis for the real and potentialinvestors in order to buy, sell and invest on the stock market. By increasing the sourcesunder the authority of the management, the number of beneficiariesrelated to the company would also increase and the result of such conditions is the appearance of conflict of interests. Managers that are in the center of this conflict are to reduce it through presenting financial information. They are motivated to present company's condition and mostly due to themanagement authorities in presenting reports,they also have the opportunity to perform this procedure. Thus, the existence of controlling and monitoring mechanisms to keep the interests of the investors is necessary, accounting procedures and principles in keeping the interest activists of stock market and controlling the managers' opportunistic incentives,suggests conservatism's limiting principle. believe that conservatism is a disciplinary mechanism inhibiting profit-seeking incentives of the managers and decreasing their power to speed up the good news disclosure and delay in bad news disclosure. According to the above approach,it's expected that the more conservative one company's accounting procedure is,keeping its bad news hidden would decrease. Accordingly, it's predicting that the more conservative companies in comparison with other ones will face less crisis resulting from sudden publishing of bad news. Rayan (2006),states that the accrual accounting gives a significantright to the managers in determining earnings in different timeframes. In fact, under such accounting system, managers have a significant control on the recognition time of some costs as advert, research and developing costs.
Although too much monitoring may cause replacing,the management incentive is being kept. If the discretionary earning management of accounting reduces,there's the possibility for the real earning management to be increased (Chen, 2009). In two dominant forms,earning managementwhich has often suggested in previous researches, is manipulating Real Earning Management and Accrual Earning Management. Performing each of the options of earning management includes limitations and costs. At first a manager is to one option in order to access earning management, and if facing limitations he will choose another one or use them both simultaneously (Etemadi et al, 1389).
In order to access to the desired level of earnings, managers can wait to the end of the year and use the discretionary accruals for the management of reportedearnings, but this method may be risky in which the earnings being manipulated should be more than the present discretionary accruals, because the authority about discretionary accruals have been accepted through the public principles of limited accounting. We supposed that the main economic events of one company,makes the managers ability limited for the discretionary earning reporting.Thus,there's a possibility in which managers are unable to attain the end of the year desiredearning through discretionary accruals,they can reduce the risk via manipulating the real operational activities during the year. Managing the real earning is less in danger of limitation. Another advantage of replacing real activities to manipulate the earning is that the accountants and legislators pay less attention to such behaviors. Barton (2002) believes that real earning management is not something without any cost. For, there's the possibility in which cash flows will indirectly (negatively) effected by an activity in future that are being done to raise the earnings. Most of the present evidence round the management of real earningsis focusing on the decrease of discretionary costs' opportunistic, such as costs of research and development. Although reducing the discretionary costs can raise the earnings in meeting the needs of some particular goals,it causes the risk of decreasing the future cashoperational events. Because generally,this activity reduces output events of cash and instead,has a positive effect on the current operationalcash flow. In the following years,this direct effect may become particular. In the case of extra producing,although this method improves the Profitability margin, but the company will bear costs and reduces the normal level of company'soperationalcash flows. In the case of manipulating in sales, it can be said that the managers will prepare discount field close to the end of the financial year in order to increase the rate of sale.Conditional conservatism with in time recognition activitiesof losses and delay in recognizing the economic earnings will reduce opportunities for a successful earning management. Guay and Verrecchia (2006) believe that conditional conservatism (a vast view of watts J u l y 10, 2 0 1 4 conservatism)affects the opportunistic behavior of the managers to the financial statements.  also believe that conservatism can reduce the opportunistic behavior in accounting and company's incentives for earning management. Hence,the goal of current study is to investigate conservatism and limitations of earning management,on the other hand,we are to find out whether conservatism causes the earning management or not. The above mentioned question could be posed as if there is a negative relationship between accounting conservatism and earning management. We also have two general ways for earning management:earning management based on real activities and earning management with the use of discretionary accruals. In this research, the relation between conservatism and earning management based on accruals will be exclusively investigated.

Significance of the study
Management to access the expected level of the earning has three options:real earning management,deceptive accounting and accruals management. Deceptive accounting is related to the selection of accounting procedures inconsistent with accepted principles. Mojtahedzade (1389)believes that,earning management through manipulating discretionary accruals and the accounting procedures being usedis based on accepted principles of accounting. But it's to present the real economic performance vague. Realearning management happens when the managers do activities to deviate them from the best performance,so that they can report more earnings in two forms of earning managementthat have often presented in previous researches of Etemadi et al (1389) and Chen (2009)which is the manipulationof real activities and discretionary accruals.
Performing each of the earning management includes limitations andcosts. In fact, the manager firstly turns to one option to achieve earning management and if faced limitations, the direction would change or both of them will be used simultaneously. To confirm these studies,  believe in the theory in which conservatism could reduce opportunistic behavior in accounting, stimulating company's incentives for earning management. Generallyearning management has been described as a purposeful interfering factor in financial reporting process intending to get some personal interests. This definition focuses on the opportunistic side of the management; this means that management is on earning management with profit-seeking incentives and it results in the decrease of accounting numbers informational content. On the other hand, some researchers have an informing look at earning management and describe it as the manipulation of number of earnings by the managers through which the private and confidential information of management about future performance of the company transfers to the investors. Accordingly, earning management not only reduces earnings informational content but also helps the investors to better interpret the reported numbers of earning.Balachandran and Mohanram (2011) believe that conservatism in accounting is usually considered as a choice(by legislators,standard developers and companies) of accounting procedures resulting to lowering the value of the assets and earnings. It seems that the important case around conservatism in accounting is whether this accounting oath according to the results it has, is finally beneficial for different users or not. Hence, studying the relation between conservatism and different variables has become the topic for so many of the researches and by little thinking in the results we can derive two completely different viewpoints about conservatism.Supporters' viewpoint:ofthis view,conservatism has an informational role and increases the quality. Ithas so many fans who consider a lot of informational advantages for conservatism and the most important one is the earning management decrease and also decreasing the agent costs and information asymmetry. Thus, by using this view of conservatism in financial reporting,it can help the investors and other users of financial statements to make good decisions (Lafond and Watts, 2008;Lara et al, 2012). In contrast, the opponents believe that there are no informational advantages in this viewand reduce the quality of information. Presenting conservative financial statements could cause the wrong accommodation of revenues and costs that follow less stability of the revenues and causes aunreal earning (Ghaemi et al, 1389; Balachandran and Mohanram, 2011). Hence,according to the previously done studies and also the importance of earning management and on the other hand, lack of studies in this field inside the country,the present study is investigating conservatism and limitations of earning management.

Background Researches
Shourvarzi et al (1391) studied the relation between discretionary accruals and accounting conservatism in Iran's stock market. Results showed that there is a meaningful relation between discretionary accruals and accounting conservatism and other studies presenting lack of a meaningful relation between these two.
Mehrani, EbrahimiKordlar and Hallaj (1390) studied the relation between unexpected accruals and conservatism in 1380 to 1385 using Basu model. The results show that the unexpected accruals of modified Jones model has a positive and significant relation with conservatism, but there is no significant relation about the effects of unexpected accruals of modified Jones model and non-linear cash flows on conservatism.
Khodadadi et al (1389),in discretionary accruals and accounting conservatism, studied the management authorities on accruals in accounting conditional conservatism. In the previous researches we can see that the conditional accounting conservatism presented in earnings,is mainly because of the accrual part of the earnings and also classified the accruals to two discretionary and non-discretionary ones. And they concluded that the conditional accounting conservatism being presented in relation with accruals is mainly a reaction to the discretionary section of the discretionary accruals.
Khajavi, Vali pour and Askari (1389) studied the effect of conservatism on earning stability in the years 1377 to 1386. Evidences show that there is a negative and significantrelation between conservatism and earning stability.
Ahmed and Duellman (2013) in a managerial overconfidence and conservatism of accounting and using both conditional and non-conditional conservatism studied the relation between managerial overconfidence and accounting conservatism. Results show that there is a reverse relation between managerial overconfidence and accounting conservatism. They also found that the externalmonitorships have no effect on the managerial overconfidence. J u l y 10, 2 0 1 4 Lara (2012) studied the relation between conservatism and limitations of earning management and concluded that conservatism is a tool to monitor the accounting procedure choice by the management; so, it can reduce the earning management opportunities based on accruals. Pae (2007)studied the effect of management view about accruals on conditional conservatism. After analyzing accruals to expected and unexpectedones, he found that: 1) the conditional conservatism presented in accruals is mainly for the unexpected accruals. 2) The negative relation between conditional and non-conditional conservatism is mainly for the unexpected accruals. 3)Companies with a higher leverage, show more conservatism mainly because of having unexpected accruals.

Rerseach hypothesis
In order to answer the above mentioned question and also after investigating the experiences and basic studies about possible answers, we will explain the formulation. In this study, a hypothesis has been formulated in order to answer the research question: Hypothesis: there is a significant relation between accounting conservatism and limitations of the accrual earning management. Ahmad and Duellmaninspired byGivoly and Hayen in conservatism measurement with accruals method, presented their criterion as the following:

Independent variable: accounting conservatism
Earning before accruals(net income) minus operational cash flow plus deplation cost, homogenized through dividing by the average of the revenue of a 3-year timeframe with the central year t, by (-1), (deullman, 2007). Whatever the conservatism amount calculated with this criterion is higher and more positive, it shows more conservatism. Therefore, calculating the average (prorating) during three timeframes, makes it sure that the effects of any big and temporary accruals is reduced; for, the accruals becomereverse during a one to two year timeframe (Lara, 2012). In this study, the recent criterion is used to measure the conservatism accruals.

Dependent variable: earning management based on accruals
In order to measure the earning management based on accruals, the modified Jones model is being used as follows: TAj,t: total company's accruals j in year t is made in the following relation: Accruals=changes in the current revenues-changes in the current debt-changes in cash and cash equivalent +changes in the current share of long term debt-Tangible and intangible fixed assets depletion costs ∆ j,t : differences in company's revenue j in year t and t-1 PPEj,t: Property,Plant and Equipment of the company j in year t The above mentioned model (1381 to 1385 in this research) has been estimated and the variables are being calculated, then with the use of the following model (modified Jones) normal accruals (as a percentage of total assets of the previous cycle) have been calculated. J u l y 10, 2 0 1 4 j,t = β 1 1 Asset j,t−1 + β 2 ∆REV j,t − ∆AR j,t Asset j,t−1 + β 3 PPE j,t Asset j,t−1 In the above formula NA shows the normal accruals and AR is the account receivable. In order to calculate abnormal or discretionary accruals the following formula is being used: AA j,t = TA j,t Asset j,t − NA j,t In the above formula AA is the discretionary accruals which being treated as the management financial reporting decisions. So, the dependent variable is obtained.

Control variables:
Business cycle: to what extent the business size is bigger, companies use more manipulation in accruals for earning management. And it's expected to have a positive relation with the discretionary accruals. In order to measure the business cycle,the account receivable cycle and the inventory are subtracted byaccount payable cycle using the early year balances (Lara, 2012).

Account payable cycle -(account receivable cycle + inventory cycle) = business cycle
Assets turnover:assets turnover is a ratio that makes it possible for us to find out to what extent efficiency of the company's sources are used and to what extent the management has used its limited sources as optimal; thus, it is obtained from the earnings divided by total assets at the beginning of the year (Lara, 2012).

Assets turnover= earnings before intersts and taxes average total assets
Sales growth: is equal to the change of the changing rate in annual sales and it is obtained through the difference between the current and last year sales divided by the last year's sale.

Methodology
The current research is of the correlative and descriptive researches, and in terms of goals, is an applied study. The population of the study based on its practical characteristics in the capital market, includes all the accepted companies in stock market. And the population of this research has companies which are not financial brokers and their end of the financial year is at the end of month March each year and the information related to them is also available. According to the above conditions, 86 companies were selected. In order to select the research samples, the systematic elimination sampling has been used. In this method, the necessary condition to select the sample is first described, and then the samples lacking those conditions are eliminated. The reason to use this method and describing such conditions is to homogenize the population with the whole society and the possibility to generalize results to the population. Eviews7 software is also used to analyze the data, and

Calculating the regression coefficient of modified Jones model
In this study, Excel software is used to sort the data and calculate some, and also Eviews-7 software is used to perform necessary regressions to estimate the statistical models and the models to make the discretionary accruals operational. As explained before,the modified Jones model is as follows: In order to obtain coefficients β1,β2 and β3, modified Jones model at first will be estimated through using a 5-year data during the years 1381 to 1385 and for 86 sample companies( totally 430 year-company).It should be noted that, in order to calculate accruals, the information existing in the year 1380 was also used. After performing the above regression and getting the coefficients, the level of discretionary accruals was firstly calculated and then by subtracting it from the total accruals, the discretionary accruals level which is being used as an index of earning management was obtained.

Hypothesis testing model
We can use the following model for the hypothesis testing: , = 0 + 1 , + 2 , + 3 , + 4 , + , J u l y 10, 2 0 1 4 In the above model, coefficient β1 shows the relation between the conservative independent variable and the earning management dependent variable using the accruals. For the hypothesis testing the coefficient β1significance and t-test is being used. The null hypothesis and alternative hypothesis in t-test is as follows: H0: β1=0 H1: β1≠0 If the null hypothesis of the t-test is rejected, it shows the significance (not being null) of coefficient β1 , so the research hypothesis would be accepted. The significance test of control variables' coefficient is also the same. For the hypothesis testing we will perform the regression with the use of hypothesis testing model. To do so, we should first specify an appropriate regression model.At the beginning we should demonstrate the effects related to the timeframe regression is performed and then the generalized F-test is done that the results are as follows in table 4-4: According to the text,f statistic is not less than 5% and the fixed and random effects have no subjectivism and the effects of sectionmodel will remain as non-effect. But since the probability gained of F statistic is less than 5% in a timeframe, it should be specified that the model of timeframe would be random or fixed. So, by randomizing the related effects to the timeframe, regression is performed and Hausman test is done which the results can be seen in the table 4-5: Null hypothesis of Hausman test indicates the randomness of the effects and according to the possibility gained for the statistic chi square,that is less than 5%, null hypothesis of Hausman test is rejected in confidence level of 95% and therefore the alternative hypothesis, will be acceptedbased on the consistency of time effects. Hence, the regression ‫صحیح‬model, is the effect model of fixed time. So, for the null hypothesis, the effect model of fixed time will be used and also to lessen the effect of outliers on the regression fitting, generalized least square method in the timeframe is being used and to remove any possible variance heteroscedasticity, White test has been used. The regression results with the mentioned method are shown in table 4-6: J u l y 10, 2 0 1 4 As it's seen in the above table, the probability gained for f statistic is less than 5%, so, we can reject Fisher test null hypothesis according to its non-significance regression model. Thus we can say that at each 95% of confidence level, regression is significant. Hence, the amount gained for Durbin-Watson statistic (2/32) is between 1/5 to 2/5 and we can say that the model errorsdon't have afirst level of serial correlation. Now,according to the above mentioned cases, we can continue investigating variables of significance gained from regression using t-test. As we explained earlier, if the probability gained of T statistic for each coefficient is less than acceptableerror level(5% in here), the t-test null hypothesis is rejected and the alternative hypothesis based on significance of the related coefficient is accepted. The probability gained for variable coefficient of conservatism (β1) is 0/0093 that is less than 0/05 and as a result this coefficient is significant and the research hypothesis at the confidence level of 95% is accepted.
The probability of t statistic for the related coefficients to the control variables wasn't also less than 5%, so, we can't say that the mentioned coefficients are significant. On the other hand, the control variables don't have any relation with the dependent variables.
Also, the modified adjusted coefficient is obtained that is almost equal to 0/11 and shows that about 11% of the dependent variable changes (earning management using the accruals) are explained via the right side variables of the model(control and independent variables).

Conclusion and suggestions
As it's explained before, we expected that conservatism has a relation with limitation in earning management. The results of hypothesis testing show that it is being accepted in confidence level of 95%. If the probability gained for the coefficient of each variable on the right side of the model of testing is less than 0/05, null hypothesis of t-test based on variable's nonsignificance (Being Zero) at the confidence level of 95% is not accepted, and we can claim that there's a relation between the related variable and dependent variable. According to the probability gained for the conservatism variable which is less than 0/05, the null hypothesis of t-test based on variable's non-significance (being Zero) at the confidence level of 95% is rejected, and we can claim that the conservatism variable coefficient is significant (not Zero); thus at the confidence level of 95% the relation between independent variable(conservative) and dependent variable(earning management) is accepted. So, we can claim that there's a relation between conservatism and earning management using the accruals.The coefficient signobtained for the conservatism variable also indicates that there's arelation between these two variables. According to this coefficient being negative, we can say that there is a negative relation between conservatism and earning management using accruals. Thus,with the rise ofand increase of conservatism, earning management is decreased. According to these cases, it can be said that the research hypothesis is accepted at the confidence level of 95%. The probability also obtained for the control variables is less than 0/05 and the relation between these variables and dependent variable (earning management) is not confirmed. Non-significance of the control variables also means that by eliminating each of them, therewon't be any decrease in the coefficient of determining the performed regression model. The research findings indicate that the hypothesis is accepted and that there's a relation between conservatism and earning J u l y 10, 2 0 1 4 management limitation. In the following, studies similar to the subject and content in this research and their results will be presented. Lara (2012) concluded that conservatism is a tool to monitor the selection of accounting procedures by the management, so, they can reduce earning management opportunities based on accruals.
Lafond and Watts (2008) studying the relation between conservatism and managers' profit-seeking incentives, found that conservatism is considered as an effective mechanism to decrease managers' profit-seeking intensives. Watts (2003) considers conservatism as an effective mechanism in contracts between the company and providers and stakeholders,which protects them against management opportunisms.
In addition to the above mentioned researches which have similar results with the current study,some other researches have also incompatible findings with the above results. For example, we can mention the followings.
Lara et al (2012) studied the earning management effect on the asymmetry of the earnings. These researchers concluded that earning management leads to the increase of earning conservatism size.However, the increase in conservatism is due to the operational and business environment, and finally increase in conservatism is along with the asymmetry of earnings.
Mehrani,EbrahimiKurdlar and Hallaj (1390) found that there is a positive and significant relation between discretionary accruals gained from modified Jones model and conservatism. They used Basu model in making conservatism operational.
The important point is that in the current research, conservatism entered the model as an independent variable; therefore, we can say that we investigated the effect of conservatism on earning management in this study, and have achieved same results like researches having similar procedures. Researches having diverse results, have used conservatism mainly as the dependent variable, and have investigated earning management effect on conservatism. It is possible that incompatible result of the two sets of studies is because of different role of conservatism and earning management(dependent or independent variable) in the hypothesis testing model.

Research limitations
 This research results are based on the criteria used in making research variables operational, and using other models for making operational the variables like earning management based on accruals and conservatism can lead to variety of results.
 The selected companies in the sample are the limited number of accepted companies in Tehran Stock Market; therefore, generalizing the results to all the business units being now accepted in the stock market should be done carefully.  Timeframe of the current study was 1381 to 1391, so, generalizing the results to other times should be done carefully.  This study's results are based on the accepted companies' information in Tehran stock market, and are not to generalize to companies out of stock.

Practical suggestions
 It is suggested to the standard institutions to pay particular attention to the role of conservatism in decreasing the management's opportunistic behavior in manipulating the earnings.  It is suggested to the investors to pay more attention to the signals based on the use of conservative accounting procedures, and mention the conservative procedures in accounting standards as much as possible.


It is suggested to the investors to pay more attention to the signals based on using conservative accounting procedures, and interpret them as a sign according to the non-opportunistic behavior of the management.

Suggestions for the future researches
 It is suggested that the current research is better to be done again using other management index of earning and conservatism.  According to the case that this conservatism has led to the decrease in earning management using accruals, studying the effect of conservatism on earning management using real activities should show that if the management , through reacting to the limitation created by conservatism to perform earning management with the use of accruals, turns to other earning management procedures or not.  Performing the present research once more in another timeframe.  Performing the present research in different industries and comparing the results for various industries.