The effect of earnings smoothness on manufacturing company’s performance

The purpose of this study is to determine empirically the effect of earnings smoothness on company’s performance. The company’s performance used in this study is based on two indicators of the company's operational performance (ROA) and market performance (Tobin's Q). In addition to earnings smoothness as the independent variable and company’s performance as the dependent variable, this study also uses the control variable leverage and size. The sample used in this study based on the criteria of sampling as many as 96 manufacturing companies listed in Indonesia Stock Exchange during the years of 2005-2010 so that the number of data samples 576. According to anova F test in linear regression show that models of regression can be used to predict the company's operational performance and market performance. While the results of the anova t test in linear regression show that earnings smoothness significantly affect the market performance. However, earnings smoothness does not significant affect the company's operational performance. Control variables are leverage and size results show the opposite of the independent variable smoothness profit, that significantly influence the company's operational performance (ROA) but not significantly with market performance (Tobin's Q).


INTRODUCTION
Financial statement is an important part that should be considered well in a company. The financial statement can be used to see the state of the company, whether it is in an excellent performance, in order to be able to attract a lot of investors to invest. One of important aspects that can be seen in financial statement is earnings information provided by the company. A company can be said to be in a good condition if the company has good earnings. Therefore the company's earnings should be in good quality in order to attract investors.
Earnings quality is an amount that must be maintained by the management in a period that it should remain stable in generating profits. Francis, et al. (2006) in Margani and Meinarni (2009) suggested that earnings quality is an important characteristic of financial reporting. Low earnings quality can undermine investors' confidence in the information presented in the company's financial statement.
Earnings quality can be measured by a variety of measurements. One of them is by using Earnings Smoothness. Margani and Meinarni (2009) revealed that earnings smoothness generally uses cash flow as a construct reference to unsmoothed earnings and assumes that the cash flows are not manipulated (unmanaged). Earnings smoothness reflects the idea that managers use information about forthcoming earnings to smooth earnings so as to produce more representative earnings (normalized). Representative earnings or smoothed earnings indicate that the future earnings are in high quality. More smooth earnings also describe high quality earnings (Margani and Meinarni 2009). High value of smoothness indicates less smoothed earnings or getting unsmoothed, so the earnings quality produced is good. Whereas low value of smoothness indicates the big amount of smoothed earnings or getting smoothed, so the earnings quality produced is bad. Dechow and Schrand (2004) in Mahmud, et al. (2009) also explained that the high quality of earn-ings can reflect the company's performance in the current period and can be a good benchmark for the performance of the company in the next period. Therefore the earnings quality may be the appraisal of the company's performance (Mahmud, et al. 2009). Company's performance can be measured based on two perspectives namely internal perspective and external perspective. Based on the internal perspectrive, the performance of the company can be measured using financial performance or Return On Asset (ROA). ROA shows company's ability to utilize the assets owned by the company with net profit produced. Then, based on the external perspective, the performance of the company can be measured using market performance or Tobin's Q. Tobin's Q is used to determine the valuation of market to the company. Based on the research conducted by Mahmud et al. (2009), it can be concluded that earnings quality has a positive relationship with company's performance using measurement of ROA and Tobins Q. ROA measures the financial performance, while Tobins Q measures the market performance of the company. Both of these measurements include the overall company's performance because it is measured from different perspectives (Mahmud, et al. 2009).
The formulation of the problem in this study is whether earnings smoothness affects company's performance. The purpose of this study is to investigate empirically the effect of earnings smoothness on company's performance.

THEORETICAL FRAMEWORK AND HYPO-THESIS Previous Research
Ririk Retnowati (2011) examined the effect of accounting-based earnings quality on the performance of manufacturing company in the period 2006-2007. One of the results of her research proves that earnings smoothness does not significantly affect the company's performance (ROA and Tobin's Q). The dependent variable in this study is the company's performance, (ROA and Tobin's Q), while the independent variable used is the earnings quality. Earnings quality is measured using 6 measurement techniques: earnings persistence, predictability, variability, earnings smoothness, abnormal accruals, and accruals quality.
Radziah Mahmud et al. (2009) examined the effect of earnings quality on the performance of Malaysia companies in the period [2000][2001][2002][2003][2004][2005][2006][2007]. The result shows that earnings quality (prediction, feedback value, and timeliness) has an effect on copany's performance (ROA and Tobin's Q). Margani and Meinarni (2009) revealed that earnings smoothness generally uses cash flow as a reference construct on unsmoothed earnings and assumed the cash flow which is not manipulated (unmanaged). Earnings smoothness reflects the idea in which managers use information about forthcoming earnings to smooth earnings so as to produce more representative earnings (normalized). Representative earnings or smoothed earnings indicate that the future earnings are in high quality. More smooth earnings also describe a high earnings quality (Margani and Meinarni 2009).

Theoretical Basis Earnings Smoothness
The measurement of earnings smoothness used in this research is: (1) NIBE = Net profit before company's extraordinary items divided by total initial asset of the company. CFO = Company's cash flow operating activities divided by total initial asset of the company. Retnowati (2011) explained that the measurement of company's performance can be conducted from two different viewpoints, namely internal perspective and external perspective. Seen from the internal perspective, company's performance can be measured using financial performance (ROA). Return on Assets (ROA) shows the company's ability to utilize the assets owned by the company with net profit produced, which means that the company can make use of the company's assets to generate high profits so that the company can recover the investment that is embedded in the company's assets. Meanwhile, viewed from an external perspective, the company's performance can be measured using market performance (Tobins'). Tobin's Q is used to determine the market valuation on the company, and thus the company's ability to form a stock price in the capital market can be recognized (Ririk 2011).

Leverage and Size
Jayanti (2011) revealed that leverage is a tool used to measure how much the company depends on the creditors or lenders to finance the company's assets. Company, that has a high level of leverage, indicates that the company is highly dependent on external loans to finance the company's assets. While the compan,y that has low level of leverage, indicates that the company finances its assets with its own capital. Igan (2007) explained that size (company size) is a scale that can be classified based on the calculation of the company size, one of which is log total assets. Company size is divided into three categories, namely large, medium, and small company. In general, companies that have relatively large total assets can operate in a higher level of efficiency compared to companies that have lower total assets. Therefore, companies with large total assets can generate higher level of profits well (Chelsiya 2010). Huang, et al. (2009) in Mahmud, et al. (2009) concluded that high earnings quality can lead to higher company's performance. Conversely, the company's performance declines due to lower earnings quality (Mahmud, et al. 2009). However, based on the research conducted by Ririk (2011) about the earnings quality on the company's performance in which one of its earnings quality measurement using earnings smoothness, proved that earnings smoothness does not significantly affect company's performance (ROA). Earnings Smoothness is not significant on company's performance (Tobin's Q).

Relationship between Earnings Smoothness and Company's Performance
Framework of thought underlying this study can be described in Figure 1. Based on the previous research and theoretical basis above, it can be formulated a hypothesis as follows: H1: Earnings smoothness affects company's performance.

RESEARCH METHOD Research Design
This research is a quantitative study that tests a theory through a study variables concerned with the aim to test the hypothesis (Indriantoro and Supomo 1999:27) Based on the research data sources, this study includes the research that uses secondary data sources. Secondary data is a research data source obtained indirectly through an intermediary me-dium (obtained and recorded by other parties). Secondary data is generally in the form of evidence, records and historical reports that have been arranged in the archive (documentary) both published and unpublished (Indriantoro and Supomo 1999:27) Variable Identification Based on framework that has been compiled, the variables used as the guidence of discussion in this study are as follows: Independent variable used in this study is earnings smoothness. While the dependent variable used in this study is company's performance with two indicators: ROA (Company's Operating Performance) and Tobin's Q (Market Performance). Control variable used in this study is Leverage and Size (Company size).

Operational Definition and Measurement of Variabel Independent Variable
Earnings smoothness is measured based on the size of cash flow. This measurement uses cash flow as unsmoothed earnings, and assumes that the cash flow is not manipulated (unmanaged) (Margani and Meinarni 2009 High value of earnings smoothness indicates the small earnings smoothness or getting unsmoothed, so the earnings quality generated is good. Conversely, small value of earnings smoothness indicates the big earnings smoothness or getting smoothed, so the earnings quality generated is not good.

Earning Smoothness
Company's Performance:

Leverage Size
Tobin's Q

Dependent Variable
This study uses two measures of company's performance that are based on two different perspectives, i.e. internal perspective and external perspective. Based on the company's internal perspective, the measurement of company's performance used is the Return On Asset (ROA). According Ririk (2011) ROA can be used to measure how the company's ability to manage the assets existing in the comapny so that high profit can be obtained. With high profit, then the company will be able to recover the investment that is embedded in the company's assets (Ririk 2011). ROA used in this study is ROAt +1. The calculation of ROA is as follows: While based on the external perspective, the measurement of company's performance used is Tobin's Q. According Ririk (2011) Tobin's Q can be used to measure how the market valuation agains the company. The components of Tobin's Q calculation is using the closing price of the stock, the number of shares outstanding, the total long-term debt of the company, and the total assets so that it can be seen how the company's ability to form a stock price in the stock market so that investors are interested to invest (Ririk 2011). Calculation of Tobin's Q is as follows: where: MVE = Closing Price of Stock x The number of Share Outstanding

Control Variable
Jayanti (2011) revealed that leverage is a tool used to measure how much the company depends on the creditors or lenders to finance the company's assets.
The calculation of leverage is as follows: Igan (2007) explained that the size (company size) is a scale that can be classified based on the calculation of company size, one of which is log total assets. The calculation of size is as follows:

Population, Sample, and Sampling Technique
The population of this study is manufacturing companies listed in Indonesia Stock Exchange. This study uses period of study from 2005 to 2010. The sample collection is conducted using purposive sampling method with the specified criteria. The sample criteria used in this study is manu-facturing companies listed in Indonesia Stock Exchange from 2004 to 2010 respectively, revealing a complete annual report from 2004 to 2010 respectively, having financial data necessary in research as well as expressed in rupiah currency, and fulfilling the assumptions of data not outliers.

Data Analysis Technique
This study uses multiple linear regression analysis technique to determine whether there is a significant effect of the independent variable on the dependent variable, so multiple linear regression models are used to be formulated into the regression equation as follows:

DATA ANALYSIS AND DISCUSSION Desriptive Statistics Test
Descriptive analysis is the analysis that provides description of data seen from the average value (mean), deviation standard, variance, maximum, minimum, sum, range, kurtosis, and skewness. (Imam 2011:19). The descriptive analysis of the data in this study uses a statistical test with a different test more than two sample groups (ANOVA).

Sample Description
Based on the sampling criteria, the number of samples in this study is 96 companies listed in Indonesia Stock Exchange, as a hypothesis testing toward the effect of earnings smoothness on the pecompany's performance conducted for 6 years of research, then 96 companies multiplied by 6 years of research so that the number of data is 576 data. The proportion between the mean value of earnings smoothness which is above the overall mean value during the years of study and the mean value of earnings smoothness which is below the overall mean value during the years of study is 2: 4 or with a percentage of 33.33%: 66.67 %. So it can be seen that from 2005 to 2010, the annual mean value which is below the overall mean value of earnings smoothness is a lot more than the annual mean value which is above the overall mean value of earnings smoothness. So based on the observation years that are seen from the overall mean value, there are still many manufacturing companies that have poor or less good earnings quality during the 6 years of the study.

Description of Variable
Based on the mean value of earnings smoothness and deviation standard of earnings smoothness, the deviation standard value of earnings smoothness which is getting closer to the mean value of earnings smoothness or the value range produced is small, it shows better earnings quality than the deviation standard value of earnings smoothness which is further away from mean value of earning smoothness or the value ranges of earnings smoothness produced is big. From 2005 to 2010 shows that the deviation standard value of earnings smoothness is away from the mean value of earnings smoothness or the range value produced is big in 2010. The mean value of earnings smoothness in 2010 is 1.435064 with the deviation standard of 3.4877969, so the range value is 2.0527329. The range value is greater than the other observation years. So the deviation standard value of earnings smoothness in 2010 is away from its mean value, which means that in 2010, the earnings quality is not better than the other observation years.

Description of Company's Performance: ROA Indicator
Return on Assets (ROA) can be used to measure how the company's ability in managing company's assets so as to obtain high profit. With high profit, the company will be able to recover the company investment that is embedded in the company's assets (Ririk 2011).
Based on the overall mean value of ROA during observation years, the annual mean value of ROA that exceeds the overall mean value of ROA during the observation years is in 2008, 2009, and 2010. So from 2008 to 2010 has the mean value of ROA above the overall mean value of ROA during the observation years compared to the other years. Therefore, during the years the manufacturing companies that are made sample in this study are more capable of producing greater profit than the assets owned by the company so that the company's operational performance is better than the other observation years.
Instead in 2005 to 2007 has mean value of ROA below the overall mean value during the observation years. So, in 2005 to 2007 manufacturing companies which are made as the sample in this study are not more capable of producing greater profit than the assets owned by the company so that the company's operational performance is not better than the other observation years, because the mean value is below the overall mean value of ROA during observation years.
The proportion between the mean value of ROA that is above the overall mean value during the study years and the mean value of ROA below the overall mean value during the the study years is 3 : 3 or with a percentage of 50%: 50%. So it can be seen that from 2005 to 2010 the annual mean value which is below the overall mean value of ROA is as much as the annual mean value which is above the overall mean value of ROA.
Based on the mean value of ROA and the deviation standard of ROA, standard deviation value of ROA which is close to the mean value of ROA or the value range produced is small, it shows that the companies are more capable of producing a better earnings of the assets owned by the company so that the company's operational performance is better than the deviation standard value of ROA which is far from the mean value of ROA or the value range produced is great. From 2005 to 2010, the year which shows the standard deviation value of ROA away from the mean value of ROA or the value range produced is great is in 2007. The mean value of ROA in 2007 is 4.763125 with the standard deviation of 10.0738246, so the value range produced is 5.3106996. The value range is greater than the other observation years. So the standard deviation of ROA in 2007 is far from its mean value, which means that the manufacturing companies in 2007 which are made sample in this study are not more capable of producing greater earnings of the assets owned by the company so that the company's operational performance is not better than the other observation years.

Description of Company's Performance: Tobin's Q Indicator
Tobin's Q is one of the company's performance measurements that are seen from external perspective, where the measurement has been adjusted to the company's financial transactions in Indonesia. Tobin's Q can be used to measure how the market valuation against the company (Ririk 2011).
Based on the overall mean Tobin's Q during the observation years, the annual mean value of Tobin's Q that exceeds the overall mean value of Tobin's Q during the observation years is in 2005 and in 2008. So in 2005 and in 2008 have mean value of Tobin's Q above the overall mean value of Tobin's Q during the observation years compared with the other years, then in those years the market valuation against the company is highly valued by investors so that the company's market performance is better than the other years. In contrast, in 2006In contrast, in , 2007In contrast, in , 2009, and 2010 have mean value of Tobin's Q below the overall mean value during the observation years. So in 2006So in , 2007So in , 2009, and 2010 market valuation against the company is not valued highly by investors and the company's market performance is not better than the other years, because its mean value is below the overall mean value of Tobin's Q during the observation years.
The proportion between the mean value of Tobin's Q which is above the overall average during the study years and the mean value of Tobin's Q below the overall mean value during the study years is 2: 4 or with a percentage of 33.33%: 66.67 %. So it can be seen that during 2005 to 2010 the annual mean value which is under the overall mean value of Tobin's Q is greater, with the annual mean value above the overall mean value of Tobin's Q.
Based on the mean value of Tobin's Q and the standard deviation of Tobin's Q, the standard deviation value of Tobin's Q which is close to the mean value of Tobin's Q or the value range produced is small, it shows that market valuation against the companies is valued highly by investors so that the company's market performance is better than the standard deviation value of Tobin's Q which is far from the mean value of Tobin's Q or the value range produced is great. From 2005 to 2010 shows that the standard deviation value of Tobin's Q is away from the mean value of Tobin's Q, or the value range produced is quite large in 2008. The mean value of Tobin's Q in 2008 is 0.3139038 with a standard deviation of 1.405, and then the value range produced is 1,091. The value range is greater than the other observation years. So the standard deviation value of Tobin's Q in 2008 is far from its mean value, which means that the market valuation against the company is not valued highly by investors, so the company's market performance is not better than the other years.

Description of Control Variable: Leverage
Leverage is a tool used to measure how much the company depends on the lenders or creditors to finance the company's assets. Companies that have high level of leverage indicate that the companies are highly dependent on external loans to finance then their assets. While companies that have low level of leverage indicate that the companies finance their assets much more using owned capital (Jayanti 2011).
Based on the overal mean leverage during the observation years that is 0.52, the annual mean value of leverage that exceeds the overall mean value of leverage during observation years is in 2005, 2007, and 2008. So in 2005, 2007, and 2008 the mean value of leverage is above the overall mean value of leverage compared with the other years.So during the years, the manufacturing companies that are made sample in this study reflects the company's operational performance which is not better because the capital owned by the company is less capable of financing the company's assets. While the company's performance is better because the company is capable of utilizing its debt for the purpose of expansion of the company so as to attract investors to invest in the company compared with the other years.
Conversely, in 2006, 2009 and 2010 the mean value of leverage is below the overall mean value during the observation years. So in 2006, 2009, and 2010 the manufacturing companies that are used as the sample of this study reflect better company's operational performance because the capital of the company is capable of financing company's assets, while the market performance is not better because the company is less capable of utilizing its debt for the purpose of expansion of the company making it less able to attract investors to invest into the company compared to the other observation years.
The proportion between the mean value of leverage which is above overall mean value during the study years and the mean value of leverage which is below overall mean value during the study years is 3 : 3 or with a percentage of 50%: 50%. So it can be seen that from 2005 to 2010 the annual mean value which is above the overall mean value of leverage is as much as the annual mean value which is under the overall mean value of leverage.
Based on the mean value of leverage and standard deviation of leverage, the standard deviation value of leverage which is close to the mean value of leverage or the value range produced is small, it reflects that the company's operational performance is not better because the company's capital is less capable of financing company's assets, while its market performance is better because the company is capable of utilizing its debt for the purpose of expansion of the company so as to attract investors to invest in the company, compared to the standard deviation of leverage which is away from the mean value of leverage or the value range produced is great. During 2005 to 2010, it indicates that the standard deviation value of leverage is away from the mean value of leverage or value range produced is great enough in 2008. The mean value of leverage in 2008 is 0.543229 with a standard deviation of 0.2275660, so the value range produced is 0.315663. The value range is greater than the other observation years. So the standard deviation of leverage in 2008 is far from its mean value, which means that manufacturing companies in 2008 which are made as the sample of this study reflect that the companys' operational performance is better because the company's capital is capable of financing the company's assets, while the performance of the market which is not better because the companies are less capable of utilizing its debt for the purpose of expansion of the company making it less able to attract investors to invest into the company.

Description of Control Variable: Size
Size or company size is a scale which can classify the size of the company according to a variety of ways, including log total assets, market value of shares, and others (Igan 2007).
In general, companies that have relatively large total assets can operate with a higher level of efficiency compared to companies with lower total assets. Therefore, companies with higher total assets are more capable of producing higher level of profit and this reflects that the company's performance is good (Chelsiya 2010). This study uses size as a control variable in the calculation of the log total assets. Based on the overall mean total assets during the observation years, the annual mean value of total assets that exceeds the overall mean value of total assets during the observation years is in 2008, 2009, and 2010. So, from 2008 to 2010 have mean value of total assets above overall mean value of total assets compared to the other years, then during the year, manufacturing companies which are made as the sample of the study are more capable of producing higher level of profit and it reflects that the company's performance is better than the other years.
Conversely, from 2005 to 2007 have mean value of total assets below the overall mean value during the observation years. So, from 2005 to 2007 manufacturing companies that are made as the sample of this study is not more capable of producing higher level of profit and it reflects that the company's performance is not better than the other years, because the mean value is below the overall mean value of total assets during the observation years.
The proportion between the mean value of total assets which is above overall mean value during the study years and the mean value of total assets which is below overall mean value during the study years is 3:3 or with the percentage of 50% : 50%. So it can be seen that from 2005 to 2010 the annual mean value which is above overall mean value of total assets is as much as the annual mean value which is below overal mean value of total asset.

Classical Assumption Test
Classical assumption test is used to test the normality of research data, to conduct multicoliniearity test against independent variables, and to test autocorrelation to know whether there is a correlation between the residuals in the regression model, as well as to conduct heteroscedasticity test to know the presence of variance from the residual of the regression model. Normality test results with a one-sample Kolmogorov-Smirnov test (Priest 2006: 147) shows that the only data from leverage variable which is distributed normally. Residual normality test results for residual models show abnormal distribution. After the transformation is done, it is obtained the same result, therefore the analysis using the original data, not from the result of trnaformation with the reason so as to represent the true meaning.
Multicolonierity test results show the value of tolerance and the variance inflation factor (VIF) indicates that the overall variables do not contain multicoliniearity with tolerance values of more than 0.1 and variance inflation factor (VIF) less than 10. The test result of run test for detecting autocorrelation shows that all variables contain autocorrelation because the significant value is less than 0.05. This occurs since the data processed is time series data. The test results using glejser test show that based on ROA indicator, variable of earnings smoothness and size the significant value is more than 0.05, so it does not contain heteroscedasticity. While based on Tobin's Q indicator the significant value of leverage is above 0.05, so leverage variable does not contain heteroscedasticity.

Hypothesis Test
Research Model Test is conducted using F test, the model is stated good (fit) if the result of significance/probability of F test is less than 0.05. While tesing of the effect of earnings smoothness on the company's performance is using t test. Similarly, the variable model test of earnings smoothness is said influential if the result of influence test produces the value of significance/probability less than 0.05.
Based on the results of linear regression analysis using SPSS for Windows 17.0, the results show that testing of F test with the ROA indicator is the value of F count of 42.835 with a significant level of 0.000 smaller than 0.05, then Ho is rejected which means that the model regression can be used to predict the company's operational performance (ROA) or it can be said that earnings smoothness, leverage, and size simultaneously affect the company's operational performance (ROA). While using Tobin's Q indicator, the value of F count is 11.605 with a significant level of 0.000 smaller than 0.05, then Ho is rejected which means that regression model can be used to predict market performance (Tobin's Q) or it can be said that the earnings smoothness, leverage, and size simultaneously affect the market performance (Tobin's Q).
Based on the results of linear regression analysis using SPSS for Windows 17.0, the results show that the test of t test with the ROA indicator is the value of t count of earnings smoothness variable (b) is -0.251 with significant level of 0.802 which is greater than 0.05, it means that there is no earnings smoothness (b) effect on company's operating performance (ROA). In addition the value of t count of leverage variable is -10.126 with a significant level of 0.000 which is less than 0.05, it means that there is leverage effect on company's operating performance (ROA). And the value of t count of size variable is 5,900 with a significant level of 0.000 which is less than 0.05, it means that there is size effect on the company's op-erating performance (ROA). While using Tobin's Q indicator, the value of t count of earnings smoothness variable (b) is 5.714 with a significant level of 0.000 which is less than 0.05, it means that there is no earnings smoothness effect on the market performance (Tobin's Q). Moreover the value of t count of leverage variable is 0.688 with a significant level of 0.492 which is greater than 0.05, it means that there is no leverage effect on the market performance (Tobin's Q). And the value of t count of size variable is -1.709 with a significant level of 0.088 which is greater than 0.05, it means that there is no size effect on market performance (Tobin's Q).

Discussion
Based on the test result of F test, it indicates that the regression model can be used to predict the company's operational performance (ROA) and market performance (Tobin's Q) or it can be said that earnings smoothness, leverage, and size simultaneously affect the company's operational performance (ROA) and market performance (Tobin's Q). While the test results at the time t test is conducted show that earnings smoothness significantly affect the market performance (Tobin's Q). While earnings smoothness does not significantly affect the company's operating performance (ROA). The result of control variables of leverage and size indicates the opposite of the independent variable of earnings smoothness; it significantly affects the company's operational performance (ROA) but does not significantly affect market performance (Tobin's Q).
Several things that cause earnings smoothness do not significantly affect company's operational performance (ROA), among others are; the proportion of the mean value of earnings smoothness above the overall mean value compared to the mean value of earnings smoothness below the overall mean value in which the mean value of earnings smoothness is a lot more below the overall mean earnings smoothness, the standard deviation value which is getting far away from the mean value or the value range produced greater from earnings smoothness variable and ROA, based on the classical assumption test that uses the company's operational performance as dependent variable show that the sample data of the research are not normal and prove that there is autocorrelation. Several things mentioned above can affect the results of hypothesis test in this study.
Results of this study agree with the research conducted by Ririk (2011) which states that earnings smoothness does not significantly affect the company's performance which is measured using company's operational performance namely ROA. However, this study does not agree with the study of Mahmud, et al. (2009) which states that earnings quality significantly affects the company's performance as measured by ROA.
Several things that cause leverage and size do not significantly affect the market performance (Tobin's Q), among others are: the proportion of the mean value of Tobin's Q above overall mean value compared to the mean value of Tobin's Q below the overall mean value in which average value of Tobin's Q is a lot more below the overall mean of Tobin's Q, the value of standard deviation is further away from the mean value or the value range produced great from the variable of log TA (size), based on the classical assumption test using market performance (Tobin's Q) as dependent variable shows that the sample data of the research is not normal, there is autocorrelation, and there is heteroscedasticity. Results of this study agree with the research conducted by Mahmud, et al. (2009) which states that the earnings quality significantly affect company's performance as measured by Tobin's Q. But do not agree with Ririk research (2011) which states that earnings smoothness does not significantly affect the company's performance as measured by the market performance namely Tobin's Q.

CONCLUSION, IMPLICATION, SUGGES-TION, AND LIMITATIONS
The purpose of this study is to determine empirically the effect of earnings smoothness on Company's performance which is measured using the company's operating performance (ROA) and market performance (Tobin's Q). The sample used in this study is based on sampling criteria as many as 96 manufacturing companies listed in Indonesia Stock Exchange from 2005 to 2010 so that the total sample are 576 data. The data used in this study is obtained from the financial statements of companies listed in Indonesia Stock Exchange and the Indonesian Capital Market Directory (ICMD) as a supporting tool. Testing for this study is performed using SPSS 17.0 for linear regression, using the help of One Way Anova test for data description, and also the classical assumption test.
Based on the hypothesis test with t test, it can be concluded that earnings smoothness significantly affect the company's performance as measured by market performance (Tobin's Q), and the control variables in this study, namely leverage and size have no significant effect on the company's performance as measured by market performance (Tobin's Q). While earnings smoothness has no sig-nificant effect on company's performance as measured by the company's operational performance (ROA), and the control variables used in this study namely leverage and size have significant effect on company's performance as measured by the company's operational performance (ROA).
Descriptively, this study reveals poor or less good earnings quality, because when it is tested, the normality shows that the study data are not normally distributed. However, when tested simultaneously with the F test, a research model test, show that the regression model can be used to predict the company's operational performance (ROA) and market performance (Tobin's Q), it is viewed by a significant value of 0.000 which is less than 0.05.
The limitations of this study are that this study still contains disease of classical assumption test. Testing of pooled data is not choosing the best method, random or fixed effect. Future studies should pay attention to these by utilizing testing tools that facilitate the selection of methods for the data pooled.