The Effect of the Covid-19 Pandemic on Stock Returns with Moderating Variables of Cash Holding and Earnings Management

ABSTRACT


INTRODUCTION
The Covid-19 pandemic has hit many countries in the world since January 2020 [1].Even though the World Health Organization (WHO) officially declared the Covid-19 pandemic on March 11, 2020, several countries such as Israel, Russia, and the United States have already implemented a ban on traveling abroad their citizens [2].The presence of a pandemic had a serious impact on the world economy and resulted in a sharp decline in the performance of companies in various industries globally [3].The impact of a very significant economic downturn was due to the existence of strict health quarantines in various countries, causing very limited economic activity [4].To deal with operational risks and to ensure company resilience in the face of the Covid-19 pandemic, company managers tend to increase short-term cash holding levels by making bank loans, or equity funding [2].COVID-19 initially broke out in Wuhan, China, towards the end of December 2019.Then it spread to all the provinces there.In less than two months, the core coronavirus has caused 80,000 cases and 3,000 deaths.Beginning in the third week of January 2020, Covid then entered several countries in Asia, America, Europe, Australia, and Arika.When the epidemic subsided in China, the spread of COVID-19 actually exploded in a number of American, European, and Asian countries.As of April 24, 2020, the number of confirmed cases in the US, Spain, Italy, France, Germany, the UK, Turkey, and Iran has surpassed China as the initial epicenter.While in the number of deaths, there are six countries that exceed China, namely: the US, Italy, Spain, France, Britain, Germany, Iran, and Belgium.Meanwhile, the number of victims in the Netherlands is also starting to approach that in China.The COVID-19 outbreak in China has indeed subsided.Similar conditions are also experienced.anumber of European countries which in the March-May 2020 period became the epicenter of COVID-19 such as Italy, Spain, France, England and Germany.The epicenter has shifted from China to Europe and is now sweeping across the country [5].The COVID-19 pandemic has had a very significant impact on capital markets around the world, causing investors to lose money [6].
On the other hand, companies in Indonesia experience increased credit risk [7].In the midst of the pressure of the need for cash holding and the increase in credit risk, the Government of Indonesia provides credit relaxation for companies affected by COVID-19 [8].One alternative motivation to perform earnings management is to reduce political risk [9].The political cost hypothesis states that companies that are faced with political risk will tend to carry out earnings management to reduce profits so that political costs can be lowered [10].Therefore, it is very interesting to examine whether the pressure of the urgent need for cash encourages company management to practice earning management to gain the trust of creditors or minimize profits to get relaxation facilities from the government.Previous research conducted by Qin et al. (2020) [2] concluded that the COVID-19 pandemic had a significant positive impact on the company's cash holding level.Fu & Shen.(2020) [11] stated in his research results that Covid-19 had a significant negative impact on company performance in the energy industry.This is in line with the results of research conducted by Mafrolla et al. (2017) [12] where stock market indexes in countries affected by COVID-19 fell very quickly after the virus outbreak hit.However, the research on the impact of COVID-19 was carried out in China and no similar research has been conducted in Indonesia.Research on the effect of earning management practices on stock prices conducted by Hariandja et.all, (2022) [13] concludes that earnings management practices make investors set stock prices too high then they should be.AlNajjar & Riahi Belkaoui (2001) [14] in their research results state that the management of companies with a high level of investment opportunity makes accounting choices to reduce reported earnings.Research on earnings management has been widely carried out by previous researchers, but there is still little research linking it to the COVID-19 pandemic.Many previous researchers have researched earnings management and stock price variables, but few have linked these variables to the COVID-19 pandemic.In addition, previous studies conducted in Indonesia involving the COVID-19 pandemic only measured the condition of these variables before and after Covid-19 without getting sufficient confidence whether the difference in conditions was caused by the Covid-19 pandemic.With the explanation above, the purpose of this research is to determine the effect of covid-19 on the company's cash holding, earning management practices, and stock returns in the long term.In addition, this study will also examine the role of cash holding as a moderating variable for the impact of COVID-19 on earning management practices and the role of earning management as a moderating variable for the impact of COVID-19 on long-term stock returns.
The signaling theory put forward about method management company provides information about the things that have been done to realize the wishes of the owner of the company [15].From the description put forward in on, so could said that signaling theory is a theory which put forward method management company give a signal to para investors about which has done for realizing desire these investors.According to Gao et al., (2017) [16] cash holding is the amount of cash set aside by the company to fulfill the need of its finances.During the COVID-19 pandemic, a health quarantine is applied which strictly limits the activity operational company and caused a drop in company income [2].Dang et.all (2017) [17] defines earnings management as intervening significantly in the process of reporting finance external which aims for profit for shareholders or managers.Connelly et. all (2017) [18] defines earnings management as a phenomenon that occurs when business management intervenes to report finance and internal transactions by making the necessary changes to both mislead shareholders about the company's financial performance or to influence results contract based on the performance finance company.Shares are proof of participation or ownership of a person or entity in something company or company limited.A share is a certificate that states that the holder share is an investor from a company that publishes effect [19].Portion ownership is determined by how much investment is invested in the company [20] Hypothesis H1: Pandemic Covid- 19  The model study depicted in the description above is as follows in figure 1.

RESEARCH METHOD Election Sample and Respondent
The research population is the companies listed on the Stock Exchange Indonesian Effect.Sample selection is done by selecting companies that are recorded in the board recording main in Exchange Effect Indonesia.Exchange Effect Indonesia divides the industry into several subsectors.The research method used is a quantitative method.Where is Data Research that is used is the company's financial data listed on the main recording board Exchange Effect Indonesia from a period quarter I 2015 until the secondquarter year of 2021.Data were taken from Bloomberg with criteria: available data complete financial statements within the period used in the study, and no currently in pressure financial heavy more from 2 years in period study.From the method of taking samples, they obtained 191 sample companies spread over 40 subsector industries.

Measurement
In this study, the COVID-19 pandemic was treated as a treatment [2].Grouping of companies as affected companies or not affected by the covid-19 pandemic based on the analysis of Pefindo, (2020) [21] and combined with Analysis Results Survey Impact Covid-19 to perpetrator business [22].
Data and information about level cash holding, mark discretionary accruals, and stock returns of companies that are in the group treatment and comparison for the period before and after the pandemic covid-19 taken from the database Bloomberg.Variable Calculation Cash Holding level cash holding company calculated from the formula total cash / total assets Treated dummy variable indicating the degree of impact of covid-19, value 1 for high and very high impact companies, and currently and 0 for a company which caught impact light and no affected Period variable dummy which indicates the period before and after pandemic covid-19, mark 1 symbolize the period after the pandemic and 0 for the period before pandemic covid-19 Size size company which calculated from logarithm natural total asset Lev ratio total asset on total Liability Growth growth total assets company which i s calculated from formula = (total assets period t -total asset period t-1 )/total asset period t-1 ROA is ratio Among net profit on a total asset which is calculated with formula = net profit/balance end total asset FCF free cash flow, calculated from formula = Current Cash Operation -Expenditure Capital -Expenditure for repair asset  = _ change accounts receivable business company i on year t PPE it = asset permanent form company i on year t for look for coefficient ️ on equality in on so used regression linear with model following : TACC it /Ta it -1 = 1 (1/Ta it -1 ) + 2 (REV it /Ta it -1 ) + 3 (PPE it /Ta it -1 ) Treated variable dummy which indicates level impact covid-19, mark 1 for a company which caught impact tall and very high, and currently and 0 for a company which caught impact light and not affected Period variable dummy which indicates the period before and after pandemic covid-19, marks 1 symbolize the period after the pandemic and 0 for the period before pandemic covid-19 Size size company which calculated from logarithm natural total asset CAtoTA Cash ratio and equivalent Cash / Total Asset Growth growth total assets company which is calculated from formula = (total assets period t -total asset period t-1 )/total asset period t-1 ROA ratio Among net profit on a total asset which is calculated with formula = net profit/balance total end asset

RESULT AND DISCUSSION
The research was conducted using the multiple linear regression method using dummy variables.The data used is panel data, namely: the merging of time series data (a certain time span) and crosssection (many companies).9 above states that at the levin-linchu level there are several variables that are not stationary and there are variables that contain unit roots, so it is necessary to look at the variables at the level of the presence of the LM test.The results show that all panels do not move at the level of the LM test under various conditions.
From the unit root test done, several variables only fulfill one of the unit root tests performed.The focus of this research is to examine the impact of covid-19 on variables dependent.A variable that only fulfills the wrong one unit root test no Becomes the focus main in a study so that variable-variable is maintained.

Parallel Trends Test
Parallel Trends Test Variable Dependent Cash Holding

Figure 2. Parallel Trends Test Variable Depending Cash Holding
Parallel trend test which showed on picture 2 shows that no there is difference trend which significant Among group industry affected with industrial groups not affected before pandemic covid-19 Even though on some periods there is a difference in the direction of the trend, but the next period it returns show direction trend which same.Period 2017Q2-2017Q3 shows the difference in the trend where the affected industrial groups experienced a decrease in cash holding while the unaffected industrial group experienced an increase in cash holding.However, in the period 2017Q3-2017Q4 trend cash holding both of them return experience increase so the difference direction trend on period 2017Q2-2017Q3 no influence whole trend from 2015Q1-2020Q1 by significant.

Parallel Trends Test Variable Dependent Earnings Management Figure 3. Parallel Trends Test Variable Depending Discretionary Accruals
Parallel trend test which showed figure 3 shows that no there is difference trend which significant Among group industry affected with group industry no affected before pandemic covid-19 Although on a number of the period there are difference direction trends, the next period again shows the same trend direction.Period 2015Q3-2015Q4 shows the different trends in which industry groups affected experience drop mark discretionary accrual whereas group industry no affected experience increase discretionary accruals.However, in the period 2015Q4-2016Q1 trend discretionary both accruals have increased again so that the difference in the direction of the trend in the 2015Q3 2015Q4 period does not affect the overall trend from 2015Q1-2020Q1 insignificant.

Parallel Trends Test Variable Dependent Return Share
The parallel trend test shown in Figure 4 shows that there are various trends in many periods before pandemic covid-19 hit.For help To do justification, a straight line trend is made to determine the trend stock return trend before the covid-19 pandemic hit (2015Q1-2019Q4) which is symbolized with a line dashed.Trends line straight show that the group industry affected its trend which same as the group industry not affected before the covid-19 pandemic hit.

Election Model Regression linear multiple
In the model in figure 4, the study encloses the variable treated which is a variable dummy for indicates is the company the included in the group affected or not affected by covid-19 Variable treated is not affected by changes in the time variable (period).Use variable unaffected by variable change time in a model, making the model unable to use Fixed Effect Model (" FEM" ).The use of FEM can only be done if the variable is unique to variable time [23].So that there are two possibiliti es The type of regression that will be used in this research is Pooled Least Square (" PLS" ) or Random Effect Model (" ️REM" ).

Determination of the best regression model between PLS and REM is done by performing the Breusch and Pagan
Lagrangian multiplier test.It aims to find out whether the use of the PLS regression model is still effective in producing an estimation model.[24] , the estimation method General Least Square ("GLS") can produce estimates that meet Best Linear Unbiased Estimators (BLUE) even though the panel data used is not fulfilled the assumption of homoscedasticity and there is autocorrelation.Method GLS is capable maintain a nature estimator which no bias and is consistent as well as capable overcome heteroscedasticity problems .GLS method was able to maintain the nature of the estimator that is not biased and consistent and able to overcome the problem of heteroscedasticity (Aditya R et al., 2019).

Test Hypothesis
Regression results with the Random Effect GLS Regression method for the three models can be seen in Table 8, Table 9, and Table 10.Table 11 shows the results of the regression model I used to measure the influence of pandemic covid-19 to cash holding.Mark R2 _ as big as 0.3243 shows that the independent variables together affect variable dependent as big as 32.43%.Mark P>|z| show mark significance variable independent to variable independent.If P>|z| <0.05 so the independent variable is significant to the dependent variable at the level of significance of 5% if P>|z| <0.01 then the independent variable is significant to the variable dependent on the level of significance 1%, and if P>|z| <0.001 so variable independent significant to variable dependent on level significance 0.1%.
Influence pandemic covid-19 to cash holding showed with variable treated*period which is result dummy variable interaction Among variable dummy treated and variable dummy period.
Coefficient variable treated*period show mark negative where Thing the show that pandemic covid-19 own influence negative to cash holding.
Mark P>|z| variable treated*period is 0.000 or significant to the cash holding variable on level significance 0.1%.Thus hypothesis 1 which state that "Pandemic Covid-19 impact negative significant to cash holding company" was accepted.Variable cashholding_dummy on table 9 shows the influence of cash holding company on earnings management.Coefficient variable cashholding_dummy is negative which indicates that the cash condition holding company influential negative to practice earnings management, or with saying other the more tall cash holding company so management company will trigger management company for To do earnings minimization.P value>|z| variable cashholding_dummy of 0.000 which means significant to the earning management variable at a significance level of 0.1%.Thus, hypothesis 3a which states that " cash holding has an effect " significant negative on the practice of earning management (earning minimizing)" is accepted.Interaction Among variables treated, period and cashholding_dummy show the role of the cash holding variable as a variable that moderates the negative effect of pandemic covid-19 on earnings management.
Coefficient variable positive value so that debilitating impact negative pandemic covid-19 to earnings management.However thus P value>|z| shows the number 0.344 which is much larger than 0.05.The thing this shows is that the role of variable cash holding as variable moderation is not significant.Thus hypothesis 3b which states that " cash " holding will weaken the impact of negative covid-19 to practice earnings management (earnings minimize)" was rejected.Variable desc_acc_dummy shows the influence of earnings management on stock returns.The coefficient of the variable desc_acc_dummy is positive which is show that earnings management is done by the management company influential positive to return share, with sayi n g other If the company does earn maximization, the stock return will be higher than it should be and if the company does earn minimization then stock returns will be lower than they should be.But the value of P>|z| shows the number 0.109 which means that the variable has no significant effect on the dependent variable.So it can be said that hypothesis 5a which states that " ️ Earning management influential positive significant to return share" was rejected.
Variabletreated*period*dess_acc_dummy shows the role variable earning management as a variable that moderates the negative impact of the pandemic covid-19 on stock returns.Variable coefficient worth positive so that character strengthens the negative impact of the covid-19 pandemic on stock returns.However, the value of P>|z| shows the number 0.954 which indicates that the variable does not affect all stock returns.P value>|z| which exceeds the level of trust of 95% indicates that the existence variable does not affect the dependent variable.Thus hypothesis 5b which state that " ️ Earning management strengthen impact pandemic covid-19 against return shares" was rejected.From the discussion test hypothesis in on could summarize in table 14 as follows.

Discussion
The hypothesis test show that cash holding significant effect on earnings management, so the role of the variable cash holding as variable moderation is worthy to be measured.Results test the hypothesis this is in accordance with a study done by Acharya et.all (2012) [25] which state that motivation management company to maintain a position cash holding triggers the existence practice alignment profit.

Results variable hypothesis test cash holding as variable influence moderation
The negative impact of the covid-19 pandemic shows an insignificant effect.
Although cash holding owns a significant effect on earnings management, no proven difference in significant earnings management before and after pandemic covid-19 between companies affected and not affected by COVID-19.

Influence of Pandemic Covid-19 on Price Share
Testing the hypothesis that the COVID-19 pandemic harms return share produces a coefficient negative and with a mark P>|z| as big as 0.006 or significant on level trust 99%.Results findings could be interpreted as the following : a. Company experience drop return share consequence from existence pandemic covid-19, b.The company which affected covid-19 experience a drop in return share more tall compared company which not affected covid-19 c.The more tall impact covid-19 to something company, the more big drop returns the stock.These findings are in accordance with research conducted by Mafrolla et.all (2017) [12] which shows that there are abnormal stock returns in 21 countries with market trading which is the most in the world.Mafrolla et.all (2017) [12] also states that the decline in stock returns in Asian stock markets is more rapid compared to other stock markets in Europe.This decline in stock returns caused appearance risk new which is not yet could estimated how critical impact on the economy of a country.This risk triggers the investors interested in the investment with the objective for secure the fund invested.

Influence Pandemic Covid-19 to Price share with Variable Moderation Earnings Management
Results test hypothesis show that earnings management no own influence which significant to return share.Results confirmed on testing hypothesis next that variable earnings management no could Become variable moderation which strengthens or weaken the impact of negative pandemic covid-19 to return share.
Observing the previous hypothesis testing regarding the impact of covid-19 on earning management which is not significant, these findings are consistent.Practice earnings management which done by management company no experience change significant with existence pandemic covid-19, so that no capable influence evaluation investors to price share.Change economy macro which significant also no make management the company performs earnings management sporadically.

CONCLUSION
The objective of the study is to know the impact of pandemic covid-19 on cash holding, earning management, and stock returns, and to test whether cash holding could moderate the impact covid-19 on earnings management and whether earnings management can moderate the impact of covid-19 on return share.Results study this show that pandemic covid-19 own a significant negative impact on the company's cash holding, a negative impact with no significance to earnings management, and an impact negative significant to stock returns.Meanwhile, the variable cash holding cannot be a variable that moderates the impact of pandemic covid-19 on earnings management, and earnings management variables are also not able to weaken or strengthen the impact of the COVID-19 pandemic on stock returns.Earnings Management cannot strengthen or weaken the impact of the pandemic covid-19 to return share especially caused application standard accountancy which required to companies which recorded in board recording main Exchange Effect Indonesia.summary of the main topics covered or a re-statement of your research problem, but a synthesis of key points and, if applicable, where you recommend new areas for future research.
based on Model Jones which has modified : DACC it = (TACC it /Ta it-1 ) -(NDACC it /Ta it-1 ) description : DACC it = Discretionary Accruals company i on year t TACC it = Total Accrual company i on year t NDACC it = Non-Discretionary Accruals company i on year t ta it-1 = Total asset company i on year t-1 TACC it net profit -cash flow from operation NDACC NDACC it /Ta it-1 = 1 _ (1/Ta it-1 ) + 2 _ (( REV it -REC it ) /Ta it -1 ) + 3 _ (PPE it /Ta it-1 ) description : REV it = _ change income company I on year t REC it

Table 1 .
Data and information about level cash holding, mark discretionary accruals, and stock returns of companies

Table 2 .
Calculation of earning management while testing hypothesis 2

Table 3 .
Calculation the impact of covid-19 on Stock Returns Code variable for control sample individual in the data panel YEAR variable for the control period in the data panel For test hypotheses 3a and 3b Used model 3 in table 3 as following : Discretionary Accruals it = 0 + 1 Cashholdings it * Treated it *Period it + 2 Treated it * (last price t -last price t-1)/last price t-1 Treated dummy variable indicating degree impact of covid-19, a score of 1 for high and very impacted companies tall, and currently and 0 for a A data study is a data panel which is merging between cross-section data and time-series data or called panel data.Calculation of difference in difference value using multiple linear regression analysis with software STATA 14.2.

Table 4 .
Descriptive Statistics Model I Depending Variable : Cash Holding

Table 6 .
Descriptive Statistics Model III Depending Variable : Return Share

Table 7 .
Result of Unit Root Test I

Table 7
above states that at the levin-linchu level there are several variables that are not stationary, so it is necessary to look at the variables at the level of the Hadri LM test.The results show that some panels contain unit roots and all panels do not move at the presence level of the LM test under various conditions.

Table 8 .
Result of Unit Root Test II

Table 9 .
Result of Unit Root Test III

Table 10 .
Breusch and Pagan Lagrangian Multiplier Test Breusch and Pagan LM .test it can be concluded that the REM method with the General Least Square technique is more effective in predicting the model.According to Al Saedi & Alaa (2018)

Table 11 .
Results Of The Regression Model I

Table 12 .
The Results Of Regression Model II

Table 13 .
The Results Of Regression Model III

Table 13
shows the results of regression model III which was used for measuring the effect of the covid-19 pandemic on stock returns.R 2 value of 0.2433 shows that variables independent by together same affect the dependent variable by 24.33%.The impact of the covid-19 pandemic to return share showed with the variable treated*period which is a dummy variable resulting from the interaction between the treated dummy variables and variable dummy period.Coefficient variable treated*period show mark negative where Thing the show that pandemic covid-19 own influence negative to return share.Mark P>|z| variable treated*period is 0.006 which means that it is significant to the stock return variable at the level of 1% significance.Thus, hypothesis 4 which states that the "Pandemic" covid-19 influential negative significant to return stock "is accepted.