International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

SYSTEMIC RISK ANALYSIS BY USING CONDITIONAL VALUE-AT-RISK MODEL FOR SOME ISLAMIC STOCKS IN INDONESIA Sukono *, Sudradjat Supian, Dwi Susanti, Tisa Annisa Mathematics and Natural Sciences Faculty Padjadjaran University Indonesia *Email : [email protected] ABSTRACT In this paper we analyzed about systemic risk by using a model of Conditional Value-at-Risk. Here the analysis is the relationship between an individual's risk of Islamic stocks with systemic risk. Individuals risk is analyzed using the Valueat-Risk model. While the systemic risk was analyzed using the Conditional Value-at-Risk model. The relationship between individuals risk with systemic risk was analyzed using quantile regression model approach. The data analyzed include the price of some Islamic stocks traded on the Jakarta Islamic Index. The analysis showed that some Islamic stock price, the individual risk have a strong relationship with the systemic risk. In addition, there are some stock prices the individual risk is not strongly relationship with the systemic risk.

Keywords: individual risk, systemic risk, value-at-risk, conditional value-at-risk, quantile regression other stocks, and even affect the systemic risk for the financial system. Systemic risk can occur due to interlink and interdependence within an economic system. Generally, systemic risk triggered by the failure occurred due to a financial institution or a financial market or other reasons (Aulianisa, 2013). Systemic risk is very important to do predictions and analyzed. It is necessary to anticipate the possibility of the occurrence of adverse events that can result in losses for investors. According to Arias et al. (2010) and Bjarnadottir (2012), Systemic risk can be predicted by using the Conditional Value-at-Risk (CoVaR). Arias and Bjarnadottir concluded that CoVaR is a systemic risk indicator that is easy to interpret, does not demand complicated data set and can be used with other risk indicators and stress tests that together help get a better understanding of the risks threatening the stability of the financial system. According to Adrian and Brunnermeier (2011), Systemic risk prediction using CoVaR method is based on Value-at-Risk (VaR) of a financial system and the VaR of an individual stock. Adrian and Brunnermeier do CoVaR estimation using Quantile Regression Model (QRM). Based Arias et al. (2010) and Bjarnadottir (2012), as well as Adrian and Brunnermeier (2011), this

INTRODUCTION Investments in financial assets in the stock market lately become very popular among investors. Because the capital market, investors can freely to obtain greater profits. There are two investment options in the equity markets, the market share of conventional and Islamic stock market (Aulianisa, 2013). The last few year Islamic stocks growing rapidly, it is seen from the development of the Jakarta Islamic Index (JII) which began to increase by 58.38% in 2007 compared to previous years. As reported the Indonesia Stock Exchange in 2007, that the percentage increase in the JII higher than LQ45 and JCI were only 52.58% and 52.08% at the time (Indonesia Stock Exchange, 2010). When making investments, investors must expect to benefit as much as possible. On the other hand, investors generally invest with risk-averse. However, each investment must have a risk that can not be avoided, it can only be minimized. So far, investors tend to only pay attention to the value of risk by an individual stock, but there are other risks that affect the overall risk of M - 173

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

2) D Negative ( D   sup x [ F ( x)  Fn ( x)] ) , a reduction that produces the largest negative number.

paper analyzed the systemic risk in some Islamic stocks selected, using CoVaR. The purpose of this analysis is to predict the systemic risk that is conditioned on the Valueat-Risk of an Islamic stock. So the results can be used as consideration for investors who will invest in stocks analyzed.

3) D Absolute ( D  max{D  , D  } ) , is the largest number among the absolute values D+ and D–. Kolmogorov-Smirnov Z is the result of the square root of the number of samples N and the largest absolute difference between the empirical cdf and the theoretical cdf (Arnold & Emerson, 2011; Sigurd, 2011), is almost equal to the square root of the sample size N multiplied by D absolute:

METHODS In broad outline, systemic risk analysis methodology here is done step by step which includes: determining stock returns and Islamic stock index, stock return normality test and Islamic stock index, the estimated quantile regression models, the calculation of Value-at-Risk (VaR), Estimation risk model systemic using Conditional Value-at-Risk (CoVaR), and model validation test. Begins with the determination the stocks and Islamic stock index returns as follows.

(2.b) Z  N .D Absolute . According to Arnold & Emerson (2011) and Sigurd (2011), "Kolmogorov-Smirnov Z" is D absolute is converted into a standardized Z score. The meaning of the standardized score is the value of Z in the standard normal distribution. That is, the way the test is similar to test the value of D, only this time under a normal distribution with the help of the standard normal distribution table, where: H 0 ditolak jika Z -count (Kolmogorov-Semirnov) greater than Z -table at the level of significance  . The basic concept of the KS normality test is to compare the distribution of the data (which will be tested normality) by the standard normal distribution. Standard normal distribution is the data that has been transformed into the form Z-Score and assumed to be normal. So in fact the KS test is a test of difference between data normality was tested by standard normal of data (Arnold & Emerson, 2011; Sigurd, 2011). Hypothesis on the KS One Sample is as follows: H 0 : There is no difference between the tested data by a normal distribution. H 1 : There are differences between the tested data by a normal distribution. Where: If P  Value >  , then H 0 accepted, and if

Determination of Return Suppose Pt Islamic stock price at the time t ( t  1,...,T and T the number of data observations), and rt Islamic stock returns at time t . The amount of Islamic stock returns can be determined by the equation: (1) rt  ln Pt  ln Pt 1 . The return data rt then performed test of normality as follows (Dowd, 2002). Normality Test Normality test here is done by KolmogorovSmirnov approach (KS), which is a statistical test of the most fundamental and most widely used. KS test was first introduced by Andrey Nikolaevich Kolmogorov in 1933 and then tabulated by Nikolai Vasilyevich Smirnov in 1948. According to Arnold & Emerson (2011) and Sigurd (2011), According to Arnold & Emerson (2011) and Sigurd (2011), the KS test is used for the one-sample test that allows the comparison of a frequency distribution by some standard distributions, such as Gaussian normal distribution. KS test measures the proximity between F (x)

P  Value <  , then H 0 rejected. P  Value is the statistical probability of KS and  level of significance.

Estimation of Quantile Regression Models (QRM) Quantile Regression Models is a statistical technique used to estimate the relationship between the response variable by explanatory variables on a particular conditional quantile function (Peng, 2010). As on least squares method, which minimizes the sum of squared errors and suspect the model using the conditional mean function, quantile regression is minimizes the absolute weighted error is not symmetrical and suspect quantile function of a distribution conditional on the data. In general, quantile regression is very helpful when you want to analyze a specific part of a conditional distribution. For example, top quintile, median and bottom quantile of the conditional distribution. Quantile regression method does not require parametric

dengan Fn (x) when n assumed to be a tremendous value, Sigurd (2011) defines the cumulative distribution function (cdf) is as follows: assumed to be a tremendous value, Sigurd (2011) defines the cumulative distribution function (cdf) is as follows: D  sup x | Fn ( x)  F ( x) | , (2.a) where is the supremum of the number of distance D. The statistical value D (Most Extreme Differences) the KS test consists of: 1) D Positive ( D   sup x [ Fn ( x)  F ( x)] ), a reduction that produces the largest positive number. M - 174

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

assumptions (Buhai, 2005). For random variables X by probability distribution function: (2) F ( x)  Pr( X  x) . Quantile p of X defined as the inverse function: (3) Q( p)  inf{ x, F ( x))  p} . Following Koenker et al (1978), quantile regression equation is as follows: ( p)

( p)

( p)

yi   0  1 xi   i , with 0  p  1 expressed quantile p, and: ( p)

( p)

Q ( p) ( yi | xi )   0  1 the conditional quantile function.

X ti   i   i M t 1   ti .

xi , 0  p  1 ,

(9)

In this study, i showed Islamic stock while j showed the Jakarta Islamic Index (JII). Based on Adrian and Brunnermeier (2011), quantile regression equation used is as follows:

(4)

(10)

Where X ti return of Islamic stock i in period t,  i and

(5)

 i quantile regression parameters, and M t 1 vector of

VaR p  F 1 (1  p) ,

(6)

Pr( X  VaR p )  1  p ,

(7)

Pr( X  VaR p )  p .

(8)

Equation (8) can be written as 1  F ( x)  p

JII returns lag before the period t. Return of JII is expressed in the equation:

X tJII   JII   JII M t 1   tJII ,

(11)

and

X tJII   JII |i   JII |i M t 1   JII |i X ti   tJII |i .

(12)

Where X tJII return of JII at the period t, X ti return of Islamic stock i at the period t,  JII and  JII quantile regression parameters, and M t 1 vector of JII returns lag before the period t. Coefficient values of  i ,  i ,  JII ,  JII obtained from equation (12) and (13), then the values can be used to predict the Value-at-Risk (VaR), which is expressed in the following equation: , (13) VaR i ( p)  ˆ i  ˆ i M

and

t 1

t

where F is the cumulative distribution

and

VaRtJII ( p)  ˆ JII  ˆ JII M t 1 .

function (cdf) of the total loss, and F 1 is the inverse of the function which is called quantile function (Dowd, 2002).

(14)

Where VaRti ( p) is VaR of Islamic stock i on quantile p at the period t, and VaRtJII ( p) is VaR of JII on quantile p at

Estimation of Systemic Risk Model Adrian and Brunnermeier (2011) propose a new method in the measurement of systemic risk: Conditional Value-at-Risk (CoVaR). This method is based on Valueat-Risk (VaR) of a system, and Value-at-Risk (VaR) of an individual stock (Gaglianone et al., 2009). In this study was calculated the systemic risk ( CoVaR ) of some Islamic stocks listed in the Jakarta Islamic Index (JII).

the period t. The estimators of ˆ i , ˆ i , ˆ JII , and ˆ JII can be calculate using equations:

ˆ i 

n M t 1 X ti  ( M t 1)(  X ti ) n M t21  ( M t 1) 2

,

ˆ i  X ti  ˆ i M t 1 ,

(15) (16)

and

Systemic Risk Method Using Conditional Value-at-Risk (CoVaR) Calculation of systemic risk in this paper uses CoVaR method, which was introduced by Adrian and Brunnermeier (2011) were estimated using Quantile Regression Models (QRM). Definition 1. (Adrian and Brunnermeier, 2011). j |i

is implicitly defined by the p-

j |C ( X i ) Pr{X j  CoVaR p | C ( X i )} .

or

CoVaR p

CoVaR p

quantile of the conditional probability distribution.

Value at Risk (VaR) Model Value-at-Risk (VaR) is a measure of the level of risk of investment loss. Value-at-Risk is generally defined as the maximum possible loss for a particular position within a known confidence level to a particular time horizon (Khindanova & Rachev, 2005; Holto, 2002; Manganelli and Engle, 2001). Value-at-Risk (VaR) of the random variable x value of loss on confidence level (1  p) , 0  p  1 , is (1  p) quantile of the distribution of the total loss F

x  VaR p

j |i

institution i.

ˆ JII 

n M t 1 X tJII  ( M t 1 )(  X tJII ) n M t21  ( M t 1 ) 2

ˆ JII  X tJII  ˆ JII M t 1 .

, (17) (18)

Where n number of data, X ti average return of Islamic stock i at the period t, X tJII average return of JII at the

is defined as the VaR of institution j

period t, and M t 1 average return lag of JII, before the period t.

i

(financial system) which relies on a state C ( X ) of an M - 175

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

Furthermore, to calculate the Value-at-Risk (VaR) of Islamic stocks, Value at Risk (VaR) of Jakarta Islamic Index (JII), and Conditional Value-at-Risk (CoVaR) of Islamic stocks, according to the method used Adrian and Brunnermeier (2011), the authors use quantile regression by confidence intervals on the first quantile 99%. Calculation of Conditional Value-at-Risk JII |i ,

(23) CoVaRi  CoVaRi  VaR i . Based on equation (16), systemic risk calculation is as follows:

CoVaRti ( p)  CoVaRti ( p)  VaRti ( p) .

Where CoVaRti ( p) : Systemic risk of Islamic stock i on quantile p at the period t, CoVaRti ( p) : CoVaR of Islamic

(CoVaR) using the coefficients of    obtained from equation (13), is as follows (Cao, 2013): JII |i ,

JII |i

stock i on quantile p at the period t, and VaRtJII ( p) : VaR of JII returns on quantile p at the period t.

CoVaRti ( p)  ˆ JII|i  ˆ JII|i M t 1  ˆ JII|iVaRti ( p) . (19)

Model Validation Model validation is the process of checking the feasibility of a model. Model validation performed to ascertain whether or not a model worthy to be used. Validation of the model here can be done by Likelihood Ratio (LR) Test. The step of the model validation by using Likelihood Ratio (LR) Test is as follows: 1) Determine the total number of samples or observations (T). 2) Calculate the value of V (total failure or a total exception) during the period of observation. 3) Determine the value of  or the level of confidence. 4) Calculate the value of log Likelihood Ratio (LR) by following equation:

Where CoVaRti ( p) : CoVaR of Islamic stock i on quantile p at the period t, and VaRti ( p) : VaR of Islamic stock i on quantile p at the period t. When we assume that

Y  {n M t21  ( M t 1) 2 }{n ( X ti ) 2 ( X ti ) 2 }  n{ M t 1 X ti  ( M t 1)(  X ti )}2 , then

ˆ JII |i 

{n M t 1 X tJII  ( M t 1 )(  X tJII )} Y {n ( X ti ) 2 ( X ti ) 2 }



{n M t 1 X ti  M t 1  X ti } Y

{n X ti X tJII  ( X ti )(  X tJII )}

ˆ JII |i 

  V LR  2 ln[ T V (1   )V ]  2 ln 1     T 

(20)

{n X ti X tJII  ( X ti )(  X tJII )} Y

ˆ JII |i



{n M t 1 X tJII  ( M t 1 )(  X tJII )} Y

{n M t 1 X ti  ( M t 1)(  X ti )}

(21)

 X tJII  ˆ JII |i M t 1  ˆ JII |i X ti .

(22)

  

T V

V  V       ,  T   

(25)

With LR Log Likelihood Ratio,  confidence level, T is number of observed data, and V is number of data errors. Value of log Likelihood Ratio (LR) compared to the critical value of chi squared statistic by degrees of freedom 1 (number of independent variables involved), on level of significance was set. If the LR values greater than the critical value of chi-squared statistics, the risk calculation models are not accurate, and vice versa if the value of LR is smaller of the critical value of chi-squared statistics, the risk calculation model is accurate (Peng, 2010).

 {n (M t 1 ) 2 }2 

(24)

Where n is number of data, M t 1 return lag of JII before periode t, X tJII return of JII at the period t, and

RESULTS AND DISCUSSION Data and Materials The data used in this study is a secondary data obtained of the http://www.duniainvestasi.com accessed on January 22, 2014. Observed data for the calculation of systemic risk in this study is the closing stock price during the 369 days (January 4, 2012 till June 30, 2013), which lasted for five working days in a week. Data includes ASII, AALI, ANTM, INTP, PTBA, SMGR, TLKM, and UNVR. Systemic risk assessment carried out using the method of Conditional Value-at-Risk (CoVaR) which was introduced by Adrian and Brunnermeier (2011). Data processing is performed by Microsoft Excel software, STATA 10, and SPSS 19.

X ti return of Islamic stock i at the period t. Systemic risk assessment carried out after obtaining the value at risk of Islamic stocks ( VaRti ( p) ), Value-at-Risk of JII ( VaRtJII ( p) ), and Conditional Value-at-Risk of Islamic stocks ( CoVaRtJII ( p) ). Adrian and Brunnermeier (2011) and Hardle et al. (2012) define the CoVaRti ( p) as follows: Definition 2. (Adrian and Brunnermeier, 2011).

CoVaR i represent the difference between the VaR of the financial system (j) which depend on the state in an institution i, and VaR of the financial system (j). M - 176

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

estimating of VaR, the first needs to be estimated coefficient parameters ˆ i and ˆ i follow equations (15) and (16). As for the estimation of VaR for JII return were done according to equation (14). Similarly, the first parameters previously estimated coefficients need of

Return Calculation and Normality Test Calculation of stock returns (performed using equation (1), and then returns the data to the normality test. Normality test is intended to ensure that the return data are all in the interval  i   i with the  i mean and

 i standard deviation of stock returns i . This means that

ˆ JII and ˆ JII follow equations (17) and (18). Once the

there is no incidence of extreme return data, which can interfere with the analysis of systemic risk. Normality test was performed by using the method of KolmogorovSmirnov One Sample, by equation (2). The results are given in Table-1 below.

parameter

estimator

coefficients

ˆ i and ˆ i ,

and

coefficient parameters ˆ JII and ˆ JII , then used to estimate the VaR of each Islamic stock returns, and also VaR of JII. Value-at-Risk (VaR) estimation is done by substituting the return data into equation (13) for each VaR of Islamic stocks, and into the equation (14) for the VaR of JII. Value-at-Risk (VaR) estimation results of Islamic stocks and VaR of JII are given in Table-3 as follows. Table-3 Values of ˆ , ˆ , VaR of Islamic Stocks and JII

Having ascertained return data are normally distributed, next the return data used to estimate quantile regression models as follows. Parameter Estimation of quantile Regression Models Estimation of the parameters of quantile regression model (QRM) for each Islamic stock is done according to equation (11), while the JII refer to equation (13). Estimation performed using Likelihood method, by using statistical software of STATA 10. The results of

No

Stocks

1 2 3 4 5 6 7 8 9

ASII AALI ANTM INTP PTBA SMGR TLKM UNVR JII

ˆ

ˆ

VaR 0.000573 0.001347 0.001322 0.000635 0.000622 0.000654 0.000978 0.000705 0.000650

parameters estimation  i and  i for each Islamic stock

Estimation of Conditional Value-at-Risk In this section intends to estimate the Conditional Value-at-Risk (CoVaR). However, it must be

are given in Table-2 as follows.

done first before parameter estimation of  JII|i ,  JII|i ,

Table-2. Parameters  i and  i Islamic Stocks No

Nama

1 2 3 4 5 6 7 8 9

ASII AALI ANTM INTP PTBA SMGR TLKM UNVR JII



i



and  JII|i using equation (12). Variables involved,

i

namely the return JII ( X JII ), vector lag return of JII ( M t 1 ), and Islamic stock returns i ( X i ). Estimation is done using the statistial software of STATA 10. The estimation results are given in Table-4 below.

Estimation of Value-at-Risk Islamic Stocks and JII Value-at-Risk (VaR) Estimation of Islamic stocks were done according to equation (12). Before M - 177

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

Table-4. Values of  No

Stocks

1

ASII

2

AALI

3

ANTM

4

INTP

5

PTBA

6

SMGR

7

TLKM

8

UNVR



JII|i

JII|i ,





JII |i , and

JII |i



Observing the Table-6, it appears that the value of CoVaR for each Islamic stocks relatively similar on first quantile. Whereas, if the observed value of R-square of each quantile regression equation is relatively large. This means that the optimal quantile regression equation on the first quintile is by 99% confidence interval.

JII |i

 JII |i

Systemic Risk Estimation of Islamic Stocks This section is intended to be estimates of systemic risk on Islamic stocks. Estimation is done by referring to equation (23). Systemic risk estimation results of the Islamic stocks given in Table-7 as follows. After getting the value of VaR, CoVaR, and systemic risk ( CoVaR ) of each Islamic stocks, then performed an analysis of these value of CoVaR . Analysis was performed to determine whether there is a strong relationship between the VaR by systemic risk CoVaR .The following comparison of VaR, CoVaR, and systemic risk CoVaR of each Islamic stock.

After estimating the parameters of  JII|i ,

 JII|i , and  JII|i , the next determined parameter estimator ˆ JII|i , ˆ JII|i , and ˆ JII|i using equation (20), (21) and (22). The results are given in Table-5 as follows.

Table-7. Values of VaR, CoVaR, and Systemic Risk ( CoVaR ) Systemic Risk No Stocks VaR CoVaR ( CoVaR )

Table-5. Estimator Values of ˆ JII|i , ˆ JII|i , and ˆ JII|i No

Stocks

1 2 3 4 5 6 7 8

ASII AALI ANTM INTP PTBA SMGR TLKM UNVR

ˆ JII|i

ˆ JII|i

ˆ JII|i

1 2 3 4 5 6 7 8

ASII AALI ANTM INTP PTBA SMGR TLKM UNVR

0.000573 0.001347 0.001322 0.000635 0.000622 0.000654 0.000978 0.000705

0.037438929 0.037438481 0.037438248 0.037438309 0.037438609 0.374381500 0.037435785 0.037438660

0.030938929 0.030938481 0.030938248 0.030938309 0.030938609 0.030938150 0.030935785 0.030938660

Based on the values of the parameter estimator of ˆ JII|i , ˆ JII|i , and ˆ JII|i , CoVaR estimation is done

Based on Table 7, it can be seen that the order of the Islamic stocks by the smallest to largest value of VaR, ie ASII, PTBA, INTP, SMGR, UNVR, TLKM, ANTM, and AALI. While the order of Islamic stock with a value of systemic risk smallest to largest, ie TLKM, SMGR, ANTM, INTP, AALI, PTBA, UNVR, and ASII. The results in Table 7, show that the VaR each Islamic stocks by systemic risk having a weak relationship. Stocks that have a high VaR is not necessarily a major impact on the risk in the system (not necessarily a systemic effect). Basically, the VaR each stocks is measured how much the maximum loss of each stock due to factors such as: the exchange rate, market price movements, interest rate. Systemic risk focused on how large a loss in the system caused a loss of individual stocks. This means, the systemic risk each stocks is influenced Conditional Value-at-Risk (CoVaR) than Value-at-Risk (VaR) each Islamic stocks.

by following equation (19). Values estimators of CoVaR every Islamic stocks are calculated by substituting the value of the lag returns of JII ( M t 1 ) and VaRti into equation (19). CoVaR estimator calculations performed using Microsoft Excel software assistance. While the pseudo R-Square value obtained by using the software of STATA 10. The results are given in Table-6 as follows. Table-6 Values of CoVaR for Islamic Stock Returns No Stocks CoVaR R-Square 1 2 3 4 5 6 7 8

ASII AALI ANTM INTP PTBA SMGR TLKM UNVR

0.037438929 0.037438481 0.037438248 0.037438309 0.037438609 0.037438150 0.037435785 0.037438660

0.9935 0.9935 1.0000 0.9935 0.9935 0.9935 0.9935 0.9935

M - 178

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

CONCLUSION

Model Validation Having obtained systemic risk model for each Islamic stocks then performed a validation test of the models produced. Validation is done to test whether systemic risk calculation results using by estimation method CoVaR by Quantile Regression Models (QRM) is valid or not. Steps are: i) Comparing the systemic risks that have been obtained by actual return. In this paper use stock data of January 4, 2012 till June 30, 2013, by a total observation (T) data for each as many as 369 Islamic stocks as a comparison, provided that if the value of systemic risk is greater than the actual returns, then it is counted as failure or error. ii) Calculating LR Test according to the equation (25) by confidence level in this paper is equal to 99%. The hypothesis used is as follows: H 0 : The model is suitable for systemic risk assessment H1 : The model is not suitable for calculating the systemic risk If the value of Likelihood Ratio (LR) is smaller of the critical value then accept the hypothesis H 0

In this paper has analyzed the systemic risk for some Islamic stock using the Conditional Value-at-Risk (CoVaR) approach. Based on the analysis it can be concluded that the Value-at-Risk (VaR) of ASII, AALI, ANTM, INTP, PTBA, SMGR, TLKM, and UNVR respectively are 0,000573, 0,001347, 0,001322, 0.000635, 0.000622, 0.000654, 0.000978, and 0.000705. Value of CoVaR for the stocks of ASII, AALI, ANTM, INTP, PTBA, SMGR, TLKM, and UNVR respectively are 0.037438929, 0.037438481, 0.037438248, 0.037438309, 0.037438609, 0.037438150, 0.037435785 and 0.037438660. The value of systemic risk for stocks of ASII, AALI, ANTM, INTP, PTBA, SMGR, TLKM, and UNVR respectively are 0.030938929, 0.030938481, 0.0309389248, 0.030938309, 0.030938309, 0.030938609, 0.030938150, 0.030935785 and 0.030938660. The order of the Islamic stocks with the smallest to largest value of VaR are ASII, PTBA, INTP, SMGR, UNVR, TLKM, ANTM, and AALI. While the order of Islamic stock with a value of systemic risk CoVaR smallest to largest are TLKM, SMGR, ANTM, INTP, AALI, PTBA, UNVR, and ASII. Islamic stocks with the largest value of VaR is AALI, while the value of systemic risk biggest is ASII. For VaR large value does not necessarily result in greater systemic risk. It can be said also that the VaR with systemic risk did not show a strong relationship.

which states that the model is valid. Conversely, if the Likelihood Ratio (LR) is greater of the critical value, then the hypothesis H 0 rejected, and the prevailing hypothesis H1 which states that the model is not valid. The results of the validation test calculations are given in Table-8 as follows.

Acknowledgments: Acknowledgments submitted to the Institute of Research and Community Service, Padjadjaran University, which has given the facility to conduct this research.

REFERENCES Adrian, T. & Brunnermeier, M.K. 2009. CoVaR. Working Paper. Federal Reserve Bank of New York, Capital Markets, 33 Liberty Street, New York, NY 10045, http://nyfedeconomists.org/ adrian, email: [email protected].

Based on Table-8, of the entire Islamic stocks showed that accepted, meaning Quantile Regression Models (QRM) suitable for systemic risk assessment. Calculation of systemic risk in the first quintile with 99% confidence interval is valid.

Aulianisa, F. 2013. Pengaruh Faktor Fundamental dan Risiko Sistemik Terhadap Harga Saham Syariah di Pasar Modal Syariah (Studi Empiris JII di BEI Tahun 2007-2010). Jurnal Ekonomi Islam La_Riba, Volume VII, No. 1, Juli 2013. Arias, M., Mendoza, J.C. & Perez-Reyna, D. 2010. Applying CoV aR to Measure Systemic Market Risk: the Colombian Case. Working Paper. the M - 179

International Conference on Mathematics, Science, and Education 2014 (ICMSE 2014)

Financial Stability Department at the Banco de la Republica (Central Bank of Colombia). Bjarnadottir, F. 2012. Implementation of CoVaR, A Measure for Systemic Risk. Working Paper. KTH Royal Institute of Technology, Stockholm. Buhai, I.S. 2004. Quantile Regression: Overview and Selected Applications. Roger Konker’s Lecture Notes, from the recent Netherlands Network o Economics Workshop in Groningen, December 2003. Cao, Z. 2013. Multi-CoVaR and Shapley value: A Systemic Risk Measure. Working Paper. Banque de France DSF-SMF. E-mail: [email protected]. Dowd, K., 2002, An Introduction to Market Risk Measurement, John Wiley & Sons, Inc., New Delhi, India. Härdle, W.K., Wang, W., Zhu, L. & Yu, L. 2012. CoVaR with Very High Dimensional Risk Factors. Working Paper. Center for Applied Statistics and Economics, Humboldt–Universität zu Berlin. Holto, G.A., 2002, History of Value-at-Risk, Working Paper, Contingency Analysis, P.O. Box 961, Boston, USA. http://www.contingencyanalysis. com, diakses 10 Februari 2014. Khindanova, I.N. & Rachev, S.T., 2005, Value at Risk: Recent Advances, Working Paper, University of California, Santa Barbara and University of Karlsruhe, Germany. http://www.econ.ucsb.edu/ papers/wp3-00.pdf. (Downloaded in November 25, 2013). Peng, H. 2010. Introduction to Quantile Regression. Indianapolis: Department of Mathematical Sciences, Indiana University-Purdu University. Indonesia Stock Exchange. 2010. Buku Panduan Indeks Harga Saham Bursa Efek Indonesia. Jakarta: Indonesi Stock Exchangee Building. Bencana. Laporan Penelitian.

M - 180