ExtremeVaR of South East Asian Stock Indices with Extreme Distribution- Based Efficiency

Transcription

1 Available online at ScienceDirect Procedia Econoics and Finance 5 ( 2013 ) International Conference on Applied Econoics (ICOAE) 2013 EtreeVaR of South East Asian Stock Indices with Etree Distribution- Based Efficiency Chaitip Prasert a, Chaiboonsri Chukiat b, Boonsue Sarun * a Associate Professor, Faculty of Econoics, Chiang Mai University, Chiang Mai, Thailand b Lecturer, Faculty of Econoics, Chiang Mai University, Chiang Mai, Thailand. * Research Assistant, Touris Econoic Research Centre, Faculty of Econoics, Chiang Mai University, Chiang Mai, Thailand. Abstract The paper eaines etreevalue-at-risk (etreevar) odel using edian with daily stock indices of selected South East Asian countries consisting of SET inde (Thailand), KLSE inde (Malaysia), FTSI inde (Singapore), and JKSE inde (Indonesia). Additionally, the eperient using etree value theory (EVT) was tested by Generalized Pareto Distribution (GPD) which has the characteristics of cuulative distribution function (CDF), or just distribution function, describes the probability that a subjective-valued rando. The output results indicated that using edian of KLSE etreevar in Malaysia was the AEC efficient equity for investing in these arkets The The Authors. Published Published by Elsevier by Elsevier B.V. Open B.V. access under CC BY-NC-ND license. Selection and/or peer-review under under responsibility responsibility of the Organising of the Organising Coittee Coittee of ICOAE 2013 of ICOAE Keyword: Second-order stochastic doinance (SSD); etreevar; AEC arket indices; Generalized Pareto Distribution 1. Introduction The structure of AEC arket indices is changing rapidly. AEC stock arkets have becoes bigger in nuber, larger in size and ore specialized. The research objective ais to evaluate the etreevar based The Authors. Published by Elsevier B.V. Open access under CC BY-NC-ND license. Selection and/or peer-review under responsibility of the Organising Coittee of ICOAE 2013 doi: /s (13)

2 Chaitip Prasert et al. / Procedia Econoics and Finance 5 ( 2013 ) on Generalized Pareto Distribution (GPD) of tie-series daily returns on arket indices by using edian of in SET - Bangkok SET (Thailand), KLSE - Kuala Lupur Coposite Inde (Malaysia), FTSI - Singapore Straits Industrial (Singapore), and JKSE -Jakarta coposite Inde (Indonesia), The epirical study focuses on etree variation of daily returns on arket indices of four selected AEC countries during the period of Etree distribution - Based efficiency The distribution-based approach usually equivalent definitions involving various odifications the cuulative distribution function and its inverse, such as integrated (inverse) CDF, quintiles and conditional value at risk. For Peak Over Threshold or Generalized Pareto Distribution (GPD) ethod utilizes data over a specified threshold. Jiahn-Bang Jang (2007) defined the ecess distribution as [ F( h) F( h)] F h () = Pr(X-h < X > h) = 1 F( h) Where h is the threshold and F is an unknown distribution such that the cuulative distribution function (CDF) Dentcheva and Ruszczyński (2003) introduced the following linear progra with distribution-based stochastic doinance constraints. a f ( ) E( X ) (1) n s.t. ik k sij yi, i, j 1,..., (2) k 1 1 s v, i 1 ij i j 1,..., (3) 0, s ij i, j 1,..., (4) (5) 2 Where j E[( y i y) ] F Y ( y j ) is the epected shortfall of y. The constraints fro the function basically ensure that E[( X ) ] E[( y) ],, which equivalent to SSD doinant is of over y, see Dentcheva and Ruszczyński (2003, 2006) for ore details. Another distribution-based test recently published in Kopa and Chovanee (2008) eploys the conditional value at risk defined as z CVaR ( z) E( z > VaR (z) ), (6) Where VaR 1 (z) is the value-at-risk of z, that is F z ( ) 3. Data description Daily returns on AEC arket indices during period of were collected as shown in Table 1 and

3 122 Chaitip Prasert et al. / Procedia Econoics and Finance 5 ( 2013 ) presented by graphically in Figure 1, Figure 2, Figure 3 and Figure 4.The Unit Root test was conducted based on the test developed by Phillips and Perron (1988) proposed a nonparaetric ethod for scheing for higher order serial correlation in tie series data Table 1: Data description of coefficient of variation by using edian on AEC arket indices during period of Country Standard Deviation Median CV CV edian Thai % Malaysia % Singapore % Indonesia % 4. Graphical f() Figure 1. (a) ; Graphical presentation diagnostic plot of SET daily closing of stock-inde returns during period of f() Figure 2. (b) ; Graphical presentation diagnostic plot of KLSE daily closing of stock-inde returns during period of f()

4 Chaitip Prasert et al. / Procedia Econoics and Finance 5 ( 2013 ) Figure 3. (c); Graphical presentation diagnostic of FTSI daily closing of stock-inde returns during period of f() Figure 4. (d); Graphical presentation diagnostic plot of JKSE daily closing of stock-inde returns during period of Research results Table 2 Output Results of Phillip-Perron Unit Root Tests during period of Equity Markets With intercept and trend Critical Value Order level in AEC At Level At Level At Level SET I(0) KLSE I(0) FTSI I(0) JKSE I(0) Note. Significant at 1% level; source: Fro coputed The results of Phillip-Perron unit root tests confired daily returns of SET, KLSE, SSE, and JKSE. However, the epirical study found evidence for the non-eistence of a unit root I(0). The null hypothesis of Phillips-Perron test is the proposition that iplies no effect or no relationship between daily returns for the null hypothesis that has a unit root I(1) against a stationary I(0) alternative across four national equity indices to assess intra-daily volatility dynaics in AEC as shown in Table 2. The findings indicated evidence intra-daily effect fro arket to arket of four AEC national equity

5 124 Chaitip Prasert et al. / Procedia Econoics and Finance 5 ( 2013 ) arkets. A conclusion of four AEC diagnostic plots for the threshold odel are shown in figure 1,2,3 and 4 that results including SET, KLSE, FTSI, and JKSE. These ean the accuracy correspondences to each the probability plot and quintile plot is unconvincing because the curve is about linear. In addition, the return level illustrates the large certainties that correspond eactly to the odel. The etreevars result using edian by Generalized Pareto Distribution (GPD) covered five period of intra-yearly volatility as depicted on table 1. Assessent of relative etree returns of frequency odels used epirical Generalized Pareto Distribution (GPD) odel calculed edian of daily returns natural log of stock based on threshold ethod as depicted on Table 1. Results of etreevar characterized that the Kuala Lupur Coposite Inde (KLSE) of Malaysia was the ost efficiency for decision-aking under risk of investors which CV edian equal %. Additionally, the subordinate of etreevar were Jakarta Coposite Inde (JKSE) of Indonesia, Singapore Straits Industrial (FTSI) of Singapore and SET Inde of Thailand that equal %, % and respectively. 6. Conclusion Writing a study of etreevar ethods by using edian of returns on AEC arket indices raises a question regarding the scope of the efficiency of relative etree returns for the selected period of the study. Results indicate that forecasting ethods based on GPD odel covered three paraeters, and, typically, odels for the presence of evidence for the predictability of relative etree returns and the eistence of a 'gainer-loser' effect fro arket to arket. Secondary data was used to produce evaluating of the returns covering the five-year period on AEC arket indices of SET, KLSE, FTSI, and JKSE. This original paper presents the first eperientation introduced etreevar of selected four national equity arkets in etree value estiators. According to coputed output results, statistical techniques confired that a change in percent KLSE or Kuala Lupur Coposite Inde (Malaysia) was the best efficiency for investent in ASEAN Econoics Counity or AEC. Research results during this period confired the best forecasting ethod based on the boundaries of the ter GPD odel that covered three paraeters, and. For the best equity arket in AEC, KLSE was the equity arket because CV edian of KLSE equal % which is the best value if copare with others. 7. References Best Philip Ipleenting Value at Risk. Blanco C. and Blostro S VaR Applications: Setting VaR-based Liits. Financial Engineering Associates, Inc. Chaitip Prasert.; Chaiboonsri C. and Chaithip Arreyah The Value-at-Risk (VaR) Of South East Asian Countries: Forecasting Long Meory in Etree Value Estiators. China-USA Business Review 2011, 10 No.9 (Septeber): Ebrechts P Etrees and Integrated Risk Manageent. London: Risk Books and UBS Warburg. H.G. Schodt.; Stochastic Doinant in Portfolio Analysis and Asset Pricing. Degree of Doctor fro the Erasus University Rotterda. Stuart Coles. (2001). An Introduction to Statistical ing of Etree Values