THE APPLICATION OF VALUEATRISK IN MEASURING RISKS OF VIETNAMESE STOCK MARKET

THE APPLICATION OF VALUEATRISK IN MEASURING RISKS OF VIETNAMESE STOCK MARKET THE APPLICATION OF VALUEATRISK IN MEASURING RISKS OF VIETNAMESE STOCK MARKET THE APPLICATION OF VALUEATRISK IN MEASURING RISKS OF VIETNAMESE STOCK MARKET THE APPLICATION OF VALUEATRISK IN MEASURING RISKS OF VIETNAMESE STOCK MARKET THE APPLICATION OF VALUEATRISK IN MEASURING RISKS OF VIETNAMESE STOCK MARKET

Dissertation submitted in partial fulfillment of the

Requirement for the MSc in Finance

FINANCE DISSERTATION ON

THE APPLICATION OF VALUE-AT-RISK

IN MEASURING RISKS OF VIETNAMESE

STOCK MARKET

NGUYEN THUY DUNG

ID No: 20000233 Intake 3

Supervisor: Dr Tran Manh Ha

September 2020

EXECUTIVE SUMMARY

Financial market is highly important for the growth and development of anyeconomies globally as the facilitators of the sources of funds and the uses offunds However, an inherent characteristic of the financial market is its volatilityassociated with a diversified set of risks Yet, the problematic puzzle of measuringthe complex risks of the financial market remains a challenge for not onlyacademic scholars but also financial market players To this end, this paperattempted to develop closer analysis to the presence of Value-at-Risk (VaR) inquantifying the risks associated with the financial market Within the scope of thispaper, the focus would shed the light into VaR application in Vietnamese stockmarket With the consideration of the VN30 index as the valuable snapshot of thestock market in Vietnam for the period from 21 April 2019 to 20 April 2020, thispaper implements four VaR approaches, including Parametric Value at Risk(PVaR), Historical Value at Risk (HVaR), Modified Value at Risk (MVaR) andConditional Value at Risk (CVaR) Each method might have specific weaknessesthat can be overcome with the advantages of other ones The main findings of thispaper are hoped to provide practical insights on the application of VaR intomeasuring the risk for stock market in Vietnam based on the in-depth analysis ofprevious literatures on this matter

TABLE OF CONTENTS

EXECUTIVE SUMMARY i

TABLE OF CONTENTS ii

LIST OF FIGURE iii

I INTRODUCTION 1

1 Research rationale 1

2 Research background 1

3 Research objectives 2

4 Research contribution 3

5 Synopsis of research 3

II LITERATURE REVIEW 5

1 Background of financial risks 5

2 Value-at-Risk (VaR) 14

III VIETNAMESE ECONOMY AND STOCK MARKET 23

1 Overview of Vietnamese economy 23

2 The development of Vietnamese stock market 26

IV METHODOLOGY 30

1 Research design 30

2 Data collection and data analysis approach 31

V FINDINGS AND DISCUSSION 34

1 Descriptive statistics 34

2 Risk assessment using VaR 38

VI CONCLUSION 42

REFERENCES 43

APPENDIX 48

LIST OF FIGURE

Figure 1 Comparison between VaR and CVar 22

Figure 2 Vietnam’s GDP Growth from 2013 to 2019 23

Figure 3 FDI growth in Vietnam from 2015-2019 24

Figure 4 Political Stability Index in Vietnam 25

Figure 5 Bond market growth from 2010-2015 27

Figure 6 Vietnam Stock Market Structure 28

Figure 7 Stock market capitalization as % of GDP in Vietnam 29

Figure 8 VN30 Daily Compounded Rate of Return 34

Figure 9 VN30 index descriptive statistics 35

Figure 10 Descriptive statistics for individual stocks in VN30 index 36

Figure 11 Mean Compounded Rate of Return for individual stock 37

Figure 12 Summary of the VaR result for VN30 index 38

Figure 13 Individual VaR results at 95% confidence level 39

Figure 14 Individual losses of VN30 index components measured by VaR at confidence level of 95% 41

I INTRODUCTION

1 Research rationale

Financial market is highly important for the growth and development of anynation in the world Acting as the intermediary to connect the sources of fundsand the uses of funds, one of the financial market inherent characteristics would

be its uncertainty and volatility Risks have been among key consideration forfinancial theorists since the longest time Indeed, the movements of commoditiesand financial securities deal with various uncertainties resulting from a wide range

of attributes To this perspective, risks can be considered as an unavoidable aspect

in the financial market and in business world As a matter of fact, thecomplication of risks in the financial market can be far more complication,contributing to further difficulties for the players in the financial market Series offinancial crisis and scandals booming from the beginning of the century have put

on the questions of how such uncertainties and risks associated with the financialmarket can be managed and supervised The management of risks requirecomprehensive efforts in identifying risks, assessing and measuring its impacts aswell as determining the proper risk mitigation method In response to suchalarming signals, researchers and scholars have long been gravitated towards thedevelopment of a proper sophisticated model to measure and tackle risks One ofthe most important and traditional method of risk measurements frequentlyutilized by the scholars is the usage of Value at Risk model (VaR) At its core theVaR model measure the potential losses led by unfavorable market movements

As a matter of fact, over the past few periods, the VaR model has become astandard tool utilized in the risks management process by not only financialmarket players but also management in other business sectors

2 Research background

Vietnam has become one of the continually growing nations in the Asianregion as well as globally The development of Vietnam has been emerging afterVietnamese Government’s efforts in reforming its traditional economy Stronggrowth in the national economy is largely strengthened by the contribution of thecirculation of capital between the savers and investors Hence, the health offinancial market infrastructure become increasingly critical for the development ofthe national economy It is essential to identify that the financial market inVietnam is heavily relied on the facilitation of the stock market while bond

market is relatively inexperienced and dominated by the institutional players Thestrong composition of the stock market in Vietnam highlights the significance ofstrengthening this market against the presence of increasing risks In addition, it isimportant to highlight that the stock market in Vietnam is relatively young andinexperienced while the infrastructure and regulatory frameworks have yet beenable to keep up with the emerging expansion of the financial market and nationaleconomy Moreover, the ongoingly growing participants of foreign investors aswell as other impacts of globalization process enhances the complexity of thefinancial market Such attributes further contribute to higher level of volatility andriskiness faced by investors However, to sustain higher level of efficiency incirculating the capital and investments of the financial market as well as attractinvestors to participate in such process, it become urgent significance for Vietnam

to be able to identify proper approach towards risk identification and mitigationswithin the stock market as well as other aspects of the financial market

- To implement the VaR model in measuring risks in stock market inVietnam;

- To understand the current risks and volatility in the Vietnamese stockmarket;

- To determine the existence of diversified conclusions among differentVaR approaches;

- To further outline potential recommendations on risks management forthe stock market in Vietnam

Within the scope of this paper, the study focuses on the Vietnamese stockmarket only By which it means that the risks associated with trading other form

of financial securities in Vietnamese financial market like bond trading,derivatives markets, etc would be neglected On the other hand, it also identifiesthat any comparative studies between Vietnamese and foreign stock marketswould not be within the scope of this paper

4 Research contribution

This paper would be expected to contribute to the academic world inmultiple aspects At first, there have been limited studies attempting to measuringthe level of risks in the financial market in Vietnam In fact, such implications inother emerging markets have been increasingly conducted by scholars andfinancial theorists Hence, this paper contributed to the discussion of risks andrisks management for financial market in Vietnam with respects to the usage ofValue at Risk model and its different approaches On the other hand, as theVietnamese stock market has been growing expanding over the last few years andcontributing massively to the financial market, its stability and growth remain aconsiderable concern for not only the players in these markets but also thegovernment and regulators Hence, a secondary contribution and significance ofthis paper would to assess the current risk level of Vietnamese stock market tohelp regulators to have the clear overview of current risks scheme in the financialmarket With the growing consideration of Value at Risk model in riskmanagement, which was further emphasized by the Basel III, this paper wouldfurther engage in a large picture of risks management in financial world and theapplicability of Value at Risk (VaR) model in the stock market

The main findings of this paper are hoped to provide practical insights onthe application of VaR into measuring the risk for stock market in Vietnam based

on the in-depth analysis of previous literatures on this matter Based on which,potential drawbacks and strengths of VaR as the measurement of stock risks inVietnam would be further revealed With that being said, the contribution of thispaper would potentially help future studies as well as financial market players abetter view on the risk landscape of Vietnam’s stock market

5 Synopsis of research

The later research would be conducted as follow For the first chapter, theresearch would give a brief introduction to the research rationale as well as the

determination of the research objectives and research questions Moreover, thisfirst chapter also gives the overview of the current stock market background inVietnam as well as different concerns with respects to the usage of Value at Riskmodel in financial market Moving on to the second chapter, this chapter isconducted as a comprehensive literature review of a wide range of issues inrelation to risks and the Value at Risk model In other word, this second chapterhelp establish a solid foundation and background for the later study and analysis.Next, the third chapter focuses on the analysis of the current economicbackground in Vietnam along with different development and characteristics ofthe financial market, especially the stock market in Vietnam Having profoundideas on the current macroeconomic status and the growth of stock market inVietnam would support a strong foundation on the later analysis as these factorsare highly correlated to the movements in risks and returns of the stock market.Later on, the fourth chapter provide a detailed information on the research designwith respects to a wide range of attributes and consideration for researchmethodology ranging from the research design, sampling method, data collectionand data analysis method, etc The next chapter summarizes and give a morecomprehensive discussion of the data collected from the Value at Risk model andattempt to answer the research questions and achieve the research objectives.Furthermore, in-depth descriptive statistics and discussion would be presented inthis specific chapter For the final chapter, this paper would give the conclusionremark sum up the entire paper and findings from previous chapters

II LITERATURE REVIEW

1 Background of financial risks

i The concepts and measurements of financial risks

Risk, which is also referred as volatility or uncertainty are importantconcept and aspect in not only the finance world but also the physical world Theconcept of risks has become one of the most popular concepts yet difficult tocapture and conceptualize in finance field Yet, it is noteworthy that physicalworld does not exist without uncertainty Furthermore, company must take onrisky investment opportunity to grow and prosper

In general, risks are much perceived as the degree of uncertainty whichmight be experienced in almost every aspects of life (Shkolnyk, 2019) To thisextent, risk can be characterized by two components including the level ofuncertainty and the relative exposure with respects to the object facing the formeruncertainty (Pasaribu, 2010) The concept of risks was further simplified intomathematical ideology with the work of Meyer (1985) on the uncertaintyprinciple However, to ideologize the risk concept in financial market, it would benecessary to take on a broader and more comprehensive point of view

In finance world, risk is an inherent aspect of every financial securities andinvestments In other words, when individuals participate in the financial marketthrough making investments on certain financial securities, they are exposed to awide range of risks As a matter of fact, all financial market players frominvestors, portfolio managers, investment bankers, securities rating agency, etc.concerned about the uncertainty of the return on their investments (Raza et al.2014) Such uncertainty might be contributed by a wide range of factors whichcan be either specific attributes associated with the financial securities themselves

or macroeconomic drivers affecting the entire financial market as a whole To thisextent, risks in financial market might depict the potential of unfavorable gapbetween the expected outcomes by the financial market players and the actualoutcomes (Shkolnyk, 2019) Thus, it can be perceived that any incident or activitythat might contribute to a possible unfavorable return might be termed as risks infinancial market

Attempting to quantify risk, scholars often consider the historicalmovements in the financial securities prices and behaviors as the basis ofmeasuring the extent to which the investment experience volatility (Hull, 2018)

Volatility in mathematical ideology measure the dispersion of stock pricemovement by either variance or standard deviation More specifically, using thestatistical approach, financial theorists measure risk level associated to theinvestment by incorporating the standard deviation of the financial securities over

a specific period of time (Shkolnyk, 2019) By measuring the variation offinancial securities prices, standard deviation depicts the volatility of financialsecurities prices through measuring the total gaps between the prices and thehistorical means during specific period (Hull, 2018) In addition, the level of riskmight be measured with skewness reflecting the asymmetry of the financialassets’ returns distribution (Pasaribu, 2010) In the wake of calculation ofskewness for financial assets’ returns, investment with a positive skewness mightdepict less risky investment while a negative skewness might reflect a higherreturn and higher risk portfolio (Shkolnyk, 2019) Taking similar approach tomeasure risks using the distribution of the financial assets’ return, scholars alsotake into consideration the usage of Kurtosis Another critically importantliterature body on financial risks associated with the introduction of the CapitalAsset Pricing Model (CAPM) (Jurczenko, 2018) At its core, the CAPM reflectsthe financial market characteristics a single factor, named market beta; to this end,the market beta measures volatility of the overall market On the other hand, themarket beta act as the basis for measuring volatility of other assets with specificrisk premium depending on the specific characteristics of the financial assets(Hull, 2018) In the light of CAPM, an asset with higher beta might indicate that itmight expose to higher risk than those with lower beta Under CAPM, beta ismeasured by the division of covariance between portfolio or individual stockreturns and the relative market return and the variance in market portfolio return(Hull, 2018) Nevertheless, such approaches in measuring risks using historicaldata might violate the fundamental financial theory of random walk To thisperspective, the random walk hypothesis explained that past information or stockprices cannot be used to predict the future movements (Chitenderu et al 2014).This in turn contributes to the comprehensiveness of risk measuring in the financeworld

ii Financial risks classification

Moreover, the comprehension of risk is further contributed by thesophisticated classification of risks There are various ways for scholars to

categorize different types of risks faced by financial market participantsdepending on their characteristics In general, risks can be classified into thebusiness risk, non-business risk and financial risks Firstly, business risk refers tothe type of uncertainty that business take on to achieve higher profitability andmaximize the benefits (Hull, 2018) In contrast, non-business risks can beidentified as the types of risks that are out of the control of firms, meaning thatbusiness has no choice on whether or not it would accept the risks Financial risks,

on the other hand, associate with financing activities or financial transactions andinvestment (Simons, 1996)

A more common approach to characterize investment risks in financialinvestments would be to separate the systematic and unsystematic risk In thewake of such risk classification, players in the financial markets face bothsystematic and unsystematic risks associated to their portfolios (Hull, 2018) Theterminology of systematic risk can be identified as those risks that areuncontrollable by the investors and have broad effects on the vast majority of thefinancial market Hence, these systematic risks have macroeconomiccharacteristics in nature While the systematic risks are broadly known as themarket risk, unsystematic risks can be broadly perceived as the specific risksaffecting a particular industry, sector or party rather than the entire market Inother words, the core classification factors between systematic and unsystematicrisks are their impact horizon which can be either the entire financial market orspecific to certain industry or investors (Chitenderu et al 2014)

With regards to each of the above risk classifications, scholars also furtherspecified different types of risks in relation to their individual characteristics(Simons, 1996) An important contributor for the systematic risks is the presence

of market risks, which arises from the changes in the market price of the financialsecurities leading to potential negative influences in the portfolio returns andvalues (Hull, 2018) Adverse movements in the market might lead to higherexposure to potential losses in values of the portfolio held by the investors Tobegin with, investors might face with risks associated with unfavorablemovements in interest rate Such volatility in interest rate affecting the investment

is referred as the interest rate risks At most, interest rate risks affect the bearing securities, especially those securities with fixed interest rate Interest raterisk might affect the investors’ portfolio in a wide variety of aspects At first,

interest-interest rate movements might affect the prices of shares, commodity, investmentleading to unfavorable changes in the values of the portfolios (Sinaee and Moradi,2010) On the other hand, the variation in interest rate might also impact theinvestors in terms of reinvestment opportunities (Hull, 2018) The fact thatinterest and dividend earned by investors would be reinvested at an unfavorablerate in the future might lower the return of the portfolio

On the other hand, one important market indicator that might contribute tothe movements in portfolio returns might include the variation of inflation rate(Quang et al 2018) Indeed, inflation rate affects deeply the entire market andeconomic leading to various changes across different sectors (Quang et al 2018).Inflationary risk or the purchasing power risk can affect the movements infinancial securities in various ways Firstly, the inflation rate is highly correlatedwith the health of the entire market as a whole; hence, any movement ininflationary background might lead to weakened or strengthened economicsituation (Vu and Tran, 2019) Theoretically, in the case of weakened economicsituation, inflation might lead to lower output and reduction in profitability ofcorporate resulting in unexpected negative effects on the asset returns (Sinaee andMoradi, 2010) Increase uncertainty of inflation rate movements might also affectthe riskiness of the financial assets which indirectly lead to higher required rate ofreturn on the financial assets (Quang et al 2018) The movement inflation ratealso correlates with the movements in prices of consumers goods and othercommodities (Simons, 1996) This depicts another risk faced by investors infinancial market wherein the commodity prices move in an unfavorable direction

in relation to the financial market (Sinaee and Moradi, 2010) The movements inprices of commodities like gold, oil, energy, etc are inseparably associated withthe financial market movements (Sinaee and Moradi, 2010) At its core the riskslinked with unfavorable movements in commodity prices are highly impactful asthey are the principle founding block of the economy as well as the entire marketbehavior (Simons, 1996) To this extent, higher commodities prices might affectthe production costs and corporation profitability while impacting on the prices ofconsumer goods (Chitenderu et al 2014) Undeniably, such movements wouldeventually hit the financial market with its impacts on prices of financialsecurities On the other hand, commodities like gold might act as the alternativeinvestment for other financial securities

In the growing presence of globalization trend in the financial market,another crucial market risk affecting the values and returns of the portfolio might

be depict by the movement in currency or foreign exchange rate (Vu and Tran,2019) Unfavorable movement in the exchange rate positions might lead todepreciation in the values of the portfolio and erode the potential returns of theinvestors Foreign exchange risks might deeply impact investors whose portfolioconsist of certain range of foreign investments and securities (Pasaribu, 2010).However, the actual effects of the exchange rate movements on the value offinancial commodities and the stock market might be far from universal asdifferent scholars depicted controversial and contradicted results with regards tothis specific idea (Suriani et al 2015)

Another popular systematic risks might relate to a broader view with regards

to the political stability of a nation Any political instability might hit hard on thefinancial market as well as the entire economy as it is the basis for development ofnational economy (Beaulieu et al 2005) Politic risks might arise from a widerange of political power execution, war, terrorism, expropriation, governmentbodies’ decisions, etc leading direct and indirect impacts on the nationaleconomy

iii The association of risks and return

Another important consideration in the financial theory would berelationship between the risk and returns As a fundamental financial norm, it wasfrequently cited and believed that a higher return on investment would more likely

to associate with an increase in risks faced by the investors (Hull, 2018) Hence,the statutory trade-off between risk and return has become the core idea within thefinancial world for making investment decision and the process of financialsecurities pricing (Abdullah et al 2011)

In explaining the risk-return relationship, a frequently associated conceptwith the risk and return tradeoff is the ideology of risk aversion wherein itdepicted that risk-averse investors require for additional compensation wheninvesting in higher risk financial assets (Shamsabadi et al 2012) By which itmeans that for risky investment, risk-averse investors would expect higher returns

to offset the risks that they might be exposed to (Abdullah et al 2011) In otherwords, higher risks investment would result in higher return under the effects ofrisk aversion Another important theoretical concept explaining such relationship

can be explained with the Capital Asset Pricing Model (CAPM) (Shamsabadi et

al 2012) The concept of the CAPM stipulate a positive relationship betweenfinancial securities risks and return, which holds true to the fundamental trade-offbetween risks and return

While much of the theoretical approaches to explain the relationshipbetween risk and return has come to an agreement on the positive effects of thesetwo variables, the empirical evidences from the financial market world mightdepict mixed and controversial results compared to existing financial theories As

a matter of fact, while some studies found supporting evidences on the associationbetween higher risks and higher return, others produced number of contradictevidences against this concept Indeed, a considerable number of studies havebeen carried out across different financial markets in order to assess the validity ofrisk-return trade off

To begin with, assessing the association of higher uncertainty with betterreturns in the Indian stock market, Dhankar and Kumar (2006) employed theMarket Index Model to investigate such impacts within this market The empiricalfound strong implication supporting the risk and return tradeoff meaning that inthe Indian financial market, higher risks positively contributed to higher returns.Taking a different approach by assessing the validity of the Capital Asset PricingModel in Greek financial market, Michailidis, Tsopoglou, and Papanastasiou(2006) oversaw the weekly stock returns of more than 100 listed companies for a5-year period from 1998 to 2002 against their corresponding beta as the measures

of risks In alignment with previous studies, it was suggested that for financialmarket in Greek also supported the core hypothesis of higher risk link to higherreturn Similar findings supporting the contributive relationship between these twovariables can also be evident by study from Leon, Nave and Rubio (2007) withfinancial markets in Europe A recent study conducted by Chiang and Zhang(2018) employing the TARCH-M model in the case of Chinese stock marketsuggested a significant and positive contribution of higher risks to higher return.Moreover, the paper indicated that this financial tradeoff did present across bothlocal and global risk-return relations In other words, a considerable number ofempirical evidences showed support towards the fundamental hypothesis of therelationship between risks and returns persisting in the real-world financialmarket Indeed, the fact that the similar findings could be experienced goingbeyond geographic boundary and timeframe might further strengthen the

consideration supporting such belief within the financial world and amongscholars.

Nevertheless, scholars also found contradiction evidences on the presence ofrisk and return tradeoff in the real-world financial markets Such argumentsfurther enhance the controversial on the discussion of this specific ideology Tothis perspective, Celik, Mandaci and Cagli (2009) investigated the presence of therelationship between these two variables in the Istanbul Stock Market from 2002

to 2008 The findings, however, suggested that the linear relationship between riskand return was not significant On the other hand, studying the Tehran StockMarket, Sinaee and Moradi (2010) indicated that the relationship between returnand risks were a non-linear relationship while there were limited evidencessupporting the positive contribution between the two variables Similar casescould be experienced in the study by Hasan, Kamil, Mastafa and Baten (2011)when invsetiging the Dhaka Stock Exchange Similarly, another study conducted

by Kayshik, Taneja and Kaur (2010) to assess the validity of risk-adjusted beta onmore than 120 companies over the period of 2004-2009 suggested that theincrease in systematic risks was not directly correlated with the higher return,which in turn reject the fundamental relationship proposed in traditional CAPMhypothesis By which it means that CAPM actually did not hold true in practicalcases A recent study presented by Rui, Rasiah, Yen, Ramasamy and Pillay (2018)investigating the Malaysian Stock during 2007 to 2015 attempted to assess therelationship between the two variables from the application of CAPM Evidently,when testing the association between these two variables, Rui et al (2018) foundlimited evidences supporting the validity of CAPM in Malaysian Stock Exchangewhile higher beta value did not relate to higher returns

In sum, it can be seen that even though the relationship between risks andreturn has long been a fundamental idea in the financial world and receivedcontinually discussion and attentions by scholars, its practical applicationremained relatively vague and ambiguous The contradiction results from a largenumber of studies might provide a potential gap for future studies and findings

iv Risk diversification

When it comes to risk in financial market, a popular concept comes hand with risk is the diversification of risks As indicated in the earlier section,investors tend to categorize risks into either systematic risk or unsystematic risk;

hand-in-wherein, the investors have limited to no power over the control of the systematicrisk (Aliu et al 2017) While it was much agreed by finance theorists that higherrisks tend to associate with potential higher returns; however, depending on thelevel of risk adverse and behavior, each investor might have different approach toinvestment decision (Lekovic, 2018) To this extent, the ideology of investmentdiversification has been widely accepted as a crucial investment strategy with theattempt to reduce the investment uncertainty and risks faced by the investorswhile minimizing the impacts on the expected returns of the investments (Aliu et

al 2017) As a matter of fact, risk diversification has become the fundamentalconcept in a broad range of fields beside finance and gain massive attentions fromdecisional theorists, consumer scholars, genetics scholars and economists(Mahmoud, 2017)

From a broad perspective, the ideology of risk diversification in financeconvey the attempts of investors by combining a set of financial securities withdifferent level of unsystematic risks to achieve superior returns compared to asingular securities portfolio (Lekovic, 2018) As a matter of fact, the present ofportfolio diversification can be traced back to the 4th century with an old-timeinstruction on allocation of fund before it actually took shape and gained massiveattentions from scholars by the mid of 20th century (Olibe et al 2008) The simpleform of portfolio diversification can be depicted as the efforts of the investors toconstruct their portfolio by adding different financial securities withoutconsidering the correlational influences among these financial assets (Lekovic,2018) With that being said, the simple form of diversification was much based onthe concept of the larger number of financial assets held the better (Medo et al.2009) It was argued that such simple form of portfolio diversification can helplower the risks at the cost of increasing portfolio management expenses (Lekovic,2018) To this perspective, it also indicates that through diversification by holdingmore financial securities without concerning potential effects among them wouldeventually help lower the uncertainty but might also decrease the efficiency level

By which it means that such diversification would hurt the investors in the endwith increasing costs of holding a massive portfolio while the desired riskreduction attempts remain ambiguous (Medo et al 2009; Lekovic, 2018)

A more advanced version of the simple portfolio diversification model wasdeveloped around the mid of the 20th century with the introduction of Markowitz’s

works wherein the author recommended that portfolio diversification should bemade with financial securities of different industries and sectors (Aliu, et al.2017) To this extent, it was explained that different industries might holddifferent fundamental risks and uncertainties; hence, by combining with thesecurities of another industries, the overall covariance would be less than that ofsingle-industry portfolio (Olibe et al 2008) In other words, by diversificationwith securities in different industry, investors would expect lower correlationamong the components of the portfolio implying the better effects fromdiversifications (Aliu et al 2017) This form of the diversification was referred asthe Modern Portfolio Theory (MPT) dated back in the 20th century The core idea

of theory was that by choosing to combine a broad range of low-correlatedsecurities, investors would be benefited from higher diversification impacts

At most, both simple and modern approaches to portfolio diversification putemphasis on diversifying unsystematic risks In the presence of ongoinglybooming globalization in the financial market, investors might attempt todiversify the systematic risks with proper approach to international diversification(Lekovic, 2018) In other words, rather than simply investing in domesticfinancial market, the investors might choose to expand the portfolio withinvestment in international financial market to help overcome the systematic risksassociated with a nation (Medo et al 2009) The high degree of cross-country andmultidirectional movements capitals in both equity and fixed income financialsecurities further eased the flow in favor of international diversification(Mahmoud, 2017) Furthermore, it was indicated that for those nations with highercountry risks and political risks, the effects of risk diversification throughinternational portfolio might be critically higher (Abid et al 2014) However, itwas also highlighted that the international financial market also faced with what-so-called the home bias puzzle, wherein investors have higher interests indomestic or local financial securities over international alternative fordiversification (Eun et al 2010)

It can be seen from the above arguments that the majority of scholars foundfavorable impacts of portfolio diversification on the reduction of risks.Nevertheless, the arguments on the optimal risk diversification method as well asthe size of diversification can be far from universal (Eun et al 2010, Lekovic,2018) On the other hand, it was also indicated that the combination of different

diversification method might contribute to better risk diversification effects aswell as higher benefits for the investors (Aliu et al 2017) Regardless of theongoing discussion, the concept of risk diversification remains as a crucial andimportant idea in the financial market

2 Value-at-Risk (VaR)

i Overview of VaR

The ideology of risk and management of risks have always been an integralpart considered by all players in the financial market To this extent, the vastscholars and theorists have been ongoingly developing different risk mitigationapproach The Value at Risk (VaR), which is commonly referred as the VaRmodels, is one of them It was argued that the pioneering pieces of work on theVaR model dated back in the early of the 20th century with a series of intuitivediscussion on the advantageous benefits of portfolio diversification (Lidija et al.2014) The historical development of VaR model could be much explained by theemerging growth of volatility of the financial market as well as within thebusiness context The concept of VaR model was developed by two independentscholars during 1952 with the attempts to maximize the level of rewards againstthe set level of risks (Lidija et al 2014) Nevertheless, during this early period,VaR remained a mere conceptual framework published and used as the supportingliterature for portfolio theory The presence of lenders and creditors alloworganizations to borrow and have better access to financial resources as well ascapital to serve their operational needs (Bogdan et al 2015) While firms becamehighly leveraged, their volatility would also increase (Quang et al 2018) At thatpoint of time, investors as well as firms were gravitated towards the specificmethodology to help them evaluate the potential losses faced With the increasingaccessibility and availability of financial information to serve the needs of VaRmodel, it became more popular within financial market as well as in businesscontext The growing number of firms incorporate this model into practicesfurther depicted its popularity and significance As a matter of fact, the VaRmethodology has become so popular that the Basel rules on banking governanceidentified it as a key risk identification and mitigation within the depositoryfinancial service sectors (Barjaktaroviü et al 2013)

The method of Value at Risk (VaR) is an important tool for risksmanagement in the financial sectors, especially for risks assessments of financial

securities, commodities, portfolios, etc The core question of the VaR modelfocuses on the determination of the approximation of maximum amount ofpotential losses under certain market condition (Angelovska, 2013) In otherwords, VaR model estimate the possible losses in relation to the confidenceinterval Another interpretation of the VaR model’s result is how much theinvestor’s portfolio might lose in its value within a certain period of time at agiven level probability (Simons, 1996) This is to say the most important attributes

in the VaR model consist of the determination of potential losses which can beexpressed in either relative or absolute value, the timeframe wherein risks aremeasured and the probability for such losses to occur (Gorbunova, 2016) Toemploy the VaR model, one must identify three critical components including acommon measurement unit, which is usually a monetary unit, a timeframe and aprobability for potential losses to occur (Angelovska, 2013) The growingpopularity and growing perceived importance of VaR has encouraged theoristsand scholars in enhancing its comprehensiveness and adding furtherconsiderations as well as components in the traditional VaR model to develop awide spectrum of variation of VaR methodology which can be used to calculateand measure the risks Yet, there have been no consensus on the best methodology

of VaR analysis (Al Janabi, 2007)

Researchers and managements are gravitated towards the usage of VaRmodel as the critical risk assessment method for multiple reasons To start with, acore advantage of the VaR model is its applicability and flexibility (Al Janabi,2007) Indeed, as it was argued by different scholars, the VaR model can beutilized to measure the risk associated with wide spectrum of assets and projects(Pasaribu, 2010) Going beyond its application in financial market, the usage ofVaR model can also be found in different fields like project management,corporate finance, and so on In addition, the VaR model does not concern only asingular risk factor but rather providing a wholesome consideration of risks andoffering an overall risks assessment of the portfolio (Rockafellar and Uryasev,2002) By considering the probability of potential losses, VaR generates thelikelihood of corresponding losses which might be critically essential for thedecision-making process (Angelovska, 2013) The flexibility of the VaR modelcan further be indicated for its applicability in the unit of measurement (Bogdan et

al 2015) The model can be expressed in the monetary units, which is relativelyapplicable for stock market measurements

Nevertheless, there is no model that is completely perfect VaR containscertain limitations which need to be considered and addressed throughout theconduct of this paper It was argued that the VaR model, regardless of itscomprehensiveness, is not a coherent risk measure resulted by its lackingsubadditivity (Gorbunova, 2016) Such characteristics, however, go against a coreunderstanding of the portfolio management with risks diversification efforts Onthe other hand, the presence of different models and approaches to the VaR modelgenerating different values might lead to increasing confusion and failure ofselection for the right VaR model (Quang et al 2018) It can be critically fatal forthe businesses and investors to be rational about choosing the right methodology

in application

ii Different approaches to VaR

As indicated earlier, there have been ongoing approach to develop differentapproaches towards the VaR analysis The vast approaches to VaR analysisemphasize on the investigation of the statistical distribution of the stock marketreturns (Simons, 1996) Critical methodologies on VaR analysis might consists ofthree core approaches, namely Parametric VaR, Historical Simulation VaR andMonte Carlo Simulation VaR, which would be discussed in more detailsthroughout the next sections

a Parametric VaR

To begin with, Parametric VaR might be the most common approach ofVaR adopted by VaR users Implied by its name, the core consideration ofParametric VaR model is founded on the estimation of the variance-covariancematrix of returns (Simons, 1996) To this perspective, the parametric approach toVaR analysis depicts a core assumption on the distribution of assets return Underthe parametric VaR assumption, the returns of the commodities should follow anormal distribution (Angelovska, 2013) As a matter of fact, it was argued thatthere exists other distribution that might wholly be captured by their parameters,the normal distribution would be more likely to be incorporated within theparametric approach This is a method that uses historical information to calculatesuch as arithmetic mean, correlation, standard deviation, and so on, varyingdifferently among the methods used for capturing risk value (Angelovska, 2013)

However, it was argued that the mean and standard deviation were morecommonly used for this approach (Bogdan et al 2015) To capture the ideology ofthe parametric VaR, Simons (1996) introduced an equation in measuring the risks

of portfolio implied by the variance of the portfolios In which, the portfoliovariance is calculated based on the formula of individual financial assets included

in the portfolio The equation of parametric VaR can be summarized as below

P is the volatility level of the portfolio;

ai is the amount of portfolio share of the asset i;

σ2

i is the volatility level of asset i;

pi,j is the correlation between returns of i and j assets

It is essential to highlight that the advantageous aspects of parametric VaR

is its simplicity As a matter of fact, this approach is relatively quick and easy forcalculation with the usage of simple descriptive statistics figures (Bogdan et al.2015) Nevertheless, as the parametric VaR assume a normal distribution for itsanalysis, when the assets’ return depicts a non-normal distribution as well as thesignificant outliers present, the results of parametric VaR might not be a goodreflection (Simons, 1996) It was argued that such assumption might lead tomisinterpretation of the actual returns (Pasaribu, 2010)

b Historical VaR

Historical VaR utilizes the historical market returns or prices, represented inthe form of distribution generated from the degree of reliability selected by theinvestors (Bogdan et al 2015) As the historical VaR analysis incorporates themarket prices of the financial assets, it is referred as a non-parametric approach toVaR model As indicated by its name, determining the proper timeframe for theanalysis is a crucial aspect to conduct historical VaR analysis (Pasaribu, 2010).Within the selected timeframe, the market figures obtained would bereconstructed with respects to their size and the degree of probability calculatedrisk value

In the past, it would be difficult to acquire the information incorporated inthe historical VaR analysis making it a less popular approach to VaR; however,over time when it becomes easier to obtain financial information and data, this

method is increasingly used thanks to its uncomplication (Rockafellar andUryasev, 2002) To further extents, the presence of different data analytic toolslike SPSS, R or Stata as well as data streaming platform such Reuters, Bloomberg,etc further compliment the usage of historical VaR

Moreover, the historical simulation approach to VaR model is believed toprovide superior unconditional coverage among existing method of VaR, whichhelp enhancing its presence in the finance field and risks management (Sollis,2009) To this extent, in measuring the effectiveness of VaR in practice, the usage

of back testing framework relies heavily on the unconditional coverage; hence,historical VaR by providing solid unconditional coverage would depict superioreffectiveness and validity compared to other methodology (Pasribu, 2010) Yet, inthe presence of conditional measured required for conditional coverage in thedataset, the advantages of historical VaR might be missed In addition, in practice,the incorporation of wide range of sophisticated conditional risks models mightprovide more comprehensiveness and validity in practice (Sollis, 2009) On theother hand, since the historical VaR is free from distributional assumptions as anon-parametric method, it might deliver better performance compared toparametric VaR model (Bogdan et al 2015) Another advantage of this method isalso its simplicity for calculation, especially in the presence of recent innovativedata streaming system (Sollis, 2009) To this extent, the historical approach wouldnot require the estimation of volatility or correlation among the interested stocks.Hence, it makes the historical VaR to be highly adoptable for any portfolio andany financial securities for full evaluation of risks

Nevertheless, the usage of historical simulation for the VaR model mightalso carry several drawbacks At first, it is worth noting key assumption ofhistorical VaR is that it suggests that the future returns would reflect the pastevents (Bogdan et al 2015) The vast scholars indicated that such assumption wasrelative vague, reflecting incomplete consideration of the movements in thefinancial assets’ prices (Al Janabi, 2007) This in turn violates the core idea ofirrelevant theoretical bodies To enhance the validity and effectiveness ofhistorical VaR model in measuring the risks faced by the investors, it becomesessential for the model to acquire a large dataset (Rockafellar and Uryasev, 2002)

In practice, the majority of study consists of 251-day sample size, which can beconsidered to be large enough dataset Moreover, another core assumption of this

approach is that the returns calculated have constant volatility and covariance,which result in problematic in assessing the sensitivity of the results (Quang et al.2018) Considering the timeseries data, the historical VaR has higher tendency toput emphasis on the effects of any recent impactful crises while othermacroeconomic factors might be neglected (Vasileiou, 2017).

c Monte Carlo VaR

The next method to VaR model is the Monte Carlo Simulation VaR (MCVaR) Unlike the above methods, this approach to VaR was based on the model-creation approach by incorporate Monte Carlo Simulation in generating theportfolio’s changes probability distribution and mathematically representing theeffects of risks on the portfolios (Quang et al 2018) To certain extent, the MCVaR is relatively similar to the historical VaR as they both determines thepotential loses regarding a fixed confidence level (Bogdan et al 2015) However,while historical VaR requires a large dataset to enhance its effectiveness inmeasuring risks, MC VaR can achieve similar effects with a relatively smallersample size (Li et al 2013) Additionally, to conduct the Monte Carlo VaR, usersmight choose a distributional hypothetical approach that is best describe the returnand changes in value of the portfolio and use the predetermined distribution insimulating the returns and risks (Corkalo, 2011) By which, it means that instead

of using historical changes like the previous approach, the Monte Carlo VaRimplemented a specific distribution estimated by the users in determining the level

of risks After simulating the risks or returns of the portfolio, the VaR calculated

as a percentile in accordance to the chosen confidence level

Overcoming the weaknesses of the parametric VaR, MC VaR allows users

to have a broader range of distributions, including both normal and non-normalones (Wu et al 2020) Compared to the previous methods, MC VaR is far morecomprehensive and sophisticated; in return, it is indicated as the most effectiveand accurate measurements of risks (Chen and Chen, 2013) Hence, its maindisadvantageous against other method is the complexity and the time requirements

to conduct a proper and insightful analysis When choosing such models, usersmust considerably reflect the trade-off between costs and effectiveness amongthese three models to identify a proper selection based on their needs andresources However, thanks to its sophistication, the evaluation of Monte Carlosimulation might be highly adaptable to any changes in economic forecasts

Nevertheless, by choosing the comprehensiveness, Monte Carlo simulation VaRrequires heavy computational efforts and estimation of particular hypotheticalprobability to draw on a meaning sample for the analysis (Quang et al 2018) Inthe presence of a massive portfolio, especially for internationally diversifiedportfolio consisting of a wide range of multi-currency financial securities, theintensiveness for the computations would be further enhanced As a matter of fact,the estimation of hypothetical probabilities is also important for the Monte Carlosimulation; hence, if these attributes are not determined properly, it would createmeaningless information.

d Modified VaR

Parametric traditional VaR performs well in accounting for the tails of thedistribution and precisely estimate the tail below the risk quantile determined bythe confidence interval However, such parametric approach might provideinsufficient consideration of the risks and returns in the case of non-normaldistribution of portfolio returns As indicated earlier, such disadvantages depict acore weakness in using the traditional VaR approach for employing a symmetricaldistribution function The usage of other non-parametric method like historical orMonte Carlo simulation approach might be helpful in the light of providing moreprecise prediction of risks and returns associated with the portfolio (Quang et al.2018) Nevertheless, they remain considerably flawed in explaining andestimating the risks associated with financial assets with intensive non-normaldistribution In practice, the presence of normal distribution for financial assets isconsiderably limited; hence, it becomes urgent to develop another approach toaddress the VaR model for non-normally distributed portfolios By taking intoconsideration higher moments of the return distribution into VaR calculationthrough employing skewness and kurtosis, the newly introduced VaR modelutilize the Cornish-Fisher expansion for the calculation of VaR model (Brian et al.2020) Such approach to VaR using Cornish-Fisher expansion is referred as theModified VaR model, which was first introduced in the early of 2000s with theworks from Ferve and Galeano (2002)

As a matter of fact, the usage of Modified VaR model has been increasingemployed by practitioners and adacemic scholars to determine the level ofvolatility in portfolio management (Cavenaile and Lejeune, 2012)

Another superiority of the Modified VaR is that in the case of normalitypersist with the portfolio return distribution, the results of Modified VaR would bethe same as the result of traditional parametric VaR approach The Modified VaRmodel is highly applicable for portfolio return with negative skewness distribution

or a more commonly cited name as fat-tails distribution (Brian et al 2020) Byemploying the Cornish-Fisher expansion for the fat-tailed distribution in the VaRmodel, the modified VaR approach would provide the higher estimated lossesthan the traditional method In contrary, if the portfolio’s return is positivelyskewed, the modified VaR model would result in a lesser result than traditionalmethod In other word, the modified VaR model would more likely to produce areliable result with the consideration of the spread of returns throughout thedistribution Nevertheless, the modified VaR carries certain weaknesses compared

to the previously introduced method First of all, it was indicated that the modifiedVaR model would not be able to provide reflective result of risk estimation for acomplex and sophisticated returns structure and non-continuous returns (Cavenaileand Lejeune, 2012) Unlike the historical VaR which conduct the risk estimationbased on the historical movements of returns and data, the modified VaR utilize theCornish-Fisher expansion to estimate risks based on the shape of the tail for thereturns (Brian et al 2020) By which, it indicated that the modified VaR provide amore mathematical approach to risk evaluation based on the extreme returns thathave not been observed or measured yet

At most, rather than being considered as an approach in VaR model, theCondition VaR tends to be separated as an individual theoretical body from thetraditional VaR (Sarykalin et al 2008) Indeed, the debates on the right tacticaltools for risk management between VaR and Conditional VaR have been givenmuch attention from the vast scholars concerning the presence of risks Theselection between these two approaches were much of the tradeoff among a widerange of factors from mathematical computation, properties, stability of the

estimation, costs, simplicity, etc (Sarykalin et al 2008) Conditional VaR can beestimated using a broad spectrum of financial engineering tools and to becalculated as the mean value of some percentage of the worst-case loss scenarios(Sharma, 2012) Yet, when using the same confidence interval, conditional VaRtends to produce a lower bound than other VaR methods

Figure 1 Comparison between VaR and CVar.

Initially, the CVaR was introduced to determine the amount of loss thatmight happen in tail events or; in other words, unfavorable events led by risksassociated with the investments On the other hand, VaR focuses on the possibility

of losses that may occur beyond the threshold of the confidence interval (Sharma,2012) It can be seen from the figure above that the CVaR deviation has a broaderrange than the deviation of VaR meaning that it would give a morecomprehensive and precise estimation of risks In other words, CVaR wouldaccount for potential losses going beyond the limitation of VaR model

Moreover, compared to VaR, CVar has superior mathematical properties as

it incorporates both continuous and convex function into the model for calculation(Brian et al 2020) To further extent, risk management through the usage ofCVaR would provide higher level of efficiency as it can be optimized with convexand linear programming methods (Sarykalin et al 2008) In contrast, due tocertain constrains, the VaR can be relatively difficult to optimize and improve In