Portfolio optimization analysis of federation of Euro-Asian stock exchances (FEAS)

Transcription

1 Calforna State Unversty, San Bernardno CSUSB ScholarWorks Theses Dgtzaton Project John M. Pfau Lbrary 2003 Portfolo optmzaton analyss of federaton of Euro-Asan stock exchances (FEAS) Selm Larlar Follow ths and addtonal works at: Part of the Corporate Fnance Commons Recommended Ctaton Larlar, Selm, "Portfolo optmzaton analyss of federaton of Euro-Asan stock exchances (FEAS)" (2003). Theses Dgtzaton Project Ths Project s brought to you for free and open access by the John M. Pfau Lbrary at CSUSB ScholarWorks. t has been accepted for ncluson n Theses Dgtzaton Project by an authorzed admnstrator of CSUSB ScholarWorks. For more nformaton, please contact scholarworks@csusb.edu.

2 PORTFOLO OPTMZATON ANALYSS OF FEDERATON OF ' EURO-ASAN STOCK EXCHANGES (FEAS) A Project Presented to the Faculty of Calforna State Unversty, San Bernardno n Partal Fulfllment of the Requrements for the Degree Master of Busness Admnstraton Selm Larlar December 2003

3 PORTFOLO OPTMZATON ANALYSS OF FEDERATON OF EURO-ASAN STOCK EXCHANGES (FEAS) A Project Presented to the! Faculty of t Calforna State Unversty, San Bernardno by Selm Larlar December 2003 Date

4 ABSTRACT nsttutonal and ndvdual nvestors all around the globe are lookng for dfferent ways to dversfy ther stock portfolo. Ths thess wll gve them a chance to understand the dfference between Euro-Asan stock market portfolos and the S&P 500. Ths thess wll also compare performance analyses among ten foundng members of the Federaton of Euro-Asan Stock Exchanges (FEAS), the S&P 500 ndey, the Ten Composte ndex and four sample portfolos, consstng of the ten foundng member countres of FEAS and S&P 500. The Ten Composte ndex s presented n detals n the subsecton called measure of overall performance. The data between 1995 and 2002 for the ten foundng countres of FEAS, S&P 500 ndex, and l Emergng1 Market ndex was used to execute these performance analyses. Frst, ths thess contans a detaled' research about stock exchanges of member countres under the organzaton called Federaton of Euro-As,an Stock Exchanges (FEAS). Second, t wll analyze the portfolo performances among the ten foundng member countres' stock exchanges. Thrd, t wll compare the FEAS portfolos wth the S&P500 and sample portfolos. Rsk and return analyss for sample portfolos shows that a portfolo consstng of 100% of the S&P 500 turns out to

5 have the jlowest Annualzed Return and also results n the lowest Annualzed Standard Devaton between 1995 and 2002, compared to other markets. The ndex portfolo weghted!by the ten foundng stock exchanges' market captalzaton offered the hghest Annualzed Return wth a moderate rsk level compare to other markets. For the ten foundng countres ther selves, the Bulgaran, Tehran and stanbul stock exchanges comparatvely out performed other foundng stock exchanges. The results of ths thess suggest t!hat nvestors should nvest n portfolos consstng of the S&P500, the Ten Composte ndex and the. ten foundng stock exchanges, rather than only nvest n ether the ten foundng stock exchanges or SLP 500. v

6 TABLE OF CONTENTS ABSTRACT... LST OF TABLES... LST OF FGURES CHAPTER ONE: CHARACTERSTCS OF EMERGNG MARKETS Lterature Revew for Emergng Market Studes... Characterstcs of Actual Emergng Markets... Typcal Restrctons n Emergng Markets... CHAPTER TWO: THE FEDERATON OF EURO ASAN STOCK 1 EXCHANGE MARKETS (FEAS), TREND [ ANALYSS AMONG TEN EMERGNG MARKETS, S&P 500 AND ALL EMERGNG MARKETS... Hstory of Federaton of Euro-Asan Stock Exchanges and Foundng Stock Exchanges (FEAS)... Goals and Objectves of FEAS... General nformaton about Ten Foundng Member Stock Exchanges... Lsted Companes... Market Captalzaton... Turnover Rato... Trend Analyses of Lsted Companes... Trend Analyses of Market Captalzaton. Trend Analyses of Turnover Ratos... Trend Analyses of Prce ndces... x x v

7 CHAPTER THREE: MACRO ECONOMC AND MARKET NFORMATON ABOUT TEN EMERGNG! MARKETS Grosb Natonal Product... Average nflaton... Budget Defct... Unemployment Rate... Market Segmentaton and nstruments... Stock Exchange ndces... CHAPTER FOUR: METHODOLOGY Measure of Performances... Measure of Return and Rsk Return... Rsk and Dversfcaton... Measure of Prce Movement Relatonshp... Measure of Prce Movement Relatonshp... Need for Beta Measure of Overall Performance... 1 Defnton of ndex...! Benefts of Creatng ndces... Establshng an ndex... j Prce-weghted Seres...! Value-weghted Seres... j Geometrc Mean of Percentage Changes... 1 Concluson Remark for Choosng Computaton Method ! Composte ndex... v

8 Performance Evaluaton Measures: Jensen ndex, Sharpe Rato and Treynor ndex... Jensen ndex... 1 The Sharpe Rato...! Treynor ndex...! Lmtatons of The: Jensen ndex, Treynor nde'x and Sharpe Rato... Crehton of Sample Portfolos... CHAPTER FVE: DATA ' Market Captalzaton Short Term Government' Bond Rates Prce ndces... ]... Composte ndex CHAPTER SX: ANALYSS OF FNDNGS Correlaton Coeffcent Analyses... Cross Secton Analyses... 1 Rsk and Return Comparson... Comparson of Sharpe Measures... J Comparson of Treynor Measures... ; Comparson of Jensen Measures... 1 Treynor Versus Sharpe Measure... Portfolo Analyss... j Rsk and Return Comparson... Comparson of Sharpe and Treynor ' Measures ! Comparson of Jensen Measures v

9 Treynor Versus Sharpe Measure CHAPTER STEVEN: CONCLUSON... 72! APPENDX la: FEDERATON OF EURO-ASAN STOCK J EXCHANGES MEMBER EXCHANGES APPENDX B: CONSOLDATED FEAS MEMBERS 2002 J STATSTCS APPENDXjC: HSTORCAL OVERVEW OF TEN FOUNDNG j STOCK EXCHANGES APPENDX D: MACRO ECONOMC AND MARKET NFORMATON j ABOUT TEN EMERGNG MARKETS APPENDX;E: MONTHLY PRCE NDCES FOR TEN FOUNDNG STOCK EXCHANGES REFERENCES 96 v

10 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table LST OF TABLES 1. 1 Emergng Market Elgblty Test for Ten Member Stock Exchanges ' Foregn nvestment Celng for Lsted 1 Stocks n Ten Foundng Member Countres Tax Rates n Ten Foundng Exchanges'! Countres... ' Lsted Companes Analyses on Market Captalzaton ($mo) Turnover Rato (%) ; Trend Analyses of Lsted Companes Trend Analyses of Market Captalzaton! ($mo) J Trend Analyses of Turnover Rato (%) Trend Analyses of Prce ndces (End of Perod Levels) Rsk and Return of The Ten Stock Markets, The Ten Composte ndex and S&P j. Beta Values... 13l. Short Term Government Bond Rates Correlaton Coeffcent Analyses Correlaton Coeffcent Matrx for Ten Stock Markets, The Ten Composte ndex, and S&P Cross Secton Analyses of The Ten Stock ; Markets, The Ten Composte ndex and S&P j. Rankngs Based on Two Performance : Measures x

11 Table 18.; Result Analyss on Sx Portfolos Table 19. Rankngs Based on Two Performance Measures X

12 LST OF FGURES Fgure l.j Locaton of Stock Exchanges' Fgure 2.j llustraton for Rsk and, Dversfcaton Fgure 3. Concluson Graph x

13 CHAPTER ONE CHARACTERSTCS OF EMERGNG MARKETS Ths chapter frst presents the lterature revew for Emergng[Markets; second, t revews the defnton and the two basc crtera of Emergng Markets n the ten foundng 1member stock exchanges' countres: whether or not they meet the requrements of beng an emergng market, and f ttey fall wthn the typcal restrctons of that market 1 L terature Revew for Emergng Market Studes The jlterature on emergng stock exchanges s classfed nto three categores. The frst category concentrates on dstrbuton analyss of returns. The second category dagnoses the adequacy of the asset prcng models by usng emergng market data, and the thrd category attempts to explan nterdependency of stock markets by usng stock market correlatons. Few! approaches have determned the return dstrbutons of emergng equty markets comparng wth developed equty markets. Those approaches have resulted fve dstrbutonal characterstcs, whch are hgh volatlty, hgh long-term returns, hgh autocorrelaton, tme-varaton of skewness and kurtoss and low 1

14 correlatjon wth both developed markets and other emergng markets (;Nu & Cu, ). Research on standard global asset prcng models show that these models fal to explan the cross secton of average returns n emergng countres. Based on analyses for predctablty of the returns, returns for emergng markets are more lkely than developed countres to be affected jby local nformaton (Harvey, 1995). Some researchers also try to establsh a relatonshp between emergng markets and contagon. Durng the second half of the 1990's, economc turndowns n emergng markets were a major characterstc of the economc landscape (Dungey & Zhumabekova, 2001; Edwards & Susmel, 2001; Forbes &[Rgobon, 2002). Those knd of turndowns were by no means'a new phenomenon, the specal attenton to the recent experences was the percepton of a heghtened possblty of contagon - the spread outward of pressures from one!crss country to other countres (Meyer, 2001). A typcal example of ths knd of contagon s the! collapse1of Thaland's currency that has trggered a chan of crses n other Asan emergng markets. Another example s the Russan fnancal crss that puts pressure on world fnancal markets to ndustral economes. The possble soluton to prevent that knd of fnancal crses 2

15 for emergng markets n the future s to buld robust domestc ^fnancal nsttutons and found domestc economc [polces (Meyer, 2001). Snce emergng markets are becomng more and more! accessble, research based on emergng market data are sgnfcant. Furthermore, two forms of nvestment nstruments would be avalable to nvestors n developed countres; closed-end county funds and Amercan Depostory Receptsj (ADRs). The frst nstrument, closed-end county funds, ate nvestment companes that nvest n portfolo, of assets n a foregn country and sell shares of these assets n the domestc market, lke n the Unted States. j Ths nstrument not only helps nvestors gan experence n a foregn country wthout the need of pckng ndvdual stocks n the foregn market, but also provdes better lqudty due to transactons executed domestcally. The second nstrument, Amercan Depostary ReceptSj (ADRs), gves rghts to foregn shares to be traded n dollars over U.S. stock exchanges or over-the - cojnter. ADRs are unque nstruments to solve many of!the problems arsng from nvestment restrctons, nformatonal problems assocated wth nvestng n foregn [securtes, and transacton costs (Nu & Cu, 2002). 3

16 Characterstcs of Actual Emergng Markets Accordng to the FEAS rules, membershp n the federaton s open to emergng stock exchanges n Europe and Asa j(feas Year Book, 2001/2002). The term Emergng Market needs to be explaned to fully understand ths FEAS rule. Emergng Market mples a stock market that s n transton, ncreasng n sze, actvty, or level of sophstcaton. Most often the term s defned by a number of parameters that attempt to assess a stock market's relatve]level of development and/or an economy's level of development. Accordng to the Standard and Poor's' standards, f a market for the stock exchange meets at least one of the two general crtera, ths market would be consdered an "Emergng Market." Standard and Poor's clearly determned those two crtera: a) The market should be located n a low or mddle ncome economy as defned by the World Bank, and b) The nvestable market captalzaton should be low relatve to ts most recent Gross Net ncome! (GN) fgures. The1 frst crteron s based on the World Bank's classfcaton of low and mddle-ncome economes. n 2000, The World Bank classfed economes wth a GN per capta l'pwer than $9,226 as low or medum ncome countres 4

17 (Standard! and Poor's Emergng Stock Markets Factbook, 2002). j The 'second crteron s based on small nvestable market captalzaton relatve to gross domestc product n a market. Non-nvestable holdngs nclude, but are not! lmted to, large block holdngs and parts of companes that arenaccessble due to foregn nvestment lmts. As llustrated n Table 1, the ten foundng members of the FEAS satsfy the World Bank crtera of beng low-ncome economes. For the second crteron, the nvestable market captalzaton-to-gn rato must be n the top 25% of emergng markets for three consecutve years to1 graduate from the Emergng Market Seres. Table l. Emergng Market Elgblty Test for Ten Member Stock Exbhanges 1 ' Stock Exchange Crtera Low/Mddle ncome Area GN Per Capta Amman $3, Bulgaran $5, Caro & Alexandra $3, Dhaka $1, jlstanbul $7, 'Karach $1, [Lahore $1, Muscat $1, 'Tehran $5, [Zagreb $7,

18 Accordng to ths standard, the ten foundng members stock exchange countres were found elgble to stay n the Emergng Market Seres (Standard and Poor's Factbook Emergng stock Markets, 2002). jtypcal Restrctons n Emergng Markets Thb subsecton presents typcal restrctons n emergng'markets such as captal controls (flexblty n enterng/extng to the market), foregn nvestment celngs(for lsted stocks, and tax regulatons. * 1 Captal controls (flexblty n enterng or ' extng to the market):, Flexblty n enterng or extng Emergng Markets vares from country to country, the ' research proves that nvestors can easly buy and sell stocks n those ten foundng member stock exchanges. l There are no sgnfcant restrctons for ' foregners n those stock exchanges, gvng nvestors more flexblty to make ther nvestment decsons among emergng markets : (FEAS Year Book, 2001/2002).! Foregn nvestment Celng Regulatons For Lsted Stocks: 6

19 j Foregn nvestment Celng Regulatons for lsted stocks n Emergng Markets are mportant j restrctons that nvestors should take nto ' consderaton when makng ther nvestment decsons. n 2002, researchers at Standard and! Poor's have reported those regulatons n a smple table as shown n Table 2. - Table 2. Foregn nvestment Celng for Lsted Stocks n Ten Foundng Member Countres 1 Regulaton for Celng Amman 100% n general Bulgaran, j Caro & Alexandra 1 Dhaka 1 stanbul 100% n general 100% n general 100% n general; 10% on bankng companes for a sngle entty 100% n general Karach 100% n general Lahore 1 100% n general Muscat! Up to 49% ownershp f company approves. Tehran 100% n general Zagreb 100% n general Ths table also shows that there are no lmtatons for nvestors n the ten emergng markets studed, except n the Dhaka and Muscat stock exchanges. n the Dhaka stock exchange, celng restrcton (10%) apples for stocks n the bankng ndustry. The Muscat stock exchange lmts ownershp of foregners to 49%, f the company j -approves! the nvestment. 7

20 1 Tax Regulatons Tax wthholdng s another sgnfcant ssue for! foregners nvestng n emergng markets. nvestors have a tendency to chose low-tax or zero-tax markets among world markets to avod hgher taxes. Standard and Poor's Emergng Stock Markets Factbook 2002 reports nformaton j regardng wthholdng taxes n Emergng Markets. j Table 3 summarzes wthholdng taxes for ten! foundng member stock exchanges. Table 3. ^ax Rates n Ten Foundng Exchanges' Countres. STOCK EXCHANGE 1 nterest (%) Taxes On.. Long Term Captal Dvdens (%) Gans (%) 1 Muscat 0.0% 0.0% 0.0% Tehran 0.0% 0.0% 0.0% Caro & Alexandra 0.0% 0.0% 0.0% stanbul.* 0.0% 5.5% 0.0% Amman 0.0% 10.0% 0.0% Zagreb 0.0% 15.0% 0.0% Karach 10.0% 10.0% 0.0% Lahore 10.0% 10.0% 0.0% Bulgaran 15.0% 15.0% 15.0% Dhaka! 15.0% 25.0% 0.0% Government Securtes are exempt from taxaton f held to maturty The!Caro and Alexandra, Tehran and Muscat Stock Exchanges are tax havens for nvestors wth zero tax wthholdng on nterests, dvdends and long-term captal 8

21 gans. Th;e stanbul stock exchange s the fourth tax haven requrng; only 5.5% tax on dvdends. Amman and Zagreb! takes the; ffth and sxth place requrng 10% and 15% tax on dvdends, respectvely. Snce Karach and Lahore are n the same country, Pakstan, tax rates are the same, 10% on dvdends and nterest ncome. Table 3 also shows Bulgaran! and Dhaka are at the bottom of the lst by requrng; relatvely hgh tax rates on nterest ncome, dvdends and captal gans. 9

22 CHAPTER TWO THE FEDERATON OF EURO ASAN STOCK EXCHANGE 'markets (FEAS), TREND ANALYSS AMONG TEN! EMERGNG MARKETS, S&P 500 AND ALL! EMERGNG MARKETS Ths chapter ntroduces the Federaton of Euro Asan Stock Exchanges. ntally, 12 foundng member stock exchanges (Amman, Bulgaran, Dhaka, Caro and the Alexandra, stanbul, Karach, Lahore, Tel-Avv, Muscat, Tehran, Ukranan, and Zagreb) were chosen for ths project, [but because of nsuffcent nformaton, the Ukranan and Tel Avv stock exchanges were elmnated from the'sample. The hstory, goals and objectves of the federaton wll be explaned n the frst two subsectons General nformaton about the ten foundng member stock exchanges, and the comparatve trend analyss among those stock exchanges wll be llustrated n the thrd subsecton of ths chapter. The followng map n Fgure 1 also shows the locaton of each member stock exchange of the FEAS! 10

23 Xagrab Stock Excuh&o ;S«>ssE*<hr«o _ ' '- -^.,8u>ga:fw^ < $«>",*. Ercbarge Georgan 1 - «,, J Stocl Exchange &,.x'- StekuStock: xchan(?«j SaseR >ctar-g Armenan -Stocx takulrjxbank Currency... Exchange-- Cxcjjangp. ' < -Tehran ' ^JSsorExchange: ftrana 'ft\s<con>"{ st-rbus 5tock- Soc*' Exchange Exchange ; Falestna Securtes Exchange _. *! 4z< - 4 : j Cawoft Ah'trxfra &esk \ Extthans fs.. /, Amtft&t Stock Exchange. kajtkhswn» < H. - SsoekExehanrge V <* /, * Kyrgyz Stock Exchanga- -^ftoshksnt ReaufcjJkan Stock SXcnange <tr State Gommecgty & Raw Materals :: Exchange of Turkmenstan -;<? Muscat Securtes Market X j Lahore Stodt Exchange ^ Karach Stock Exchange- </ [ Mortpww Stock Exnlwrrp j _4; j Dhaka Stock Exchange / Fgure 1.1 Locaton of Stock Exchanges Hstory of Federaton of Euro-Asan Stock Exchanges and Foundng Stock! Exchanges (FEAS) The 1 federaton was establshed May wth 12 foundng members. The federaton has evolved and now has 23 Member Stock Exchanges as seen n Appendx A. Membershp n the federaton s open to emergng stock exchanges n Europe and Asa. The major purpose of the FEAS s to create far, effcent and transparent market envronments among the FEAS members and ther operatng regons. 1 Harmonzaton of rules, regulatons and adopton of new technologes to facltate the objectves of FEAS are major purposes of the federaton (FEAS Year Book, 2001/2002). 11

24 The 23 member exchanges represent the federaton from 21 countres consstng of over 7,000 traded companes wth a market captalzaton of $109 bllon. Appendx B shows that the market captalzaton of the federaton hts a peak level and reached $200. mllon n n 2002, the! federaton had ts lowest market captalzaton, whch was $109 mllon. Ths table also shows that the ten foundng 'member stock exchanges represent the majorty of the 23 member countres n terms of total volume for stock exchanges, bond markets, as well as other markets. The total volume of stocks n ten foundng stock exchanges represents 99% of the total volume of stocks n all member stock exchanges. Regonal statstcs show that 'Other' volume, ncludng currency, T-blls, repo/ reverse repo and dervatves among other nstruments, represents 81% of the total market volume for all fnancal nstruments that have been traded n member stock exchanges. Appendx C shows that the oldest stock exchange s ' the Alexandra Stock Exchange, whch was offcally establshed n 1888 followed by Caro n The followng stock exchanges are ranked by establshment date: 1 The Karach Stock Exchange (KSE) was founded on J September 1947, 12

25 ; The Dhaka Stock Exchange (DSE) was ncorporated on March 1954, 1 The Tehran Stock Exchange opened ts doors on [ Aprl 1968, [ The Lahore Stock Exchange (LSE) was establshed j n 1970,! The Amman Fnancal Market for stocks was J establshed n 1976, The Muscat Securtes Market (MSM) was establshed and share tradng began n May 1989, ' The stanbul Stock Exchange, formally,! ntegrated at the end of 1985, The frst Bulgaran Stock Exchange (FBSE) was, establshed on 8 November 1991 and started l. tradng n May 1992,! The Zagreb Stock Exchange (ZSE) was ncorporated 1 n 1991 as a jont-stock company by 25 commercal banks and nsurance companes. Goals and Objectves of FEAS Objectves of the FEAS are lsted below (FEAS Year Book, 2001/2002) : j Encouragng collaboraton between member ' countres to develop the each securtes market. 13

26 Actng as the representatve of member stock exchanges around the world. Promotng the development of more ntegrated 1 nternatonal stock exchanges n the regon. ; Offerng lstng and tradng opportuntes for securtes ssued n the regon. The!federaton ams to utlze a common trade platform'model as well as mplement a data center to promote cross-market statstcs. Other specal projects under FEAS are: 1 To promote the growth of stock exchange operators through extensve tranng programs, To promote development of small to medum economc enterprses on a natonal level wthn member markets, and J To arrange regonal tranng n the area of T! for both T professonals and non-t professonals (FEAS Year Book, 2001/2002). General nformaton about Ten Foundng, Member Stock Exchanges Ths subsecton ntroduces general nformaton about the ten foundng stock exchanges based on ther market 14

27 captalzaton, lsted companes and turnover rato. Explanaton for each category s shown as follows: Lsted Cqmpanes As shown n Table 4, the Caro Stock Exchange leads wth 1151' lsted companes. The Karach and Lahore stock ' exchanges are second and thrd wth 711 and 592 lsted companes. exchanges; Other stock exchanges follow those two stock Bulgaran, 354, Tehran, 327, stanbul, 288, Dhaka, 229, Muscat, 220, Amman, 212, and Zagreb, 71. For comparson purposes, about 2,800 companes are lsted on 1 the New York Stock Exchange (NYSE). Table 4.;Lsted Companes! Amman j Bulgaran Caro &! Alexandra' Dhaka stanbul! Karach j Lahore Muscat Tehran Zagreb, EM 17,572 19,574 18,864 25,582 25,975 25,687 24,880 27,560 S&P-500! Mn j Max 17,572 19,574 18,864 25,582 25,975 25,687 24,880 27,560 Average ' 1,751 1,914 1,866 2,525 2,572 2,528 2,460 2,

28 Market Captalzaton Market captalzaton s bascally defned as the total dollar value of all outstandng shares, t s calculate'd by multplyng the number of shares tmes the current market prce. Ths term s referred to as market cap. Tabe 5 shows that the stanbul Stock Exchange reached the hghest market captalzaton ($34.4 mllon) n 2002 whle the Bulgaran Stock Exchange had the lowest market captalzaton, $712,000. Other foundng stock exchanges acheved the followng market captalzatons: Caro ($26.4 mllon), Tehran Stock Exchange ($14.3 mllon),! Karach Stock Exchanges ($10.2 mllon), Lahore Stock Exchange ($10.1 mllon), Amman Stock Exchange ($7mllon), Muscat Stock exchange ($5.1 mllon), Zagreb ($3.8 mllon), and Dhaka ($1.2 mllon). Table 5 also shows that the average market captalzaton of the ten foundng]stock exchanges, $921 bllon, was lower than the S&P 500's market captalzaton, $8,254 bllon. 16

29 Table 5. [Analyses on Market Captalzaton ($mo) Amman 4,670 4,551 5,446 5,838 5,827 4,943 6,316 7,093 Bulgaran N/A 11 N/A Caro & Alexandra 8,088 14,173 20,830 24,381 32,838 28,741 24,335 26,415 Dhaka ' 1,338 4,551 1,537 1, ,186 1,145 1,228 stanbul : 20,772 30,020 61,090 33, ,716 69,659 47,150 34,401 Karach 9,286 10,639 11,899 5,836 7,064 6,602 4,944 10,204 Lahore 1-9,234 5,463 5,989 6,947 4,724 10,179 Muscat 1,978 2,662 7,108 4,392 4,302 3,463 2,606 5,152 Tehran! 6,552 17,024 15,159 15,167 21,858 7,538 9,698 14,344 Zagreb 581 2,975 4,246 3,190 2,584 2,742 3,319 3,805 EM 1,893,625 2,223,895 2,133,165 1,775,267 2,948,429 2,608,486 2,572,064 2,684,562 S&P-500 4', 588,269 5,747,638 7,290,191 9,908,953 12,223,58111,586,78710,433,301 8,254,166 1 Mn - 1, Max 4', 588,269 5,747,638 7,290,191 9,908,953 12,223,58111,586,78710,433,301 8,254,166 Average ' 594, , , ,943 1,280,563 1,193,972 1,092, ,022 TurnoverRato Ths rato s the percentage of outstandng shares traded durng a perod of tme and was calculated monthly for the fen foundng stock exchanges. The formula for the rato s shown as follows: Turnover Rato (%) = Total Volume of Stocks (# of shares)/total Market Captalzaton f Turnover rato ndcates tradng actvty: for nstance, hgh turnover ratos ndcate a hghly lqud market and the low turnover rato ndcates a low lqud market. Table 6 shows that the Karach Stock Exchange has the hghest turnover rato of 200, whch means that the market 17

30 s more lqud compared to the other stock exchanges: Dhaka (40.), Amman (20), stanbul (20), Bulgaran (10), Caro (10,) Lahore (10), Muscat (10), Tehran (10) and Zagreb (3). Zagreb stock exchange had the lowest the lqudty compared to other foundng stock exchanges. Table 6. 'Turnover Rato (%) Amman Bulgaran Caro & ' Alexandra Dhaka stanbul f Karach Lahore ' Muscat Tehran ' N/A N/A N/A N/A Zagreb j EM ; S&P-500 j Mn Max ' Average To understand the trend of the above -mentoned bascs among the ten emergng markets, Table 8 / 9, and 10 were prepared,to show ths trend analyss between 1995 and Results from ths trend analyss are shown as follows: 18

31 Trend Analyses of Lsted Companes Tabl'e 7 llustrates an upward trend n number of lsted cbmpanes for the ten foundng stock exchanges! between 1995 and After 2001, except n Amman, Caro, l Muscat, and Zagreb, the number of lsted companes n other stock exchanges had a downward trend. The Bulgaran stock exchange lost 300 lsted companes between 1999 and Ths decrease n the number of companes lsted s due to new regulatons from the newly establshed Securtes and Stock Exchange Commsson. The new regulaton ntroduced a new requrement that all lsted stocks must have ther prospectuses Commsson n order to trade n the were no companes that were able to approved by the stock exchange. There comply wth ths requrement; Between 1995 trend n! all occurred, at therefore, tradng was suspended for a whle, and 2002, lsted companes had a postve emergng markets and an average growth rate 7% whle the number of lsted companes n the Bulgaran stock exchange grew by 45% on average, the hghest growth rate among other stock exchanges. 19

32 Table 7. Trend Analyses of Lsted Companes! Amman Bulgaran! Caro & Alexandra Dhaka stanbul Karach Lahore Muscat Tehran, Zagreb! EM 17,572 19,574 18,864 25,582 25,975 25,687 24,880 27,560 S&P Mn Max 17,572 19,574 18,864 25,582 25,975 25,687 24,880 27,560 Average, 1,751 1,914 1,866 2,525 2,572 2,528 2,460 2,683 Std Trend Analyses of Market Captalzaton Table 8 shows that the stanbul stock exchange and the Zagreb stock exchange have a unque poston compared to otherfoundng stock exchanges. The same table also llustrates that the market captalzaton of the stanbul l Stock Exchange dramatcally ncreased from $20 mllon to! $112 mllon wth an average growth rate of 54% between 1995 andl The closest growth rate to stanbul stock exchangers market captalzaton occurred n the Zagreb stock exchange wth a growth rate of 45% for the same perod. Trend analyss for market captalzaton of each J 20

33 Table 8. Trend Analyses of Market Captalzaton ($mo) 1! Growth (%) Amman : 4,670 4,551 5,446 5,838 5,827 4,933 6,316 7, OSS 6.15% Bulgaran % Caro & 1 Alexandra 1 8,088 14,173 20,830 24,381 32,838 28,741 24,335 26, % Dhaka 1,338 4,551 1,537 1, ,186 1,145 1, % stanbul 20,772 30,020 61,090 33, ,716 69,659 47,150 34, % Karach! 9,286 10,639 11,899 5,836 7,064 6,602 4,934 10, % Lahore! - 9,234 5,463 5,989 6,947 4,724 10, % Muscat! 1,978 2,662 7,108 4,392 4,302 3,463 2,606 5, % Tehran 5,552 17,024 15,159 15,167 21,858 ' 7,538 9,698 14, % Zagreb j 581 2,975 4,246 3,190 2,584 2,742 3,319 3, % EM 1,893,625 2,223,895 2,133,165 1,775,267 2,918,429 2,608,486 2,572,064 2,684, % S&P ,588,269 5,747,638 7,290,191 9,908,953 32,223,58L 11,585,787 10,433,3d 8,254, % Mn Max 4,588,269 5,747,638 7,290,191 9,908,953 12,223,581 11,595,787 10,433,3d 8,254,166 Average [ 544, , , ,933 1,280,563 1,193,972 1,092, ,022 stock exchange shows that the Bulgaran Stock Exchange s an nfant stock exchange compared to other stock exchanges. Total Market captalzaton n all emergng markets grew only 5%, on average, between 1995 and Other foundng stock exchanges wth hgh market captalzaton compared to the emergng markets are Tehran, Muscat and Caro Stock exchanges, wth growth rates of 12%, 15%, and 18% respectvely. Compared to the ten foundng stock exchanges' market captalzaton, S&P 500's market captalzaton grew only 9% durng the same perod.! 21

34 Trend Analyses of Turnover Ratos Turnover ratos are unque ndcators to analyze the lqudty of stock markets. Table 9 shows that turnover rato n the S&P 500 ranged between 3.82% and 9.46%, between and Overall turnover ratos for the emergng markets reached 84%, ts peak pont n 2000 as llustrated n the Table 9. Due to the new regulaton, a new requrement was ntroduced that all lsted stocks must! have ther prospectuses approved by the Commsson n order to,trade n the stock exchange. The Bulgaran stock exchange1has the weakest turnover rato, at 13%. Turnover rato for the stanbul Stock Exchange ranged from 20% to 226% between 1995 and Due to the devaluaton of the local currency aganst the U.S. dollar n Turkey, and the chan reacton n the lack of trade volume n the market, turnover; rato dramatcally decreased to 20, from 161 between 2001 and

35 Table 9. 'Trend Analyses of Turnover Rato (%) Amman ; Bulgaran Caro & 1 Alexandra: Dhaka ; stanbul Karach, Lahore Muscat Tehran ] N/A N/A N/A N/A Zagreb ; EM S&P-500! Mn Max Average Std 1 Trend Analyses of Prce ndces The1 Bulgaran, Karach and Tehran Stock exchanges have performed better compared to the S&P 500 between 1995 and ;. The average performance for those stock markets are 8.26%, 8.8%, and 21.53% respectvely, whch are above the S&P 500's average return of 5%. Due to new reforms and! re-entry programs to MF, the Lahore stock exchange had the worst growth rate of -12% among other markets. The emergng^markets' ndex also retaned a negatve fgure durng the same perod (see Table 10 for detal). 23

36 1 Table 10.1 Trend Analyses of Prce ndces (End of Perod Levels)! Growth (%) Amman (ASE) % Bulgaran (SOFX-50) N/A N/A N/A N/A % Caro & Alexandra (CASE-30) % Dhaka (DSE)! % stanbul (SE-100) % Karach (KSE-100) j % Lahore (LSE-101) % Muscat (MUSCAT- : ALL) % Tehran 1 (TEPX) % Zagreb (CROBEX) % EM Composte ndex % S&P-500 ' % Mn Max ' 3,269 4,615 5,365 4,003 5,759 3,591 3,554 5,044 Average 730 1,117 1, ,306 1, ,213 Std 24

37 CHAPTER THREE MACRO ECONOMC AND MARKET NFORMATON ABOUT TEN ' EMERGNG MARKETS To understand the dynamcs of each of the orgnal 12 foundng member stock exchanges, macroeconomc data of each stock exchange's country summarzed n a matrx format are llustrated n Appendx D. Ths chapter compares the ten emergng markets wth each other n terms of stock exchange ndces, GNP, nflaton rate, budget defct, 'unemployment rate, market1 segmentaton and nstruments. The matrx analyss n Appendx D helped to compare those categores. nterpretatons for each category are shown as follows: Gross Natonal Product Gross Natonal Product helps nvestors to understand the magntude of the stock exchange n a country. A number of prevous studes show that fnancal deepenng promoted the growth of GNP n emergng countres (Standard and Poor's Emergng Stock Markets Factbook, 2002). The researchjsuggests a strong connecton between stock market development and economc growth. Accordng to another study, "t s also clear that an actve equty market s! an mportant engne of economc growth n developng 25

38 countres or emergng markets" (Harvey, 1995). Comparatve analyss jn Appendx D shows the dfference between emergng [markets n terms GNP. The stanbul Stock Exchange leads wth a GNP of $199,437 mllon and Caro Stock Exchange 'follows t wth a GNP of $98,725. Average nflaton Purchasng power affects nvestment decsons n the! domestc'market and the comparatve nflaton rates n the matrx show dfferences between the markets. The Bulgaran, Zagreb and stanbul Stock Exchanges have nflaton rates over 50%: 102%, 86%, and 76% respectvely.! Budget Defct n terms of captal outflow and nflow, the budget defct of each market has an mportant mpact. Budget defctsllustrate whether a country has excess funds or lack of funds. Because the magntude of budget defct has! a strong-affect on borrowng and/or lendng rates n the market, nvestors should focus on ths fgure to make an effcent nvestment decson n a market. The stanbul and Dhaka Stock Exchanges are n countres wth comparatvely hgh budget defcts, $9,772 mllon and $2,732 mllon respectvely, however, those budget defcts! are relatvely small compared to the defct n 26

39 Amerca, (whch s $40 bllon by the end of The Tehran and Amman stock exchanges are n countres wth hgh budget surpluses, $5,518 mllon and $5,838 mllon respectvely. Other countres have reported budget defcts,; Pakstan (Karach & Lahore) has a defct of $221.8 mllon; Egypt (Caro & Alexandra) reported a defct of $118.4mllon; Oman (Muscat) has a budget defct of $299 mllon; Croata (Zagreb) has a defct of $39 mllon. ; Unemployment Rate Unemployment rate provdes nvestors wth suffcent nformaton about the general pcture n the economy and the matrx n Appendx D compares unemployment rates among the ten foundng emergng markets. Accordng to the matrx, Zagreb and Dhaka Stock Exchanges are countres n whch the unemployment rate s extremely hgh compared to other countres, at 21% and 35%. For nstance, Pakstan (Karach1 & Lahore stock exchanges) has the lowest unemployment rate, at 6.3% compared to other foundng stock exchanges' countres. The Amman (14%) Bulgaran (15.3%),, Caro (12%), stanbul (10%), and Tehran (14%) stock exchanges are n countres wth moderate unemployment rates. 27

40 Market Segmentaton and nstruments To effcently make an nvestment decson n emergngjmarkets, nvestors should understand market segmentaton. Market segmentaton not only ndcates the depth of the market but also ntroduces nvestment alternatves to nvestors n the market to dversfy portfolos. To understand market segmentaton, some terms from the[matrx analyss n Appendx D need to be defned: Frst Market or PO market s the market for new companes whle the secondary market s for exstng companes. Off-floor transactons represent the transactons between dealers and brokers placed out of the market. Dervatve market s the market n whch secondary products of stocks, currences and bonds are traded among nvestors. Equty and fxed ncome markets are markets for certfcate of deposts and annutes such as nsurance and mortgages. Bond markets are the place for fxed borrowng nstruments for governments and corporatons. The Amman, Bulgaran and stanbul Stock Exchanges have dfferent markets where nvestors can access dfferent nstruments rather than typcal stocks and bonds. Those nstruments are foregn securtes, depostory recepts, muncpalty bonds and mortgage bonds (only n The Bulgaran Stock Exchange). 28

41 [ Stock Exchange ndces EacA of the ten foundng stock exchanges uses a dfferent base for ther ndex calculaton. Some stock l exchanges use only certan companes n ther calculatons whle otler use all companes. For nstance, the Amman Stock Exchange uses all companes n the ndex computaton (ASE-All}, whle Bulgaran Stock Exchange has 50 companes for SOFX-50 ndex. The column for ndces of the ten foundng:member countres can help to determne the dfferences between stock exchanges n terms of ndex, structure. For nstance, SOFX-50 determnes that ndex calculaton s based on 50 stocks n the Bulgaran Stock Exchange1. The calculaton method for most of those ndces s based on market captalzaton. 29

42 CHAPTER FOUR METHODOLOGY Ths chapter ntroduces the methodology that s used n ths thess. Frst, processes for the methodology wll be lsted n fve steps, and second, each step n ths lst wll be explaned n detal. These steps nclude: j ' rsk and return calculatons, ncludng j dversfcaton concept, and measure of prce ; movements, the comparson of rsk adjusted performance for, the ten foundng stock exchanges, overall performance of those stock exchanges ; a creaton of sample portfolos to analyze.rsk ; dversfcaton n the ten foundng stock! exchanges. Measure of Performances j The followng procedures were used to compare the performance of the selected ten foundng member stock exchanges and to measure the rsk nvestors face when nvestng n these exchanges: j Rsk and Return Analyses to measure the monthly j. performances of each ndex from 1995 thru 2002, 30

43 J Correlaton Coeffcent Analyses to measure the relatonshp between prces movements between each country and S&P 500. Sharpe, Treynor and Jensen performance measures j to analyze the rsk adjusted return performances. of the chosen stock exchanges. A weghted average Ten Composte ndex consstng of.ten foundng stock -exchanges was created to measure the performance of those ten stock exchanges n a portfolo structure. 1 Four sample portfolo to compare performances among domestc, foregn and a combnaton of domestc and foregn nvestment Measure of Return and Rsk Return Thej concept of return provdes a convenent way to express the fnancal performance of an nvestment. Two methods are typcally used to calculate performance - return n dollar terms and return n percentage terms. n dollar terms, the return s the total dollars receved from the; nvestment. n percentage terms, the return s calculated on a percentage bass to avod the scale problems: of dollar returns. Ths thess used monthly percentage terms to get accurate solutons n performance 31

44 analyses by avodng scale problems. Frst, the monthly returns were calculated for the years 1995 and Second, the Average Monthly Return of each stock exchange was calculated over the seven years under nvestgaton: Algebrcally: l RtJ= (Pt - Pt_) / Pt_ Rt =. Return of market for month t Pt = Prce ndex of market for month t Pt_]_ = Prce ndex of market for month (t-1) n j Rmt = 2 (Rt) / 96 ' t= t = 1 to 96 (number of months for 8 years) Rmt = Average Monthly Return of market The;Average Monthly Return was then converted nto an Annualzed Return (AR) by multplyng by 12. Therefore, nvestors can effcently compare returns of chosen stock markets on an annual bass. The followng equaton was used to calculate AR: AR,= Rm * 12 AR = Annualzed Return of market Rsk and Dversfcaton The;basc premse underlyng the relatonshp between rsk and return s nvestors who lke returns but do not lke rsk- Ths means that nvestors wll nvest n 32

45 rsker -J than- average assets f, and only they expect to receve above average returns on those rsky assets. The rsk can,be measured n dfferent ways, and dfferent conclusons about an asset's rskness can be reached dependng on the measure used. There are two methods n whch the rsk can be consdered: on a stand-alone bass, where the asset's cash flows are analyzed by themselves, or n a portfolo context, where cash flow from number of assets are combned and then consoldated cash flows are analyzed;(relly, & Brown, 2000)., n one stock context, a stock's stand alone rsk can be analyzed from two standponts; on a stand-alone bass, where the stock s consdered solated, and on a portfolo bass, where the stock s held as one of the number of stocks n the portfolo. Therefore, an asset's stand-alone rsk s the rsk an nvestor would face f he or she held only ths one asset. No nvestment wll be undertaken unless the expected rate of return s hgh enough to compensate the nvestor for the perceved rsk of the nvestment (Relly, & Brown, 2 000). n portfolo context, a stock's rsk can be dvded nto two!components: A dversfable rsk component, whch can be dversfed away, or a market rsk component, whch reflects,the rsk of a general stock market declne. Ths 33

46 market rsk cannot be elmnated by dversfcaton. Only market rsk s relevant. Dversfable rsk s rrelevant to most nvestors because t can be elmnated (Relly, &! Brown, 2000). Fgure 2 helps nvestors learn how addng more stocks to a portfolo affects the portfolo rsk. Accordng to ths table, the portfolo s affected by formng larger and larger portfolos of randomly selected stocks from 34

47 n ths thess, the Emergng Market ndex represents market ndex, whle each of the ten foundng stock exchanges represents ndvdual assets. The sample graph n the table llustrates that the rskness of a portfolo consstng of large company stocks tends to declne and! approach;some lmt as the sze of the portfolo ncreases. Accordng to the sample graph n Fgure 2, the standard devaton of a one-stock portfolo or an average stock s!approxmately 35%, whle a portfolo consstng of all stocks, whch s called the market portfolo, would have a standard devaton of about 20.4%, whch s shown as the horzontal dashed lne. Almost half of the rskness nherent n an average ndvdual stock can be elmnated f the stock s held n a reasonable, well-dversfed portfolo. Based on nformaton n Table 11, the same relatonshp exsts among the Muscat, 52.2%, Lahore 33.1% Dhaka 26.7%, Zagreb 27.5 and Emergng Marker ndex, 18.8%. The four, ndvdual stock exchanges have hgher standard devatons than the Emergng Market ndex's standard devaton, 18.8%., whch also ncludes those four ndvdual stock exchanges' ndex. n ths thess, each stock exchange was consdered an ndvdual asset, whle 35

48 Table 11..Rsk and Return of The Ten Stock Markets, The Ten Composte ndex and S&P 500 Market Number of Observatons Average Monthly Return Standard Devaton Correlaton Coeffcent wth S&P COMPOSTE % 75.30% 0.32 STANBUj 96,1.70% 60.70% 0.45 MUSCAT! % 52.20% 0.13 Karach! % 34.90% 0.05 LAHORE % 33.10% 0.13 ZAGREB, 64 '..0.50% 27.50% 0.42 DHAKA % 26.70% 0.06 BULGARAN % 23.90% 0.07 CARO % 17.10% 0.08 S&P % 16.70% 1 TEHRAN! % 16.30% 0.03 AMMAN % 12.30% 0.03 the Ten Composte ndex, Emergng Market ndex and four sample portfolos were consdered portfolos. t s dffcult, f not mpossble, to fnd stocks whose expected returns are not postvely correlated. Most stocks tend to go well when the natonal economy s strong. Even very large portfolos end up wth a substantal amount of rsk, but not as much rsk than f all the money was nvested n only one stock. One of the purposes! of ths thess s to evaluate dfferent portfolo structures consstng of the ten foundng stock exchanges, S&P 500,j and the Ten Composte ndex. The chapter ttled Analyss of Fndngs concludes the results of analyses 36

49 based on those portfolo structures. Some rsk always remans, lt s vrtually mpossble to completely dversfy portfolo rsk. The part of the rsk of a stock that canbe elmnated s called dversfable rsk or unsystematc rsk, whle the part that cannot be elmnated s called market rsk or systematc rsk. The total of j those rsks s known as total rsk of the portfolo. Dversfable or unsystematc rsk s caused by such random events as lawsuts, strkes, successful and unsuccessful marketng programs, wnnng or losng a major contract, and other events that are unque to a partcular asset or stock. Snce these events are random, ther effects on portfolo can be elmnated by dversfcaton. Bad events n one asset wll tend to be offset by good event n: another. Market rsk stems from factors, whch systematcally affect all assets n the portfolo. Typcal events a're war, nflaton, recessons, and hgh nterest rates. Snce most assets n the portfolo tend to be negatvely affected by these factors, market rsk cannot be elmnated by dversfcaton (Brgham, Gapensk, & Daves, 2(0 00). nvestment rsk s bascally related to the probablty of earnng less than the return. n ths thess, (rsk concept was analyzed for the ten foundng 37

50 stock exchanges, Emergng Market ndex, the Ten Composte ndex, S&P 500 and sample portfolos consstng of the Ten Composte ndex and the S&P 500. Standard devaton s one of the ways to measure the rsk of each ndex. The smaller standarddevaton represents the lower the rsk of the ndex. Standard devaton provdes an nsght of how far above orjbelow the actual value s lkely to be. Unlke returns,[the rskness of a portfolo generally s not the weghted}average of the standard devaton of the ndvdual assets n the portfolo (Brgham, Gapensk, & Daves, 2000). The! followng formula was used to calculate Standard Devaton of Monthly Return for ten stock exchanges' ndces. ;! 11 o = a/ (2 (Rt-Rmt) 2 /n-1) t= p = Standard Devaton of Monthly Return of market 11 n = amount of months consdered (96) Rt Rmt = Return of market for month t = Average Monthly Return of market month t The. Annualzed Standard Devaton for each ndex was calculated n order to compare the rsk of the dfferent 38

51 countres on annual cross-secton bases. The followng equaton!was used for ths calculaton: Ao = V (o2* 12) : Ao = Annualzed Standard Devaton of market j Measure of Prce Movement Relatonshp Measure of Prce Movement Relatonshp Covarance and the correlaton coeffcent are two key concepts to measure the prce movement relatonshp. Covarance s a measure, whch combnes the varance or volatlty of a stock's returns wth the tendency of those returns to move up or down at the 'same tme other stocks move up or down. The covarance between two stocks tell us l whether the returns of two stocks tend to rse and fall together as well as how large those movements tend to be. Correlaton s a statstcal measure of the relatonshp between a seres of data, and the correlaton coeffcent s a measure of the degree of correlaton between the seres of data. Correlaton coeffcent vares between (-1) and! (+1). A postve sgn means that varables move together'whle the negatve sgn ndcates two assets tend to move n opposte drectons. Explanng the dea of dversfcaton wll help to understand the correlaton coeffcent analyss. Portfolo theory assumes that 39

52 nvestors are bascally rsk averse, meanng they wll! select tfe asset wth the lower rsk, but ths does not mply that everybody s rsk averse or that nvestors are completely rsk averse regardng all fnancal commtments. The majorty of nvestors attempt to dversfy ther rsk. The purpose of the dversfcaton s to reduce the standard devaton of the total portfolo return. A well-dversfed portfolo ncludes securtes that have a low coeffcent of correlaton. Tn dversfcaton, only the unsystematc rsk, whch s the rsk that s specfc to the frm, can be dversfed away n portfolo constructon. Market rsk or systematc rsk s the rsk of the entre market, and cannot be dversfed away. Macroeconomc varables such as money supply, nterest rate volatlty, ndustral producton, and corporate earnngs, would cause ths systematc rsk, whch remans n the market portfolo and cannot be dversfed away (Relly & Brown, 2000). n ths thess, correlaton coeffcents among the S&P 500,the Ten Composte ndex and the ten foundng member stock exchanges' ndces were calculated to measure the prce movement relatonshp between the U.S. and the selected! ten-member country's ndces. The formula for correlaton coeffcent s shown as follows: 40

53 rj = 2 (Rt - Rmt)(Rjt - Rmjt) / (o oj ) t= ; rj! = Correlaton Coeffcent between and j markets t = amount of months consdered (96 n ths thess) Rt j = Return of market for month-t Rmt; = Average Monthly Return of market for month t Rjt! = Return of market j for month t Rmjt = Average Monthly Return of market j for month t o * 1 = Standard Devaton of Monthly Return of market oj! = Standard Devaton of.monthly Return of market j An optmum portfolo s a combnaton of nvestments, each havng desrable ndvdual rsk-return characterstcs that also ft together based on ther correlatons. Ths deeper understandng of portfolo theory should lead nvestors to reflect back on how to use foregn stocks and bonds to'reduce the overall rsk of the portfolo. Need for;beta The,correct measure of an ndvdual stock's contrbuton to the rsk of the market portfolo s ts beta coeffcent, or smply beta, whch s calculated as follows: 41

54 Betd Stock of = S = rm o om ' (omp rm o am The market portfolo has a beta of 1.0. Addng a stock wth a beta of 1.0 to the market portfolo wll not change the portfolo's overall rsk. Addng a stock wth a beta of less than 1.0 wll reduce the portfolo's rsk; hence reduce ts expected rate of return. Addng a stock wth a beta greater than 1.00 wll ncrease the portfolo's rsk! and expected return, therefore, stock's beta s as a measure of how closely t moves wth the market. A stock 1! wth a beta greater than 1.0, wll tend to move up and down wta the market, but wth wder swngs. A stock wth a beta close to zero wll tend to move ndependently of the market. When a stock has a beta coeffcent of 1.0, f the market goes up 15% the stock wll also ncrease by 15%; f the market goes down by 15% the stock returns would decrease by 15%. A portfolo wth that knd of beta coeffcent would be as rsky as the market average. f a stock has a beta of 0.5, the stock s only half as volatle(as the market. t wll rse and fall only half as much as he market and a portfolo of such stocks wll be half as rsky as a portfolo of beta = 1.00 stocks. On the 42

55 other hand, f beta= 2.00, the stock s twce as volatle as an average stock. Therefore, a portfolo of such stocks F wll be twce as rsky as an average portfolo (Brgham, Gapensk,] & Daves, 2 000). The!beta for each market was calculated n order to measure each ndvdual portfolo's contrbuton to the rsk of the market portfolo. These calculatons also help nvestors to understand the volatlty of each market, whch s;essental to dversfy ther portfolos based on ther rsk preferences. Table 12 summarzes beta calculatons for the ten stock exchanges, the S&P 500, and sample portfolos. The Amman, Bulgaran, Caro, Dhaka, Karach,!Lahore, Tehran, Zagreb stock exchanges, and S&P 500 have.all beta less than 1.00 whle the four sample Table 12. Beta.Values 1 Market Beta Sample Portfolos Beta AMMAN AVERAGE BULGARAN AGGRESSVE CARO MODERATE DHAKAj COMPOSTE STANBUL NDEX PORTFOLO KARACH LAHORE MUSCAT TEHRAN zagerj

56 portfolos and the Ten Composte ndex have a beta hgher than Measure of Overall Performance Defnton of ndex The ndex s set at a numercal level on the base perod of startng pont aganst whch a percentage change can be compared to any partcular pont of tme. The ndex measures'the up and down movement of stocks or bonds or funds reflectng market prce and market drecton (Relly & Brown, '2000). A stock ndex wll reveal the overall trend n the equty market. t s a comprehensve measure of market trends ndcatng the general stock market prce movements. The ndex wll be the nvestor's yardstck for the level of the whole stock market, or a certan group of 1 stocks, aganst whch the performance of ndvdual stocks can be measured or judged. ndces are worldwde nstruments used by nvestors n developed as well as developng markets. Benefts' of Creatng ndces Benefts of ndces can.be summarzed n four ways: Summarzes the whole market: An ndex s j composed of companes from all sectors of the 44

57 economy, so t provdes an easy way to quantfy! the performance of the economy as well as the! market as a whole. ndces act as ndcators of 1 busness condtons snce stock markets are! beleved to be senstve to them. An ndex can also be constructed for a gven sector to measure the performance of that sector. j Leadng ndcator: Prces of companes, ' represented n the ndex, are equvalent to the! present value of future1cash flows. f future cash flows are expected to change (ncrease or 1 decrease), the ndex wll reflect these expectatons. Allows for a self-regulatng market:! Arbtrageurs can easly dentfy dscrepances! n the market and correct the market to ensure! that prces are accurate.! nternatonal nvestors can compare the j performance of the country's ndex to other ndces around the world. A strong return wll ncrease publc awareness and foregn nvestment j n ths market (Relly & Brown, 2000). ndces :are the major ndcator for the performance of the bond and/or stock market n each country. nvestors! 45

58 consder[market performance frst and the portfolo performance second. To provde nvestors wth suffcent nformaton, most of the nvestment frms and publc organzatons created ndexes. n ths thess, the Ten Composte ndex consstng of ten stock markets was created to provde nvestors suffcent nformaton about those ten stock exchanges. ^Establshng an ndex Choosng a sample, weghtng -the sample, and usng the computng ndex procedures are three major challenges to establsh an ndex. By recallng statstcs, a sample should represent the populaton, all stock performance seres. Samplng s the only way to determne somethng about those stock seres. To weght each member n the sample, fund managers and securtes analysts usually use three methods; prce-weghted seres, value weghted seres and un-weghted or equally weghted seres. Computng ndces by usng the sample and weghtng methods conssts of smple arthmetc average and geometrc averages (Relly & Brown, 2000). Prce-weghted Seres The[typcal example for ths ndex s the Dow Jones ndex and s calculated by usng the arthmetc average of 46

59 current prces. The changes n the prce of each stock nfluence the value of the ndex. One lmtaton for ths ndex s that the stock values are prce weghted, a hgh prced stock nfluences more weght than a low prced stock. Value-weghted Seres The ntal base for those types of ndces s calculated by usng the total market value of all stocks n the sample. The market value s calculated as follows: Market Value = Shares Outstandng * Current Market Prce Percentage change n the ndex s' calculated by comparng the market value of the ndex at tme (t + 1) to the ntal value ofjthe ndex at tme t. The lmtaton for ths method s companes havng a large market value have a sgnfcant affect on ndex changes, compared to a comparable percentage change for a small company. Geometrc Mean of Percentage Changes n addton to arthmetc average calculaton n the above - mentoned methods, geometrc mean of the holdng perods method s rarely used by some ndces such as Value Lne ndustral Average and Fnancal Tmes Ordnary Share njdex. 47

60 Concluson Remark for Choosng Computaton Method Because the ten foundng stock exchanges, Emergng Market ndex and the S&P 500 all use the market value weghted method wth the arthmetc average computaton procedure, ths thess wll use the same method to create a composte ndex consstng of ten foundng stock exchange j ndces. Composte ndex To measure overall performance of the ten foundng member stock exchanges, a composte ndex was created by- usng the ten stock exchanges' market captalzaton and monthly ndces provded by the headquarters of FEAS. Market captalzaton for each market was used to determne each market's weght n the composte ndex. To determne the ten Composte ndex followng formula was used: * 1 1 n 10 COMPNDXt = (w * Pt) t = tme ndex w '= weght Pt1= Prce ndex for market n tme t. The composte ndex wll be used to compare the overall performance of the ten foundng stock exchanges wth the performance of the S&P 500 between 1995 and

61 Performance Evaluaton Measures: Jensen ndex, Sharpe Rato and Treynor ndex Ths secton presents the classcal ndces used n ths study; Sharpe Rato (S), Treynor ndex (T) and Jensen ndex (J). t also ncludes a comparson of the ndces used n ths thess. Jensen ndex Jensen's alpha s the most wdely used ndex of performance among scholars and practtoners. t s defned as the dfference between the actual portfolo return and the estmated benchmark' return. The benchmark l could be 'based on ether the Captal Asset Prcng Model J (CAPM) or the Arbtrage Prcng Theory (APT) model. CAPM specfes the relatonshp between rsk and requred rates of return on assets when they are held n well-dversfed portfolos. f many factors were requred to specfy the equlbrum rsk/return relatonshp rather than just one or two, APT can nclude any number of rsk factors, so the requred rate of return could be a functon of two, three, four or tore factors (Brgham, Gapensk, & Daves, 2 000). The Jensen ndex has been used for ndvdual securtes as well as portfolos. Ths ndex s senstve only to depth and not to breadth; whle depth analyss ndcates magntude of excess returns, breadth analyss takes 49

62 magntude of resdual varance nto consderaton & Brown,,2 000). (Relly, E(R) = RF + S (E(Rm) - RF) E(R) = Expected return on portfolo RF - Rsk free rate of the market (short term government bond rate) E(Rm) = The expected return on the market portfolo of rsky assets. 1 S = The systematc rsk (beta) for securty! or portfolo The Sharpe Rato The'Sharpe Rato s defned as the rato of the excess return of the portfolo, over the rsk free return, to the standard devaton. For other applcatons, the relatonshp must be proportonal, that s, t s assumed that the future measure wll equal the same constant, typcally less than 1.0, tmes the hstorc measure. The Sharpe Rato ndcates the expected dfferental return per unt of rsk assocated wth ths same expected dfferencal return. Ths Sharpe rato s senstve to both depjth and breadth analyss. Whle depth analyss j means magntude of excess returns, breadth analyss concentrates on dversfcaton. Snce the standard devaton of return s the measure of rsk, the Sharpe 50

63 ndex s Only approprate for portfolos and not for ndvdual securtes (Wllam, 1994). Sl = (AR - RF) Ao Treynor ndex Sl = Sharp ndex of market RF = Rsk free rate of market (short term government bond rate) Ao = Annualzed Standard Devaton of Monthly! Return of market AR = Annualzed Return of market A measure of a portfolo's excess return per unt of rsk, eqal to the portfolo's rate of return mnus the rsk-free rate of return, dvded by the portfolo's beta. The Treynor ndex may also be defned as the rsk premum earned per unt of rsk taken, where beta s the rsk measure. Ths s a smlar rato to the Sharpe rato, except that the portfolo's beta s consdered the measure of rsk as opposed to the varance of portfolo returns. Ths s useful for assessng the excess return from each unt of systematc rsk, enablng nvestors to evaluate how structurng the portfolo to dfferent levels of systematc rsk wll affect returns. The Treynor ndex s a measure wth whch one may measure the performance of a 51

64 portfolo over a gven perod of tme. n order to use the Treynor ndex, the portfolo return, the rsk-free rate of return, and the beta of the portfolo should be calculated. The average return of a government bond or note over a gven perod of tme can be used for rsk free rate of return. The formula for the ndex s shown as follows [(Relly & Brown, 2000). Treynor = (Portfolo Return - Rsk-Free Return)/Beta or,! T! = (AR - RF) E J T = Treynor ndex of market j RF = Rsk free rate of market (short term j government bond rate)! Lmtatons of The Jensen ndex, Treynor ndex and Sharpe Rato Most researchers found that both the Jensen and Sharpe ndces are potentally useful, however, these j ndces [suffer sgnfcant lmtatons. The most crtcal ssues are the approprate benchmark to be used for comparson, the role of market tmng and the affect of transactjon costs. Forj Jensen, researchers argued that the Jensen's alpha s; senstve to the choce of the benchmark model 52

65 that s employed for comparson. Another argument s that the estmaton of Jensen's alpha may be based due to market tmng, whch s the ablty of fund managers to systematcally change the target rsk of the fund. When portfolo managers change the target beta for the fund by movng money among dfferent nvestments, estmaton bas can be ntroduced nto the benchmark model because t assumes a constant beta coeffcent over the perod under study. The Jensen performance measure also does not take care of transacton costs or expenses assocated wth the purchasejand sale of securtes. For;Sharpe, as compared to Jensen, ths ndex prevents!the problem arsng from the specfcaton of the benchmark model. Ths ndex also does not take nto consderaton the transacton costs or expenses assocated wth the purchase and sale of securtes. The1 Treynor ndex has smlartes wth the Jensen ndex, snce the beta coeffcent s the rsk measure. The Treynor ndex, lke the Jensen ndex, s nsenstve to breadth (.e., t gnores resdual varance). Wth beta as the rsk1 measure, the Treynor ndex s applcable for ndvdual securtes as well as for portfolos. The Treynor ndex has an advantage over the Jensen ndex. The Treynor ndex takes the opportunty to lever excess 53

66 returns nto account when rankng alternatves (Muth, Cho, & Desa, 1998)., Creaton of Sample Portfolos A portfolo represents a set of two or more assets. The return of a portfolo s equal to the weghted average of the return of the ndvdual ndces n the portfolo. Ths subttle llustrates how several sample portfolos were created to analyze possble rsk dversfcaton for nvestors. The followng sample portfolos were created to compare performance among domestc, foregn and a combnaton of domestc and foregn nvestment. Those sample portfolos also help to analyze how dfferent combnatons of ndvdual stock ndces affect portfolo rsk and return performances. By executng these analyses, nvestors can choose any of the portfolo combnatons accordng to ther rsk and return preferences. S&P 500 Portfolo: Ths portfolo conssts 100% of the S&P 500 ndex The Ten Composte ndex: The Ten Composte portfolo conssts of 100% of the Ten! Composte ndex ndex Portfolo: A portfolo comprsed of the Ten Composte ndex and the S&P 500, 54

67 Aggressve Portfolo: weghted accordng to ther market captalzaton Conssts 25% of the S&P 500 and 75 % of the Ten Composte ndex Average Portfolo:! Moderatej Portfolo: Conssts 50% of the S&P 500 and 50 % of the Ten Composte ndex Conssts 75% of the S&P 500 and! 25 % of the Ten Composte ndex Rsk and! return of those sample portfolos were analyzed on the bass of Annualzed Return.(AR), Treynor ndex, Sharpe Rato and Jensen ndex. Formulas1 used for these analyses are shown as follows: Annualzed Return: l AR(P) = (w X AR) t= j w = weght of the market captalzaton ; S (w) = 1.00! t=l Portfolo Standard Devaton: n b = (X! (Rt - Rmt)2) / n-1) t= = Standard Devaton of Monthly Return of market n = amount of months consdered (96) 55

68 Rt Return of market n t Rmt = Average Monthly Return of market n t..(market s emergng markets) 56

69 ! CHAPTER FVE DATA Ths chapter ntroduces basc statstcs for the data used to do performance analyses among the ten foundng j members of the Federaton of Euro-Asan Stock Exchanges (FEAS), S&P 500 ndex, Ten Composte ndex and four sample portfolos consstng of the ten foundng member countres of FEAS and the S&P 500. All data gathered for these performance analyses s based on the monthly observatons between 1995 and 2002.! Market Captalzaton As mentoned prevously, data about the market captalzaton of the ten countres was collected to determne the weght of each country. Ths determnaton helped to create sample portfolos and a composte ndex 1 for the fen foundng stock exchanges and the S&P 500. Data for market captalzaton of the ten foundng stock exchanges was gathered by usng FEAS Yearbooks. Market captalzaton for the S&P 500 and Emergng Market ndex (EM) were gathered from the Standard and Poor's Emergng Stock Markets Factbook 2002.! 57

70 j Short Term Government Bond Rates Short-term government bond rates were summarzed by usng the database at FEAS, and they determne the rsk free rate of the ten emergng stock markets. The rsk-free rate s needed to calculate the Sharpe ndex, Treynor ndex and Jensen's Alpha. Results n Table 13 ndcate that these rates vary between 3% and 69%. Table 13 ;. Short Term Government Bond Rates Amman, 3.0% 4.0% 7.0% 4.0% 4.0% 5.0% 6.0% 5.0% Bulgaran 10.0% 15.0% 12.0% 14.0% 15.0% 12.0% 13.0% 14.0% Caro & Alexandra 9.0% 6.0% 8.0% 8.0% 11.0% 12.0% 9.0% 8.0% Dhaka 1 5.0% 6.0% 8.0% 9.0% 7.0% 5.0% 8.0% 7.0% stanbul j 60.0% 75.0% 49.0% 69.0% 57.0% 65.0% 64.0% 56.0% Karach 11.0% 12.0% 13.0% 15,. 0% 12.0% 13.0% 15.0% 13.0% Lahore 14.0% 12.0% 15.0% 16.0% 17.0% 16.0% 14.0% % Muscat ' 18.0% 22.0% 20.0% 19.0% 15.0% 18.0% 21.0% 20.0% Tehran 9.0% 8.0% 11.0% 12.0% 13.0% 14.0% 16.0% 12.0% Zagreb 8.0% 9.0% 6.0% 11.0% 12.0% 8.0% 7.0% 9.0% Prce ndces All,' prce ndces were collected on a monthly bass j from the! FEAS database. The database ncludes prce ndces between January 1, 1995 and December 31, The Ten Comppste ndex and Performance Analyses were performed based on ths database and Appendx E llustrates those ndces n detal. All ndces for the 58

71 ten stock exchanges provded by,the headquarters of FEAS use the market captalzaton weghted method. ndces for the Amman, Bulgaran, Dhaka Caro, Muscat, Tehran and Zagreb stock exchanges use performance of all lsted companes n the market, whle ndces for the stanbul, Karach, and Lahore stock exchanges use the performance of a predetermned group of 100 stocks lsted n each stock exchange Composte ndex Research on the ten foundng stock exchanges' ndces! showed that due to lack of consstency among those stock exchanges n weghtng, sample selecton, and! computatonal procedure, t s dffcult to compare the results mpled by ndces across countres. n order to prevent ths problem, a composte ndex that conssts of the ten foundng stock exchanges' ndces was created, weghtedby ther market captalzaton. 59

72 CHAPTER SX ANALYSS OF FNDNGS!! Correlaton Coeffcent Analyses Ths chapter compares the correlaton coeffcents between the ten stock exchanges, the S&P 500 ndex and the Ten Composte ndex. The results of the calculatons are! shown n,table 14. All selected markets or portfolos have a postve correlaton coeffcent wth the S&P 500, rangng from to Ths analyss concludes that j the ten stock exchanges and the Ten Composte ndex tend to move n the same drectons wth the S&P 500; when the S&P 500 ncreases 1 unt, the Ten Composte ndex s expected;to ncrease 0.32 unts or the ndex of the stanbul Stock Exchange s expected to ncrease 0.45 unts based onanalyss shown n Table 14. Snce those coeffcents are too small, nvestng n FEAS stock exchanges mght reduce rsk substantally. n terms of rsk and return relatonshp, the Ten Composte ndex has the hghest average monthly return of 2.2% wth a standard devaton of 75.3%, whch represents the hghest rsk among other portfolos. Whle the S&P 500 had a poor monthly average performance (0.80%) between 1995 and! 2002, the Bulgaran Stock Exchange had the 60

73 Table 14 J Correlaton Coeffcent Analyses Market Number of Observatons Average Monthly Return Standard Devaton Correlaton Coeffcent wth S&P COMPOSTE % 75.30% 0.32 STANBUL % 60.70% 0.45 MUSCAT j % 52.20% 0.13 KARACHl % 34.90% 0.05 LAHORE! % 33.10% 0.13 ZAGREB % 27.50% 0.42 DHAKA ; % 26.70% 0.06 BULGARAN % 23.90% 0.07 CARO % 17.10% 0.08 S&P 5 0 0' % 16.70% 1 TEHRAN % 16.30% 0.03 AMMAN %, 12.30% hghest average monthly return of 1.9% compared to other stock exchanges for the same perod. n addton to the correlaton coeffcent analyss between the ten stock exchanges and the S&P 500, Table 15 llustrates the cross secton analyss n a matrx format for the ten stock exchanges, S&P 500, and Ten Composte ndex. Ths matrx would help nvestors to analyze how two of those1 portfolos tend to move together. Snce the correlaton coeffcent between the Bulgaran Stock Exchange; and the Karach Stock Exchange s less than 0 (-0.05) J these two portfolos are negatvely correlated; they ten'd to move n opposte drectons. Ths helps nvestors to dversfy ther portfolo by addng those two! 61

74 stock exchanges. Because the correlaton coeffcent between stanbul and Lahore s greater than 0, +0.30, those two portfolos are postvely correlated. Consequently, those two stock exchanges tend to move up. and down together. The stanbul and Caro stock exchanges also show a postve correlaton of between 1995 and The correlaton coeffcent between Zagreb and Karach shows a postve rato of n terms of negatvely correlated stock exchanges, Dhaka has negatve Table 15. Correlaton Coeffcent Matrx for Ten Stock Markets,! The Ten Composte ndex and S&P 500 Market MJ 1 HGWt CKK) TffKA smsh esmd: MERE MKHT 2A3TB 10 S&P 500 OMGBnE AYMN T.oo (0.02) :.oo CAED tr 0;5 U KA 0.13 (0.04) (0.010) [/fll.j SffNBUL 0.12' KRPCHC 0.04, (0.05) (0.14) METRE o.n (0.17) BM MBMT (0.02) 0.03 (0.014) 0.06 (0.20) (0.06) (0.06) 1.00, TEHW (0.0S>) (0.04) (0.06) AREB (0.045) loot 10 GMCSEE (0.01) (0.28) S&P ^ (0.11) 1.00 j 62

75 correlaton coeffcents wth Karach and Lahore, and respectvely. The matrx analyss to forecast the movement n Table 15 could help nvestors of ther composte_portfolos consstng of these ndvdual portfolos. By recallng the portfolo theory, a completely dversfed portfolo would have a correlaton wth the market portfolo of Therefore, f stock exchanges' correlaton coeffcents are close to +1.00, those stock exchanges should be chosen to establsh a successfully dversfed portfolo. Because the Lahore and Karach have a correlaton coeffcent of 0.88, nvestors would beneft greatly by selectng those stock exchanges for ther portfolo. A smlar bombnaton would be the S&P 500 and the stanbul Stock Exchange, whose correlaton coeffcent s Cross Secton Analyses Rsk and Return Comparson The purpose of these analyses are to compare each ndvdual stock exchange the Ten Composte ndex portfolo and the S&P 500 portfolo, n terms of annualzed return and annualzed rsk, as well as the performance evaluaton methods (Sharpe ndex, Treynor 63

76 ndex and Jensen ndex). Table 16 summarzed the result of these cross-secton analyses. Coeffcent of varaton or rsk per unt of return calculatons n Table 16 helps to compare rsk and return j relatonshps among ten foundng stock exchanges, the S&P 500 and Ten Composte ndex. Accordng to these calculatons, the Tehran Stock Exchange, The Bulgaran Stock Exchange and the S&P 500 have the lowest coeffcent varatons compared to Karach, Amman, stanbul, Zagreb, and Dhaka. Table 16 also shows that the Dhaka stock exchange,has the hghest coeffcent varaton of 5.56 compared to other stock exchanges.' Ths means Dhaka has the hghest rsk per unt of return. Comparson of Sharpe Measures Fndngs n Table 16 ndcate that the Ten Composte ndex, Bulgaran and Tehran stock exchanges outperformed the S&P 500 wth the hghest rsk premum returns of 23.6%, 40.8% and 47.1% respectvely. Karach exhbts a postverato slghtly lower than the S&P 500, 9.1%, Muscat, -44.6%, stanbul, -68.3%, Lahore, -63.7%, Caro, -53.4%, Zagreb, -12.3% and Dhaka, -7.1%, all have negatve rsk premum returns. Snce the bond, rates n each stock 64

77 Table 16. Cross Secton Analyses of The Ten Stock Markets, The Ten Composte ndex and S&P 500 Market Number of Observaton Annualzed Return Annualzed Rsk Rsk. Per Unt of Return (Coeffcent of Varaton) DHAKA % 26.70% 5.56 ZAGREB % 27.50% 5.00 STANBUL % 60.70% COMPOSTE % 75.30% 2.84 AMMAN % 12.30% 2.56 KARACH % 34.90% 2.22 S&P % 16.70% 1.80 BULGARAN % 23.90% 1.04 TEHRAN % 16.30% 0.83 LAHORE % 33.10% (3.94) MUSCAT % 52.20% (13.05) CARO % 17.10% (47.50) Market Sharpe ndex (S) Treynor ndex (T) Jensen Measure DHAKA % ZAGREB % STANBUL % 10 COMPOSTE % AMMAN % KARACH % S&P % BULGARAN % TEHRAN % LAHORE % MUSCAT % CARO % exchange 's countres out performed stock exchange's performance, negatve premum returns were retaned n those markets. 65

78 Comparson of Treynor Measures Treynor was nterpreted as a measure of performance that would apply to all nvestors regardless of ther rsk preferences. Ths ndex shows the portfolo's rsk premum '. ' - ' return and consders rsk premum return earned per unt! j of rsk.jths method assumes a completely dversfed 1 portfolo. Table 16 also presents Treynor ndex (TE) between the ten stock exchanges, Ten Composte ndex and S&P 500. Comparson of Jensen Measures Thej Jensen performance measure bascally calculates the realzed return on a securty or portfolo durng a gven tme perod and s a lnear functon of the rsk free-rate of return durng the perod. Jensen values n Table 16 shows that the stanbul stock exchange has the hghest return of 24% whle Amman has the lowest rate of 5%. Muscat, 19%, Tehran, 12%, Bulgaran, 12%, Karach, 11% Lahore,!ll% Zagreb, 9% and Caro, 9%, have all out performe'd the S&P 500. Treynor Versus Sharpe Measure For a completely dversfed portfolo, those two measures^ gve dentcal rankngs whle a poorly dversfed portfolo could have a hgh rankng on the bass of the Treynor performance measure, however a much 66

79 lower rankng on the bass of the Sharpe performance measure. Any dfference n rank would come drectly from a dfference n dversfcaton. Therefore, these two performance measures provde complementary yet dfferent nformaton. Table 17 llustrates these rankng analyses for the ten foundng stock exchanges, Ten Composte ndex and S&P 500. Snce the Dhaka and Karach stock exchanges have an dentcal rankng under two performance measures, those portfolos are consdered well dversfed portfolos compared wth other portfolos wth the rankng (Relly & Brown, 2000). Table 171 Rankngs Based on Two Performance Measures : Jll satp. /,y tp MUSCAT STANBUL CARO LAHORE ZAGREB CARO ( LAHORE MUSCAT STANBUL ZAGREB DHAKA , 7 -DHAKA, ' ; -C.071 S&P AMMAN ' '. KARACH '. o.r. ', KARACtP, COMPOSTE 0.29 S&P BULGARAN COMPOSTE TEHRAN 1.11 BULGARAN j AMMAN 1.25 TEHRAN

80 Portfolo Analyss Rsk and: Return Comparson Among sx dfferent portfolo structures, the Moderate Portfolo, consstng of 50 % of the S&P 500 and 50% of t!he Ten Composte ndex, turns out to have the hghest jrsk per unt of return or coeffcent varaton of ndex Portfolo that conssts of the Ten Compost[e ndex and the S&P 500, weghted accordng to ther majrket captalzaton has acheved the lowest coeffcent varaton' of The S&P 500 has moderately performejd and acheved coeffcent varaton of The Average portfolo and Aggressve portfolo has the same coeffcent varaton of 2.84 after Moderate Portfolo (see Table 18 for detal). Comparson of Sharpe and Treynor Measures Bas.ed on llustratons n Table 18, the Aggressve Portfolo shows the hghest return premum of 28.6% whle the S&P 500 showed the lowest return premum of 9.1%. The Average Portfolo has the second hghest return premum of 27.5% per rsk retaned. Moderate Portfolo, 25.9%, ndex Portfolo, 23.9%, and the Ten Composte Portfolo, 23.6 % have all} performed moderately compared to other sample portfolos. Treynor ndex (TE) comparsons for the four j sample portfolo structures show that Aggressve Portfolo 68

81 Table 18.' Result Analyss on Sx Portfolos Market! Annualzed Return Annualzed Rsk Rsk Per Unt of Return (Coeffcent of Varaton) MODERATE 33.00% 94.20% 2.85 AGGRESSVE 46.40% % 2.84 AVERAGE 39.80% % 2.84 S&P % 16.70% COMPOSTE 26.50% 21.70% 0.82 NDEX PORTFOLO 27.00% 22.10% Market Sharpe ndex (S) Treynor ndex (T) Jensen Measure MODERATE % AGGRESSVE % AVERAGE % S&P % 10 COMPOSTE % NDEX PORTFOLO % has the hghest return premum of 0.35 per total rsk retaned,n the portfolo. The S&P 500's performance s low and the return premum s Accordng to the coeffcent of varaton analyses n Table 18, the Moderate! Portfolo has the hghest rsk premum per unt of return, 2.85, whle the ndex Portfolo has the lowest premum of Therefore, the Moderate portfolo has the hghest rsk level to earn one unt of return. The Aggressve Portfolo and the Average Portfolo show the second closest coeffcents, 2.84, after the Moderate Portfolo. 69

82 Comparson of Jensen Measures Based on ths performance measure, the Ten Composte had the hghest return of 8.08% and Aggressve Portfolo had the lowest return of 7.6% between 1995 and The man reason why the Aggressve Portfolo had the lowest return n Jensen whle t had the hghest returns under other performance measures, s because of ths portfolos' hgher beta, whch represents the total market rsk. The hgher total rsk n the portfolo brngs down the return performance n Jensen. The ndex, Moderate, and Average portfolos also out performed the,s&p 500 (7.7%) n the Jensen performance measure, 8.07%, 7.90, and 7.8% respectvely. Treynor Versus Sharpe Measure Table 19 llustrates rankngs for the ten foundng stock exchanges, Ten Composte ndex and S&P 500. Snce all sample portfolos have dentcal rankngs n Table 19 those portfolos are consdered well-dversfed portfolos, compared wth ndvdual stock exchanges. 70

83 Table 19: Rankngs Based on Two Performance Measures 1S j 'SKnr : J' j' "j p;j./. fc-jf!,; fc'/4 ;3 j LOW S&P S&P COMPOSTE COMPOSTE NDEX PORTFOLO NDEX PORTFOLO MODERATE MODERATE Average AVERAGE HGH AGGRESSVE AGGRESSVE

84 CHAPTER SEVEN CONCLUSON Based on the nformaton n the chapter ttled Analyss\of Fndngs, the correlaton coeffcent comparson between stock portfolos would help nvestors to analyze how ndvdual portfolos affect the movement of the composte portfolo. Therefore, they can effcently dversfy ther portfolo. All ten stock exchanges and the Ten Composte ndex are postvely correlated wth the S&P The Ten Composte ndex had the hghest annualzed return of 26.5%, wth the hghest annualzed standard devaton of 75.3%. The Ten Composte ndex ancl S&P 500 tend to move same drecton. Snce the correlaton coeffcent s 0.32 between S&P 500 and Ten Composte ndex, for nstance, f S&P 500 ncreases by 10%, the Ten Composte portfolo ncreases by 3.2%. The! correlaton coeffcent matrx analyses for the ten foundng stock exchanges, S&P 500 and Ten Composte ndex suggest that the Lahore and Karach stock exchanges had the hghest postve correlaton coeffcent rato of Therefore, nvestors would beneft greatly by selectng those stock exchanges for ther portfolo. 72

85 The Annualzed return analyses summarze that the Bulgaran Stock Exchange and Ten Composte ndex show the hghest returns wth the hghest standard devatons. Snce the Treynor and Sharpe measures gve dentcal rankngs 1 for Dhaka and Karach, those stock exchanges are consdered well-dversfed portfolos. Therefore, addng Dhaka and Karach n a portfolo would help to dversfy portfolo rsk under the Treynor and Sharpe measures.! The;Jensen performance measure suggests that the stanbul,stock exchange had the hghest return of 24% whle Amman had the lowest rate of 5%. The same performance measure also shows that Muscat, Tehran, Bulgaran, Karach, Lahore, Zagreb, and Caro had hgher returns tlhan the S&P 500's return, whle the Dhaka Stock Exchange,under performed the S&P 500. The,Jensen performance measure for sample portfolos suggests that the Ten Composte performed the hghest return of 8.08% compared to other sample portfolos. The Jensen also shows that Aggressve Portfolo had the lowest return of 7.6% whle ndex Portfolo, 8.07%, Moderate Portfolo, 7.9%, and Average Portfolo, 7.8%, out performed the S&P 500, 7.7%. Accordng to Portfolo Analyses, Aggressve Portfolo performed the hghest annualzed return of 46.4% whle the 73

86 S&P 500 had the lowest return of 9.3% compared to other sample portfolos. These analyses also show that fve! sample portfolos (Aggressve, ndex, Ten Composte, Average and Moderate) and the Ten Composte out performed the S&P 500 not only based on return performances, but also based on three other performance measures. nvestng on those 5 portfolos (Aggressve, ndex, Ten Composte, Average and Moderate) are superor to nvestng n ether the ten foundng stock exchanges or the S&P 500. Fgure 3 also llustrates those comparatve analyses n a graph format. 1 By recallng the Lterature Revew for Emergng Markets Studes, due to the accessblty of the ten foundng,stock exchanges by nvestors, two forms of, nvestment nstruments would be avalable to nvestors n the Unted States -- closed-end county funds and Amercan Depostory Recepts (ADRs.) The frst nstrument, closed-end county funds, s for nvestment companes to help nvestors to nvest n portfolo assets n the ten foundng stock exchanges and sell shares of these assets n the domestc market,.e. the Unted States. 74

87 <1 tn H- cq 42 H 40 (D 38 _w n 32 0 d 30 o 1_J m H co 20 Q U 18 PJ 16 U o 14.00% ed 12.00% 10.00% CCS 8.00% (5 6.00% 3, 4.00% 2.00% 0.00% -2.00% -4.00% -6.00% -8.00% ! ! ! ! ! fl l l 9 9 7ljjfl l j f f WBXMlgy) " " ~ EM Composte 10 Composte ndex ~~~ Aggressve Average ^»^ Moder at e

88 Ths nstrument not only helps nvestors gan experence n these ten emergng markets wthout pckng ndvdual stocks n those foregn markets, but also provdes'better lqudty due to transactons executed domestcally. The second nstrument, Amercan Depostary Recepts j (ADRs), gves foregn shares the rght to be traded j dollars over U.S. exchanges or over-the-counter They are unque nstruments to solve many of the problems arsng from nvestment restrctons, nformatonal problems1 assocated wth nvestng n those ten foundng stock exchanges' securtes, as well as transacton costs (Nu & C, 2002). l 76

89 ! APPENDX A 1 FEDERATON OF EURO-ASAN STOCK EXCHANGES MEMBER EXCHANGES 77

90 FEDERATON OF EURO-ASAN STOCK EXCHANGES Amman Stock Exchange Armenan Stock Exchange Baku nterbank Currency Exchange Baku Stock Exchange Bulgaran Stock Exchange Dhaka Stock Exchange Egyptan Stock Exchange Georgan Stock Exchange stanbul Stock Exchange Karach Stock Exchange Kazakhstan Stock Exchange Kyrgyz Stock Exchange J Lahore Stock Exchange Macedonan Stock Exchange Moldavan Stock Exchange Mongolan Stock Exchange Muscat Securtes Market Palestne Securtes Exchange Tehrah Stock Exchange Trana Stock Exchange Toshkent Republcan Stock Exchange Ukranan Stock Exchange Zagreb Stock Exchange MEMBER EXCHANGES 78

91 APPENDX B 'CONSOLDATED FEAS MEMBERS 2002 STATSTCS 79

92 Consoldated FEAS Members 2002 Statstcs STOCKS _ BONDS ~ OTHER - " - j j Total Total Market Total Volume Average-Daly Volume for Average Daly s Total Volume Average Daly Cap. for 10 : Total Volume for 10 Stock Volume Total Volume 10 Stock Volume ' Total Volume for 10 Stock Volume Stock 2002 j (S Mllons) Exchanges % (S Mllons) (SMllons, Exchanges % ($ Mllons) j ($ Mllons) Exchanges % {$ Mllons) ; {Market Cap. Exchanges Jan-02 14,828 14, % ,901 10,302 95% , , % 4, , Feb-02 11,609 11, % 588 5,834 5,464 94% ,794 86, % 4, , Mar-02 5,685 5,654 99% 335 1,677 1,209 72% 95 62,546 62, % 3, , Apr-02 9,254 9, % 468 1,873 1,367 73% 96 47,011 46, % 2, , May-02 13,088 13, % 572 3,469 3,002 87% ,608 52,662 93% 2, , Jun-02 7,566 7, % 366 3,516 3,129 89% ,840 51,289 97% 2, , Jul-02 6,986 6,949 99% 319 2,476 2,017 81% ,487 42,357 75% 2, , Aug-02 5,472 5,430 99% 248 2,838 2,327 82% ,501 35,136 83% 1,932 98, Sep-02 3,963 3,924 99% 209 2,526 1,933 77% ,306 31,780 90% 1,765 85, OD O Oct-02 7,351 7,304 99% 329 3,271 2,446 75% ,021 31,163 97% 1,455 94, Nov-02 10,147 10,034 99% 467 4,398 3,206 73% ,379 25,582 82% 1, , Dec-02 7,916 7,822 99% ,135 72% ,612 3,135 12% 1, , Total 103, , ,140 39, , ,057 90% 2, FEAS Regon Market Captalzaton FEAS Regon Market Captalzaton

93 , APPENDX C HSTORCAL OVERVEW OF TEN FOUNDNG STOCK EXCHANGES 81

94 HSTORCAL OVERVEW OF TEN FOUNDNG STOCK EXCHANGES Amman Stock Exchange The Amman Fnancal Market was establshed n 1976, and started ts frst day of busness on January 1978, as a publc fnancal nsttuton wth legal, admnstratve and fnancal ndependence, operatng under the auspces of the Mnster Of Fnance. Bulgaran Stock Exchange The frst Bulgaran Stock Exchange (FBSE) was establshed on 8 November 1991 and started tradng n May n 1996, the newly establshed securtes and Stock Exchange Commsson (SSEC) ntroduced the requrement that all lusted stocks must have ther prospectuses approved by the Commsson n order to trade on the FBSE. Dhaka Stock Exchange The Dhaka Stock Exchange (DSE) was ncorporated n March 1954 as the East Pakstan Stock Exchange Assocaton Ltd. On June 1962, t was renamed the Dhaka Stock Exchange. Formal tradng began jn 1954 but was suspended when Bangladesh ganed ndependence n Wth the chan'ge n the economc polcy of the government n 1976, tradng actvtes were ultmately resumed wth nne lsted companes., Caro and the Alexandra Stock Exchange The Alexandra Stock Exchange was offcally establshed n 1888 followed by Caro n The Egyptan jstock Exchange s comprsed of two exchanges: The Caro and the Alexandra Stock Exchanges (CASE), and s governed by the same board of drectors that share the same tradng, clearng, and settlement systems. stanbul Stock Exchange n 1981, The Captal Market Law was enacted and one year later the man regulatory body The Captal Market Board was establshed. n October 1983, the Parlament approved the regulatons for the establshment and functons of Securtes Exchange, whch paved the way for the establshment of the stanbul Stock Exchange, formally ntegrated at the end of

95 Karach Stock Exchange The Karach Stock Exchange (KSE) came nto exstence on September t was later converted and regstered as a company lmted by guarantee on March Although as many as 90 members were lcensed at that tme, only half dozen were actve brokers. Lahore Stock Exchange The present Lahore Stock Exchange (LSE) was establshed n 1970 n Lahore, the provncal captal of Punjab, Pakstan under the 1969 Securtes and Exchange Ordnance Muscat Securtes Market The Muscat Securtes Market (MSM) was establshed and share tradng began n May Untl 15 January 1999 the MSM fulflled many roles: regulatng the market, organzng the exchange and actng as the central depostory. The MSM has now separated these functons nto three organzatons, each wth ts own board of drectors. Tehran Stock Exchange The dea of havng a well-organzed stock market to speed up the process of ndustralzaton of the country dates back to the 1930s when Bank Mell ran studed the market. The outbreak of WW and subsequent economc and poltcal events delayed the establshment of the TSE untl The TSE opened n Aprl ntally, only government bonds and certan state-backed certfcates were traded. Durng the 1970s, the demand for captal boosted the demand for stock. At the same tme, nsttutonal changes led to the expanson of stock market actvty. Tre restructurng of the economy followng the slamc Revoluton expanded publc sector control over the economy and reduced the need for prvate captal. At the same tme, the abolshment of nterest-bearng bonds termnated ther presence n the stock market! As a result, the TSE entered a perod of stagnaton. Ths perod ended n 1989 and snce then the TSE has expanded contnuously. Zagreb Stock Exchange The Zagreb Stock Exchange (ZSE) was ncorporated n 1991 as a jont-stock company by 25 commercal banks and nsurance companes. Today, the ZSE has 43 shareholders who n turn elect a nne-member supervsory board for a two-year term. The supervsory board apponts the Manager ojf the Exchange who s n charge of the strategc plannng and day-to-day operatons. 83

96 The ZSE currently has 39 members. Prerequstes for ZSE membershp nclude: complance wth the Securtes Law, CROSEC requrements and ZSE rules. A seat on the ZSE currently costs approxmately USS 13,000. Members are requred to comply wth the rules and regulatons of the ZSE and must regster at least one lcensed broker (FEAS Year Book, 2001/2000). j 84

97 APPENDX D MACRO ECONOMC AND MARKET NFORMATON ABOUT! TEN EMERGNG MARKETS 85

98 MACRO ECONOMC AM) MARKET NFORMATON ABOUT TEN EMERGNG MARKETS CONCEPT Amman Bulgara Dhaka varuo Alexandra stanbul Karach Lahore Muscat Tehran Zagreb ndex ASEA SOFX-50 DS E-All CASE 30 SE-100 KSE-100 LSE-100 MSM-A TEPXA CROBEXA GNP ($ Mllon) 8,340 11,995 47,106 98, ,437 61,638 61,638 14, ,031 Average nflaton (%) N/A Budget Defct (% of GDP) 0.7 N/A ,838 (273,215) (118,470) (977,241) (221,897) (221,897) (299,240) 5,518 (39,965) Unemployment Rate (%) N/A CD CTl Frst -PO Maket Y Y Y Y Y Y Y Y Y Y Secondary Market Y Y N Y Y Y Y Y N N Off-Floor Transactons Y N N N Y N N N N N Dervatves Market N N N N Y N N N N N Equty and Fxed ncome N N N N Y N N N N N Bond Market Y Y N Y Y Y Y Y N Y Stocks Y Y Y Y Y Y Y Y Y Y Mutual Funds Y N Y Y-Close Ended Y N N Y N N T-Bonds Y N N Y Y Y Y Y N Y Foregn Secrtes Y N N N Y Y Y N N N Muncpalty Bonds N Y N N Y Y Y Y N N Corporate Bonds N Y N N Y Y Y Y N N Mortgage Bonds N Y N N N N N N N N Depostory Recepts N Y N N Y N N N N N Foregn Partcpaton No restrctons No restrctons No restrctons No restrctons No restrctons No restrctons No restrctons No restrctons Restrcted No restrctons

99 APPENDX E MONTHLY PRCE NDCES FOR TEN FOUNDNG STOCK EXCHANGES 87

100 MONTHLY PRCE NDCES FOR TEN FOUNDNG STOCK EXCHANGES AMMAN BULGARAN CARO-ALEXANDRA STANBUL KARACH Date ndex Date ndex Date ndex Date ndex Date ndex Jan Jan Jan-95 3,100.0 Jan Jan-95 1,256.0 Feb? Feb Feb , Feb :0 Feb-95 1;270:0 Mar Mar Mar-95 3,170.0 Mar Mar-95 1,298.0 Apr Apr Apr-95 3,200.0 Apr Apr-95 1,350.0 May May May-95 3,260.0 May May-95 1,300.0 Jun Jun Jun-95 3,270.0 Jun Jun-95 1,345.0 Jul Jul Jul-95 3,300.0 Jul Jul-95 1,360.0 Aug Aug Aug-95 3,290.0 Aug Aug-95 1,398.0 Sep Sep Sep-95 3,260.0 Sep Sep-95 1,400.0 Oct Oct Oct-95 3,255.0 Oct Oct-95 1,450.0 Nov Nov Nov-95 3,250.0 Nov Nov-95 1,470.0 Dec Dec Dec-95 3,269.0 Dec Dec-95 1, ,497.8 Jan Jan Jan Jan Jan-96 1,503.0 Feb Feb Feb-96 4,320.0 Feb Feb-96 1,500.0 Mar Mar Mar-96 4,390.0 Mar Mar-96 1,490.0 Apr Apr Apr-96 4,400.0 Apr Apr-96 1,469.0 May May May-96 4,430.0 May May-96 1,450.0 Jun Jun Jun-96 4,500.0 Jun Jun-96 1,440.0 Jul Jul Jul-96 4,590.0 Jul Jul-96 1,430.0 Aug Aug Aug-96 4,632.0 Aug Aug;-96 1,390.0 Sep Sep-96 &7 Sep-96 4,685.3 Sep Sep-96 1,370.0 Oct Oct Oct-96 4,670.0 Oct Oct-96 1,360.0 Nov Nov Nov-96 4,650.0 Nov Nov-96 1,356.0 Dec Dec Dec-96 4,615.0 Dec Dec-96 1, , ,339.9

101 MONTHLY PRCE NDCES FOR TEN FOUNDNG STOCK EXCHANGES DHAKA LAHORE MUSCAT TEHRAN ZAGREB Date ndex Date ndex Date ndex Date ndex Date ndex Jan Jan Jan Jan-95 1,170.0 Jan-95 #NZA Feb Feb ,0 -Feb :00 Feb ;190:0 Feb-95 #N/A Mar Mar Mar Mar-95 1,210.0 Mar-95 #N/A Apr Apr Apr Apr-95 1,225.0 Apr-95 #NZA May May May May-95 1,239.0 May-95 #NZA Jun Jun Jun Jun-95 1,249.0 Jun-95 #NZA Jul Jul Jul Jul-95 1,245.0 Jul-95 #NZA Aug Aug AUg Aug-95 1,250.0 Aug-95 #NZA Sep Sep Sep Sep-95 1,260.0 Sep-95 #NZA Oct Oct Oct Oct-95 1,269.0 Oct-95 #NZA Nov: Nov Nov Nov-95 1,280.0 Nov-95 #NZA Dec Dec Dec Dec-95 1,288.1 Dec-95 #NZA , #N/A Jan Jan Jan Jan-96 1,828.0 Jan-96 #NZA Feb-96 1,000 Feb Feb Feb-96 1,840.0 Feb-96 #NZA Mar-96 1,100 Mar Mar Mar-96 1,860.0 Mar-96 #NZA Apr-96 1,300 Apr Apr Apr-96 1,850.0 Apr-96 #NZA May-96 1,450 May-96 13,1 May May-96 1,838.0 May-96 #N/A Jun-96 1,500 Jun Jun Jun-96 1,845.0 Jun-96 #NZA Jul-96 1,590 Jul Jul Jul-96 1,850.0 Jul-96 #NZA Aug-96 1,700 Aug Aug Aug-96 1,890.0 Aug-96 #NZA Sep-96 1,750 Sep Sep-9C 180 Sep-96 1,070.0 Sep-96 #NZA Oct-96 1,900 Oct Oct Oct-96 1,934.0 Oct-96 #NZA Nov-96 2,100 Nov Nov Nov-96 1,945.0 Nov-96 #NZA Dec-96 2,300 Dec Dec Dec-96 1,967.3 Dec-96 #NZA , , #NZA

102 MONTHLY PRCE NDCES FOR TEN FOUNDNG STOCK EXCHANGES AMMAN BULGARAN CARO-ALEXANDRA STANBUL KARACH Jan Jan-97 #N/A Jan Jan Jan-97 1,534.2 Feb Feb-97 #N/A Feb Feb Feb-97 1,667.1 Mar Mar-97 #N7A~ Mar " Man Mar-97 1,574:7 Apr Apr-97 #N/A Apr Apr Apr-97 1,538.8 May May-97 #N/A May May May-97 1,508.0 Jun Jun-97 #N/A Jun Jun Jun-97 1,565.7 Jul Jul-97 #N/A Jul Jul Jul-97 1,989.5 Aug Aug-97 #N/A Aug Aug Aug-97 1,744.6 Sep Sep-97 #N/A Sep Sep Sep-97 1,849.7 Oct Oct-97 #N/A Oct Oct Oct-97 1,875.0 Nov Nov-97 #N/A Nov Nov Nov-97 1,772.2 Dec Dec-97 #N/A Dec Dec Dec-97 1, #N/A , ,753.8 Jan Jan-98 #N/A Jan Jan Jan-98 1,609.2 Feb Feb-98 #N/A Feb Feb Feb-98 1,650.3 Mar Mar-98 #N/A Mar Mar Mar-98 1,553.1 Apr Apr-98 #N/A Apr Apr Apr-98 1,562.2 May May-98 #N/A May May May-98 1,040.2 Jun Jun-98 #N/A Jun Jun Jun Jul Jul-98 #N/A Jul Jul Jul Aug Aug-98 #N/A Aug Aug Aug Sep Sep-98 #N/A Sep Sep Sep Oct Oct-G8 #N/A Oct-OS Oct Oct Nov Nov-98 #N/A Nov Nov Nov-98 1,051.0 Dec Dec-98 #N/A Dec Dec Dec #N/A ,

103 MONTHLY PRCE NDCES FOR TEN FOUNDNG STOCK EXCHANGES DHAKA LAHORE MUSCAT TEHRAN ZAGREB Jan-97 1,962.0 Jan Jan-97 #N/A Jan-97 1,942.7 Jan-97 #N/A Feb-97 1,702.0 Feb Feb-97 #N/A Feb-97 1,823.4 Feb-97 #N/A Mar-97 1,199.0 Mar-97 12(2 Mar-97 #N/A Mar :8 Mar-97 #N/A Apr Apr Apr-97 #N/A Apr-97 1,916.2 Apr-97 #N/A May-97 1,217.0 May May-97 #N/A May-97 1,872.8 May-97 #N/A Jun-97 1,112.0 Jun Jun-97 #N/A Jun-97 1,859.4 Jun-97 #N/A Jul Jul Jul-97 #N/A Jul-97 1,792:9 Jul-97 #N/A Aug Aug Aug-97 #N/A Aug-97 1,681.4 Aug-97 #N/A Sep Sep Sep-97 1,225.8 Sep-97 1,643.8 Sep-97 1,225:8 Oct Oct Oct Oct-97 1,634.6 Oct Nov Nov Nov Nov-97 1,629.5 Nov Dec Dec Dec Dec-97 1,631.4 Dec-97 ' 1, , ,002,1 Jan Jan Jan Jan-98 1,646.5 Jan Feb Feb Feb-98 1,025.6 Feb-98 1,652.2 Feb-98 1,025.6 Mar Mar Mar-98 1,028.4 Mar-98 1,609.5 Mar-98 : 1,028.4 Apr Apr Apr Apr-98 1,610.4 Apr May May May May-98 1,601.8 May Jun Jun JUn Jun-98 1,604.1 Jun Jul Jul Jul Jul-98 1,557.9 Jul Aug Aug Aug Aug-98 1,517.8 Aug Sep Sep Sep Sep-98 1,533.7 Sep Oct Oct Oct Oct-98 1,566.5 Oct Nov Nov Nov Nov-98 1,560.0 Nov Dec Dec Dec Dec-98 1,531.1 Dec ,