E D H E C A L T E R N AT I V E I N D E X E S M - - PDF document

▶

Aug 25, 2022 220 likes •516 views

E D H E C R I S K A N D A S S E T M A N A G E M E N T R E S E A R C H C E N T R E E D H E C A L T E R N AT I V E I N D E X E S M A R C H 2 0 0 4 1090, ro ute de s Cr te s So phia -Antipo lis 06560 Va lb o

SLIDE 1

E D H E C A L T E R N AT I V E I N D E X E S

E D H E C R I S K A N D A S S E T M A N A G E M E N T R E S E A R C H C E N T R E M A R C H 2 0 0 4

SLIDE 2

E dhe c is o ne o f the to p five b usine ss sc ho o ls in F ra nc e o wing to the hig h q ua lity o f its a c a de mic sta ff (90 pe rma ne nt le c ture rs fro m F ra nc e a nd a b ro a d) a nd its privile g e d re la tio nship with pro fe ssio na ls tha t the sc ho o l ha s b e e n de ve lo ping sinc e its e sta b lishme nt in 1906. E dhe c Busine ss Sc ho o l ha s de c ide d to dra w o n its e xte nsive kno wle dg e o f the pro fe ssio na l e nviro nme nt a nd ha s the re fo re c o nc e ntra te d its re se a rc h o n the me s tha t sa tisfy the ne e ds o f pro fe ssio na ls. E dhe c imple me nts a n a c tive re se a rc h po lic y in the fie ld o f fina nc e . T he E dhe c Risk a nd Asse t Ma na g e me nt Re se a rc h Ce ntre c a rrie s o ut nume ro us re se a rc h pro g ra ms in the a re a s o f a sse t a llo c a tio n a nd risk ma na g e me nt in b o th the tra ditio na l a nd a lte rna tive inve stme nt unive rse s. 1090, ro ute de s Crê te s So phia -Antipo lis 06560 Va lb o nne T é l. : +33 (0)4 92 96 89 50 F a x : +33 (0)4 92 96 93 22 r e se a rc h@ e dhe c -risk.c o m

www.e dhe c - r isk.c om T his do c ume nt has b e e n pub lishe d with the suppo rt o f Alte ram

SLIDE 3

I. T HE HE T E ROGE NE IT Y AND BIASE S OF AL T E RNAT IVE INDE XE S

7 T he Bia se s o f He dg e F und I nde xe s 7 T he Pe rfo rma nc e He te ro g e ne ity o f He dg e F und I nde xe s 9

II. E DHE C AL T E RNAT IVE INDE XE S CONST RUCT ION AND MANAGE ME NT PRINCIPL E S

11 Princ ipa l Co mpo ne nt Ana lysis 11 L ist o f E dhe c Alte rna tive I nde xe s a nd the ir c o nstitue nts 12 Ba c kfilling Pro c e ss 13 Upda ting Pro c e ss 13 E dhe c I nde xe s Pre limina ry Re turns 15 E dhe c I nde x Adviso ry Bo a rd 16

III. T

HE PROPE RT IE S OF E DHE C AL T E RNAT IVE INDE XE S

17 T he Re pre se nta tivity Dime nsio n 17 T he Sta b ility o f E dhe c Alte rna tive I nde xe s 18

IV. USE

OF E DHE C INDE XE S F OR PORT F OL IO MANAGE ME NT

19 T he T ra c ka b ility o f E dhe c Alte rna tive I nde xe s 19 E dhe c Alte rna tive I nde xe s in the Stra te g ic Allo c a tio n Pro c e ss 20

APPE NDIX 1: De finitio n o f He dg e F

und Style s 22

APPE NDIX 2: Princ ipa l Co mpo ne nt Ana lysis

APPE NDIX 3: Pe rfo rma nc e o f E

dhe c Alte rna tive I nde xe s 26

T ABL E OF CONT E NT S

3

SLIDE 4

4

SLIDE 5

E XE CUT IVE SUMMAR Y I

5

Alte rna tive inve stme nt c urre ntly re pre se nts ne a rly 700 b illio n do lla rs in ma na g e d a sse ts. I nstitutio na l inve sto rs' inc re a sing inte re st in a lte rna tive inve stme nt c o nfirms the fa c t tha t it is ra pidly b e c o ming pa rt o f ma instre a m a sse t a llo c a tio n de c isio ns. At the e nd o f 2002, 52% o f the a sse ts ma na g e d b y a lte rna tive funds c a me fro m institutio na l so urc e s (c o mpa re d to o nly 19% in 1992). T he ma ssive inflo w o f c a pita l (+60% in 2000 a nd +40% in 2001) ha s thus b ro ug ht a n e nd to the re la tive ly c o nfide ntia l na ture o f a lte rna tive inve stme nt stra te g ie s, whic h c a n no lo ng e r re a so na b ly b e c o nside re d to b e a ma rg ina l a c tivity within the a sse t ma na g e me nt industry. Ma jo r a sse t ma na g e me nt ho use s e xpe c t tha t the lo ng -te rm g ro wth ra te o f a sse ts unde r ma na g e me nt b y he dg e funds sho uld b e a ro und 15%. T he re a re ho we ve r a numb e r o f o b sta c le s to the industria liza tio n o f the a lte rna tive inve stme nt industry. I ts a do ptio n b y institutio na l inve sto rs will o nly c o me a b o ut if a se rio us e ffo rt is ma de in te rms o f tra nspa re nc y a nd ra tio na liza tio n o f the inve stme nt ma na g e me nt pro c e ss a nd, a b o ve a ll, pe rfo rma nc e e va lua tio n. Due to the sc a rc ity o f info rma tio n, the lo g ic o f re pre se nta tive ne ss thro ug h ma rke t c a pita liza tio n is diffic ult to a pply to the a lte rna tive unive rse . As a re sult, finding a b e nc hma rk tha t is re pre se nta tive o f a pa rtic ula r ma na g e me nt unive rse is no t a trivia l pro b le m. T he diffe re nt inde xe s a va ila b le o n the ma rke t a re c o nstruc te d fro m diffe re nt da ta , a c c o rding to dive rse se le c tio n c rite ria a nd me tho ds o f c o nstruc tio n, a nd the y e vo lve a t diffe ring pa c e s. As a re sult o f this he te ro g e ne ity, inve sto rs c a nno t re ly o n c o mpe ting he dg e fund inde xe s to o b ta in a “true a nd fa ir” vie w o f he dg e fund pe rfo rma nc e . I nve sto rs a re the re fo re a t a lo ss whe n se le c ting b e nc hma rks. As a re spo nse to the ne e ds o f inve sto rs, the E dhe c Risk a nd Asse t Ma na g e me nt Re se a rc h Ce nte r pro po se s a n o rig ina l so lutio n b y c o nstruc ting a n “inde x o f inde xe s.” T he a im o f the me tho do lo g y use d to c o nstruc t this “inde x o f inde xe s” is to c o nstruc t a b e nc hma rk with de g re e s o f re pre se nta tivity a nd sta b ility tha t a re sig nific a ntly hig he r tha n tho se o f the inde xe s a va ila b le o n the ma rke t. T his me tho do lo g y wa s first intro duc e d in Ame nc , Ma rte llini (2003)1. T he sta tistic a l pro c e ss le a ding to the c o nstruc tio n o f the “inde x o f inde xe s” (i.e ., the E dhe c Alte rna tive I nde xe s) g ive s the m inte re sting po rtfo lio pro pe rtie s. T he se inde xe s impro ve the so undne ss o f the stra te g ic a llo c a tio n pro c e ss a nd the y c a n b e re plic a te d mo re e a sily. He nc e , the E dhe c Alte rna tive I nde xe s a re ide a l c a ndida te s to he lp inve sto rs to a llo c a te a sig nific a nt pa rt o f the ir po rtfo lio to the a lte rna tive c la ss.

L ist of E dhe c Alte r native Inde xe s and the ir Constitue nts as of Januar y 2004

1 Ame nc , N., and Mar

te llini, L ., 2003, T he Br ave Ne w Wo r ld o f He dge F und I nde xe s, Wo r king Pape r , E dhe c Risk and Asse t Manage me nt Ce ntr e . I nfo rmatio n availab le at www.e dhe c -risk.c o m

Edhec Indexes HFR CSFB EACM Altvest Hennessee Van Hedge CISDM HF Net Barclay S&P Convertible Arbitrage X X X X X X CTA Global X X X X X Distressed Securities X X X X X X X X Emerging Markets X X X X X X X X Equity Market Neutral X X X X X X X Event Driven X X X X X X X X X Fixed Income Arbitrage X X X X X X Funds of Funds X X X X X X Global Macro X X X X X X X X Long / Short Equity X X X X X Merger Arbitrage X X X X X X X Relative Value X X X X X X X Short Selling X X X X X X X X X

SLIDE 6

6

SLIDE 7

T he full fle xib ility e njo ye d b y he dg e fund ma na g e rs in te rms o f pro duc ts use d a nd stra te g ie s imple me nte d le a ds, a s we mig ht e xpe c t, to a hig h de g re e o f dive rsity in te rms o f ma na g e me nt style s. Just a s in the tra ditio na l wo rld, this dive rsity c a n a lso b e fo und a t the le ve l o f the inde x pro vide rs (he nc e fo rth re fe rre d to a s c o mpe ting inde xe s, c f. ta b le b e lo w). I t a lso a ppe a rs c le a rly tha t diffe re nt c o mpe ting inde xe s a re c o nstruc te d fro m diffe re nt da ta , a c c o rding to dive rse se le c tio n c rite ria a nd me tho ds o f c o nstruc tio n a nd tha t the y e vo lve a t diffe ring pa c e s. Suc h dive rsity po se s se rio us pro b le ms to inve sto rs. T

se e why, we first c o nside r the diffe re nt b ia se s tha t a re

c ha ra c te ristic o f he dg e funds a nd the ir impa c t o n the pe rfo rma nc e s o f the diffe re nt c o mpe ting inde xe s a va ila b le o n the ma rke t.

T able 1: L ist of He dge F und Inde x pr

vide r

s as of De c e mbe r 2003

T he Biase s of He dge F und Inde xe s

A fund’ s pa rtic ipa tio n in a da ta b a se is vo lunta ry, whic h po se s a re a l pro b le m in te rms o f the re lia b ility o f the da ta pub lishe d ("se lf re po rting b ia s"). A fund c a n in fa c t de c ide , fo r o ne re a so n o r a no the r, to re g iste r in o ne o r mo re da ta b a se s. Sinc e the funds tha t ha ve re fuse d to re po rt to o ne o r o the r o f the da ta b a se s a re , b y de finitio n, uno b se rva b le , it is no t po ssib le to e va lua te the impa c t o f this b ia s. I n a dditio n, sinc e so me re fuse to displa y the ir pe rfo rma nc e b e c a use o f po o r re sults a nd o the rs b e c a use the y ha ve a lre a dy re a c he d the ir c ritic a l size , it is e ve n diffic ult to kno w whe the r this b ia s ha s a po sitive o r ne g a tive impa c t o n the pe rfo rma nc e s a nno unc e d. T he la c k o f tra nspa re nc y a lso po se s a pro b le m in te rms o f the re lia b ility o f da ta a nd e xpo se s inve sto rs, in pa rtic ula r, to a risk o f a c ha ng e in the ma na g e r's ma na g e me nt style (this is kno wn a s "style drift", c f. L ha b ita nt - 20013). He dg e fund ma na g e rs a re unde r no o b lig a tio n to pub lish the de ta ils o f the ir po sitio ns, so it is ve ry unlike ly tha t the y will he sita te to se ize a n inve stme nt o ppo rtunity, e ve n if this me a ns mo difying the ir ma na g e me nt style te mpo ra rily (witho ut g o ing a s fa r a s de c la ring it). I t is o nc e a g a in diffic ult, o r inde e d impo ssib le , to put a fig ure o n the impa c t o f this b ia s, due to the pre va iling la c k o f tra nspa re nc y. T he vo lunta ry na ture o f the a c t pre suppo se s tha t o nly so me o f the funds will de c ide to re g iste r. Sinc e he dg e funds do no t ha ve the rig ht to a dve rtise , the fa c t tha t the y a re re c o rde d in a da ta b a se is impo rta nt in te rms o f c o mmunic a tio n, whic h is why the y ma y de c ide to re g iste r e ve n if the y do no t inte nd to g ive o ut info rma tio n o n a re g ula r b a sis. T his ma ke s it po ssib le to smo o th the re sults, a pra c tic e tha t is la rg e ly fa c ilita te d b y the c o mple xity a nd lo w le ve l o f liq uidity

f the pro duc ts ha ndle d b y the he dg e funds (i.e . a "sta le pric e s" o r "ma na g e d pric e s" pro b le m). Asne ss, K

ra il a nd L ie w (2001)4, fo r insta nc e , re ve a le d a n inc re a se in the vo la tility o f c o nve rtib le a rb itra g e re turns o f 41.5% whe n o ne switc he s fro m mo nthly da ta to q ua rte rly da ta .

T HE HE T E ROGE NE IT Y AND BIASE S OF AL T E RNAT IVE INDE XE S I

7

2 Va issié , M., 2004, Are He dg e F

unds I ndic e s Cre a te d E q ua l? , Alte rna tive I nve stme nt Qua rte rly, F

rthc o ming.

3 L

ha b ita nt, F . S., 2001, Asse ssing Ma rke t Risk fo r He dg e F unds a nd He dg e F unds Po rtfo lio s, Jo urnal o f Risk F inanc e , Spring 2001, p.1-17.

4 Asne ss, C., K

ra il, R. a nd L ie w J., 2001, Do He dg e F unds He dg e ? , Jo urnal o f Po rtfo lio Manag e me nt, F a ll 2001, Vo l.28, N°1, p.6-19. Index Provider Number of Indices Launch Date Beginning of Historical Data Web Site

Hennessee Group (Hennessee) 24 1987* 1987 hennesseegroup.com LJH Global Investments (LJH) 16 1992 1989 ljh.com Van Hedge Fund Advisors International, Inc. (Van Hedge) 16 1994* * 1988 vanhedge.com Hedge Fund Research, Inc. (HFR) 37 1994 1990 hedgefundresearch.com CISDM / MAR (CISDM) 19 1994 1990 marhedge.com HedgeFundNews.com / Bernheim Index (Bernheim) 1 1995 1999 (with monthly frequency) hedgefundnews.com/ Evaluation Associates Capital Markets, Inc. (EACM) 18 1996 1996 eacmalternative.com Hedgefund.net / Tuna Indices (HF Net) 37 1998 1976-1995* * * hedgefund.net HFIntelligence * * * * * * * * * 2002 / 2001 / 2001 / 2003 1998 hedgefundintelligence.com CSFB/Tremont Index LLC (CSFB) 14 nov-99 1994 hedgeindex.com Investorforce / Altvest (Altvest) 14 2000 1993 investorforce.com Zurich Hedge Fund (Zurich) * * * * * * 5 2001 1998 www1.zindex.com Standard & Poor’s (S&P) 10 2002 1998 spglobal.com ABN AMRO / Eurekahedge (Eurekahedge) 3 may-02 2000 eurekahedge.com MSCI Hedge Fund Indices (MSCI) More than 190 july-02 2002 msci.com Blue Chip Hedge Fund Index (Blue X) 1

ct-02

2002 bluex.org Feri Alternative Assets GmbH (Feri) 16 dec-01 2002 feri-alta.de Edhec Alternative Indices (Edhec) 13 march-03 1997 Edhec-risk.com MondoHedgeIndex (MondoHedge) 7 march-03 2002 mondohedgeindex.com Talenthedge 2

ct-03

2003 talenthedge.com Barclay Group / Global HedgeSource Hedge Fund Indices (Barclay) 18 sept-03 1997 barclaygrp.com/indices/ghs/ * In 1992 for the general public * * * * HFIntelligence / Invest-, Europe-, Asia-Hedge, Absolute Retur n * * In 1995 for the general public * * * * * 13 InvestHedge + 13 EuroHedge + 7 AsiaHedge + 12 Absolute Return * * * Depends on the strategy * * * * * * Note that Zurich has stopped maintaining its hedge fund indices since October 2003

So urc e : Va issié (2004)2

SLIDE 8

De pe nding o n the da te a t whic h the da ta b a se b e g a n, the q ua lity o f pa st info rma tio n will va ry (no ta b ly fo r funds tha t c e a se d the ir a c tivity b e fo re the da ta b a se b e g a n). T his a ffe c ts the pe rfo rma nc e o f the inde x to a g re a te r o r le sse r de g re e , de pe nding o n the numb e r o f funds tha t sto p c o mmunic a ting the ir re sults e a c h ye a r (re fe rre d to a s the a ttritio n ra te ) a nd the a ve ra g e pe rfo rma nc e diffe re ntia l o b se rve d b e twe e n tho se funds a nd the re ma ining funds. T his is kno wn a s a "survivo rship b ia s." Sinc e the HF R a nd MAR da ta b a se s b e g a n in 1994, it is like ly tha t the y will dispo se o f mo re a c c ura te info rma tio n tha n the CSF B da ta b a se (whic h o nly b e g ins in 2000) o ve r the pe rio d 1994/ 2000, a nd tha t the y will no t b e a ffe c te d in the sa me wa y b y survivo rship b ia s. F ung a nd Hsie h (2002)

5 va lue d the a ve ra g e impa c t o f

this b ia s a t 3.0%. As a c o mpa riso n, it sho uld b e no te d tha t Ma lkie l (1995)

6 e stima te d this b ia s to b e 0.5% fo r mutua l

funds. T

he va rio us da ta b a se s a re a g a in a ffe c te d in diffe re nt wa ys b y this b ia s. F

r e xa mple , the T

ASS da ta b a se ha s a hig he r survivo rship b ia s tha n the HF R da ta b a se b e c a use it ha s a hig he r a ttritio n ra te , whic h in turn is due to diffe re nt se le c tio n c rite ria fo r a dding a nd re mo ving funds. T he funds ha ve se le c tio n c rite ria tha t c a n b e ve ry dive rse , a nd the da ta pro vide d will no t b e re pre se nta tive o f the sa me ma na g e me nt unive rse . T his is re fe rre d to a s "se le c tio n b ia s." F

r insta nc e , HF

R e xc lude s ma na g e d future s fro m its da ta b a se s while T ASS a nd MAR ta ke the m into a c c o unt. Mo st funds a re pre se nt in o ne b ut no t the o the r: o f the 1,162 HF R funds a nd the 1,627 T ASS funds, o nly 465 a re c o mmo n to b o th da ta b a se s. 59% o f the funds tha t a re still in a c tivity a nd 68% o f the funds tha t no lo ng e r re po rt to HF R a re no t pa rt o f the T ASS da ta b a se (c f. L ia ng - 2000)

7. On to p
f this, inde xe s a re ra re ly re pre se nta tive o f the ir o wn da ta b a se , a s c a n b e se e n in the fo llo wing ta b le :

T able 2: Numbe r

f funds in the database s as of De c e mbe r

2003

Of the 465 funds in c o mmo n b e twe e n the HF R a nd T ASS da ta b a se s, o nly 154 (o r 33.1%) ha ve b e e n inc lude d in b o th da ta b a se s a t the sa me time . Ho we ve r, whe n a fund is a dde d to a da ta b a se , a ll o r pa rt o f its histo ric a l da ta is re c o rde d e x-po st in the da ta b a se . Sinc e the da ta b a se s a re the so le me a ns o f c o mmunic a tio n fo r mo st funds, it is re a so na b le to b e lie ve tha t the funds will o nly de c ide to pub lish the ir re sults whe n the y a re a t the ir hig he st le ve ls, in

rde r to a ttra c t a s ma ny inve sto rs a s po ssib le . I

t is the re fo re pro b a b le tha t the a ve ra g e pe rfo rma nc e s displa ye d b y the funds during the ir inc ub a tio n pe rio d will b e b e tte r tha n tho se o f funds tha t ha ve b e lo ng e d to the da ta b a se unde r c o nside ra tio n fo r a lo ng e r pe rio d. I n this c a se we ta lk a b o ut "insta nt histo ry b ia s." F ung a nd Hsie h (2002)

5 va lue d the

impa c t o f this b ia s a t 1.4 % pe r ye a r. I f the funds a re no t re c o rde d a t the sa me da te in two diffe re nt da ta b a se s, it is pro b a b le tha t the two da ta b a se s will no t b e e xpo se d to insta nt histo ry b ia s in the sa me wa y. T his risk is he ig hte ne d b y the fa c t tha t o nly 47% o f the pe rfo rma nc e s re c o rde d a re stric tly ide ntic a l.

I

8

5 F

ung , W., a nd Hsie h D. A., 2002, Be nc hma rk o f He dg e F und Pe rfo rma nc e , I nfo rma tio n Co nte nt a nd Me a sure me nt Bia se s, F inanc ial Analysts Jo urnal, Ja n/ F e b 2002, Vo l.58, N°1, p.22-34.

6 Ma lkie l, B. 1995, Re turns fro m I

nve sting in E q uity Mutua l F unds 1971 to 1991, Jo urna l o f F ina nc e , vo l. 50, no . 2 (June ): 549–572.

7 L

ia ng , B., 2000, He dg e F unds: T he L iving a nd the De a d, Jo urnal o f F inanc ial and Quantitative Analysis, Se pte mb e r 2000, Vo l.35, I ssue 3, p.309-326. Index Providers Database

Nr. of Funds in the Database
Nr. of Funds in the Indices

Van Hedge Proprietary Database 5 400 1 300 Feri

Propr. Data. + Other Avail. Data. (Van Hedge, TASS, HF Net)

5 000 41 Hennessee Proprietary Database 3 500 690 S&P Proprietary Database + Other Available Databases 3 500 40 CSFB TASS Database and Tremont Database 3 300 431 HFIntelligence Proprietary Database 3 202 2 652 Altvest Proprietary Database 2 600 All the funds Barclay Global HedgeSource 2 450 All the funds HF Net Proprietary Database 2 300 All the funds HFR Proprietary Database 2 300 1 400 CISDM Proprietary Database 2 300 1 600 MSCI Proprietary Database 1 800 1 500 Bernheim U.S. Offshore Funds Directory 900 18 Zurich ZCM + Other Available Databases 900 60 LJH Proprietary Database 800 All the funds Edhec Main hedge fund indices available on the market n.a. n.a. MondoHedge Proprietary Database 720 48 Blue X Proprietary Database 350 - 400 30-40 Eurekahedge Proprietary Database 365* 110 EACM Proprietary Database 100 100 Talenthedge Proprietary Database Not communicated 5 to 20 per index

So urc e : Vaissié (2004)2

SLIDE 9

Sinc e the va rio us inde xe s ha ve diffe re nt c o mpo sitio ns, the y will no t b e a ffe c te d in the sa me wa y b y the diffe re nt b ia se s to whic h we ha ve re fe rre d. T his c o uld, in re le va nt c a se s, disto rt the a na lysis o f the ir pe rfo rma nc e . T he fre q ue nc y with whic h a n inde x is mo difie d will a lso pla y a n impo rta nt ro le in the pe rfo rma nc e s o f the diffe re nt inde xe s. T he HF R inde xe s a re e q ua lly-we ig hte d a nd re b a la nc e d o n a mo nthly b a sis. As a re sult, the y fo llo w a c o ntra ria n stra te g y. T he CSF B inde xe s, fo r the ir pa rt, a re va lue we ig hte d a nd re b a la nc e d q ua rte rly a nd the re fo re fo llo w a mo me ntum type stra te g y. As F ung a nd Hsie h ha ve sho wn, the se diffe re nc e s a lo ne e xpla in a pe rfo rma nc e diffe re ntia l o f 7.4% b e twe e n the two da ta b a se s in 1999! He re we c a n spe a k o f a "re b a la nc ing sc he me b ia s."

T he Pe r for manc e He te r

ge ne ity of He dge F

und Inde xe s

Sig nific a nt pe rfo rma nc e diffe re nc e s fo r the sa me style a re c o mmo nly o b se rve d b e twe e n the diffe re nt c o mpe ting inde xe s. T his phe no me no n is pa rtic ula rly no tic e a b le in pe rio ds o f c risis (b e twe e n Aug ust 1998 a nd Oc to b e r 1998, c f. ta b le b e lo w). T he he te ro g e ne ity o f the info rma tio n supplie d b y the diffe re nt inde x pro vide rs is a c tua lly spe c ta c ula r . Mo re tha n 20% se pa ra te s the pe rfo rma nc e s o f the Zuric h a nd E ACM L

ng / Sho rt E

q uity inde xe s in F e b rua ry 2000 (c f. Ame nc a nd Ma rte llini – 2003)

1. An a na lysis o f the me a n a nd me dia n c o rre la tio ns b e twe e n the pe rfo rma nc e s o f the

diffe re nt c o mpe ting inde xe s c o nfirms the la c k o f ho mo g e ne ity. T he me a n c o rre la tio n b e twe e n c o mpe ting inde xe s fo r the sa me type o f stra te g y (E q uity Ma rke t Ne utra l: 0.43, L

ng / Sho rt E

q uity: 0.46) c a n b e lo we r tha n 0.5. T he inc re a sing numb e r o f inde x pro vide rs a nd c o nstruc tio n me tho ds po se s the pro b le m o f the he te ro g e ne ity o f the da ta . I t a ppe a rs c le a rly tha t the c o mpe ting he dg e fund inde xe s do no t to da y pro vide re pre se nta tivity a nd sta b ility c o nditio ns tha t wo uld a llo w inve sto rs a ho mo g e no us a nd re le va nt o ve rvie w o f a lte rna tive funds.

T able 3: Maximum Re tur n Diffe r e nc e s by Inve stme nt Style as of De c e mbe r 2003

With the he lp o f a simple he te ro g e ne ity indic a to r, we c a n a tte mpt to e va lua te the de g re e o f he te ro g e ne ity o f the diffe re nt stra te g ie s. T his indic a to r is de no te d HI a nd is c a lc ula te d a s fo llo ws:

HI = 1 – Ave ra g e Co rre la tio n

T he re fo re , a pe rfe c tly he te ro g e ne o us situa tio n fo r the inde xe s is tra nsla te d b y HI =1. Suc h a n a na lysis re ve a ls fo r e xa mple tha t the E q uity Ma rke t Ne utra l, F ixe d I nc o me Arb itra g e a nd Re la tive Va lue stra te g ie s a re pa rtic ula rly he te ro g e ne o us. T he y e xhib it re spe c tive ly, fo r the te st pre se nte d a b o ve , de g re e s o f h e t e r

g e ne ity o f 55.08%, 42.93% a nd 42.23%. Co nve rse ly, the E

me rg ing Ma rke ts a nd CT A Glo b a l stra te g ie s re spe c tive ly pre se nt de g re e s o f he te ro g e ne ity o f o nly 7.37% a nd 5.64%. T he se re sults c o nfirm tho se o b ta ine d thro ug h the c o rre la tio n a na lysis.

I

9

Investment Styles Max differences (with dates and indexes) Convertible Arbitrage 7,55% (Dec 01: EACM (-6.93%) / Hennessee (0.62%)) CTA Global 7,50% (Dec 00: Barclay ( 6.00%) / S&P (13.50%)) Distressed Securities 7,75% (Aug 98: CSFB (-12.45%) / Van Hedge (-4.70%)) Emerging Markets 19,45% (Aug 98: CISDM (-26.65%) / Altvest (-7.20%)) Equity Market Neutral 5,00% (Dec 99: Hennessee (0.20%) / Van hedge (5.20%)) Event Driven 5,37% (Aug 98: CSFB (-11.77%) / S&P (-6.40%)) Fixed Income Arbitrage 10,09% (Oct 98: HF Net (-9.89%) / Van Hedge (0.20%)) Funds of Funds 8,01% (Dec 99: CISDM (2.41%) / Altvest (10.42%)) Global Macro 14,17% (Oct 98: CSFB (-11.55%) / Altvest (2.62%)) Long/Short Equity 9,51% (Feb 00: Altvest (3.50%) / CSFB (13.01%)) Merger Arbitrage 3,18% (Jan 99: CSFB (-1.51%) / Altvest (1.67%)) Relative Value 10,50% (Oct 98: S&P (-6.90%) / Van Hedge (3.60%)) Short Selling 21,20%(Feb 00: Van Hedge (-24.30%) / EACM (-3.10%))

Source: Edhec Risk From January 1998 through December 2003

SLIDE 10

do c ume nt the he te ro g e ne ity o f the diffe re nt inde xe s, o ne ma y a lso hig hlig ht the ir c o ntra ste d e xpo sure s to a

va rio us risk fa c to rs. F

r e xa mple , HF

R fixe d-inc o me a rb itra g e inde x is sig nific a ntly mo re e xpo se d to liq uidity risk a s pro xie d b y c ha ng e s in tra ding vo lume o n the NYSE (c o rre la tio n c o e ffic ie nt = –0.18) tha n the c o rre spo nding HF Ne t inde x (c o rre la tio n c o e ffic ie nt = –0.01). I n the sa me wa y, the c o rre la tio n c o e ffic ie nt o f the pe rfo rma nc e s o f the Va n He dg e inde x with the re turns o f the S&P500 is 0.53, c o mpa re d to 0.00 fo r the CSF B inde x! F ina lly, the e xpo sure o f the Va n He dg e inde x to vo la tility risk is re la tive ly hig h (–0.47), while tha t o f the HF R inde x is sig nific a ntly lo we r (0.14). I t a g a in a ppe a rs c le a rly tha t diffe re nt c o mpe ting inde xe s a re ve ry he te ro g e ne o us. T a king the diffe re ntia te d e xpo sure s to va rio us risk fa c to rs into a c c o unt, it wo uld no t b e the sa me to ta ke o ne o r o the r o f the c o mpe ting inde xe s a s a b e nc hma rk. So whic h o ne sho uld b e c ho se n?

T able 5: Se nsivity to Br

ad- base d Risk F

ac tor s - T he Case of F ixe d Inc ome Ar bitr age

T he data use d to c har ac te r ise the diffe r e nt so ur c e s o f r isk is as fo llo ws: T he marke t risk is me asure d b y the e vo lutio n o f the S&P 500 Pric e I nde x T he vo latility risk is me asure d b y the re lative pric e variatio ns o f the VXO (i.e ., e x VI X) c o ntrac t, the unde rlying o f whic h is the implic it vo latility o f the S&P 100 T he inte re st rate risk is me asure d b y the variatio ns o f the rate o f re turn o f the 3-mo nth T re asury b ill T he slo pe o f the yie ld c urve is o b taine d b y c alc ulating the diffe re nc e b e twe e n the rate o f re turn o f a b o nd with a 30-ye ar maturity and that o f a 3-mo nth Tr e asury b ill T he c urre nc y risk is me asure d b y the e vo lutio n o f the e xc hang e rate o f the US do llar c o mpare d to a b aske t o f fo re ig n c urre nc ie s T he c o mmo dity risk is me asure d b y the re lative pric e variatio ns o f a b arre l o f c rude o il T he c re dit risk is me asure d b y the re lative variatio ns o f the diffe re ntial b e twe e n the re turns o n b o nds rate d Baa and Aaa b y Mo o dy’ s T he liq uidity risk is me asure d b y the e vo lutio n o f the vo lume o f se c uritie s e xc hang e d o n the NYSE

Just a s in the tra ditio na l wo rld, no ne o f the c o mpe ting inde xe s a re e ithe r c o lle c tive ly e xha ustive o r mutua lly e xc lusive . T he la c k o f re g ula tio n a nd its c o ro lla ry, a la c k o f tra nspa re nc y, a c c e ntua te the pro b le m dra ma tic a lly in the he dg e fund unive rse . I t is the re fo re ne c e ssa ry to c o nstruc t style inde xe s tha t a llo w fo r a re spo nse to the ne e ds o f pra c titio ne rs in te rms o f tra nspa re nc y a nd re lia b ility. T he E dhe c Alte rna tive I nde xe s pre se nte d in the fo llo wing se c tio n ha ve b e e n c o nstruc te d a c c o rding ly to me e t tho se e xpe c ta tio ns.

I

10

Fixed Income Arbitrage Market Risk Volatility Risk Interest Rate Risk Slope of the Yield Curve Currency Risk Commodity Risk Credit Risk Liquidity Risk CSFB 0.00 0.12 0.15 0.23 0.42 0.05

0.38
0.10

HFR

0.16

0.14 0.25 0.19 0.57 0.07

0.24
0.18

Van Hedge 0.53

0.47

0.09 0.02

0.13

0.14

0.16
0.05

Hennessee 0.37

0.37

0.06 O.19 0.26 0.12

0.22
0.12

HF Net

0.10

0.20 0.22 0.20 0.42 0.03

0.37
0.01

Source: Amenc, Martellini (2002a) From January 1998 through December 2000

T able 4: He te r

ge ne ity Inde x by Inve stme nt Style as of De c e mbe r

2003

So urc e : Ame nc , Marte llini (2003) 1

Investment Styles HI Convertible Arbitrage 16,28% CTA Global 5,64% Distressed Securities 13,77% Emerging Markets 7,37% Equity Market Neutral 55,08% Event Driven 8,36% Fixed Income Arbitrage 42,93% Funds of Funds 8,50% Global Macro 28,96% Long/Short Equity 15,51% Merger Arbitrage 12,23% Relative Value 42,23% Short Selling 11,28%

Source: Edhec Risk From January 1998 through December 2003

SLIDE 11

T he lo g ic o f re pre se nta tivity thro ug h ma rke t c a pita liza tio n is diffic ult to a pply to the a lte rna tive unive rse . As a re sult, finding a b e nc hma rk tha t is re pre se nta tive o f a pa rtic ula r ma na g e me nt unive rse is no t a trivia l pro b le m.

Pr inc ipal Compone nt Analysis

Give n tha t it is impo ssib le to c o me up with a n o b je c tive judg me nt o n wha t is the b e st e xisting inde x, a na tura l ide a c o nsists o f using so me c o mb ina tio n o f c o mpe ting inde xe s to re a c h a b e tte r unde rsta nding o f wha t the c o mmo n info rma tio n a b o ut a g ive n inve stme nt style wo uld b e . One stra ig htfo rwa rd me tho d fo r o b ta ining a c o mpo site inde x b a se d o n va rio us c o mpe ting inde xe s wo uld invo lve c o mputing a n e q ua lly-we ig hte d po rtfo lio o f a ll c o mpe ting inde xe s. T his wo uld o b vio usly pro vide inve sto rs with a c o nve nie nt o ne -dime nsio na l summa ry o f the c o ntra ste d info rma tio n c o nta ine d in c o mpe ting inde xe s. I n pa rtic ula r, b e c a use c o mpe ting he dg e fund inde xe s a re b a se d o n diffe re nt se ts o f he dg e funds, the re sulting po rtfo lio o f inde xe s wo uld b e mo re e xha ustive tha n a ny o f the c o mpe ting inde xe s it is e xtra c te d fro m. We c a n push the lo g ic o ne ste p furthe r a nd sug g e st using fa c to r a na lysis te c hniq ue s to e xtra c t the b e st po ssib le o ne -dime nsio na l summa ry o f a se t o f c o mpe ting inde xe s, a nd de sig n wha t c a n b e c a lle d “pure style ” inde xe s. Our me tho d is a na tura l g e ne ra liza tio n o f the ide a o f ta king a po rtfo lio o f c o mpe ting inde xe s. T he re fine me nt invo lve s re la xing the a ssumptio n o f a n e q ua lly-we ig hte d po rtfo lio . We sug g e st using fa c to r a na lysis te c hniq ue s to g e ne ra te a se t o f a lte rna tive inde xe s tha t c a n b e tho ug ht o f a s the b e st po ssib le o ne -dime nsio na l summa rie s o f info rma tio n c o nve ye d b y c o mpe ting inde xe s fo r a g ive n style , in the se nse o f the la rg e st fra c tio n o f the va ria nc e e xpla ine d. He re , we a re lo o king fo r the po rtfo lio we ig hts tha t ma ke the c o mb ina tio n o f c o mpe ting inde xe s c a pture the la rg e st po ssib le fra c tio n o f the info rma tio n c o nta ine d in the da ta fro m the va rio us c o mpe ting inde xe s. T e c hnic a lly spe a king , this a mo unts to using the first c o mpo ne nt o f a Princ ipa l Co mpo ne nt Ana lysis (se e Appe ndix 2 fo r mo re de ta ils o n this me tho d, whic h is he nc e fo rth re fe rre d to a s PCA) o f c o mpe ting inde xe s a s a c a ndida te fo r a pure style inde x. No te tha t the first c o mpo ne nt typic a lly c a pture s a la rg e pro po rtio n o f c ro ss-se c tio na l va ria tio ns b e c a use c o mpe ting style s te nd to b e a t le a st so me wha t po sitive ly c o rre la te d. T he PCA o f a time -se rie s invo lve s studying the c o rre la tio n ma trix o f suc c e ssive sho c ks. I ts purpo se is to e xpla in the b e ha vio r o f o b se rve d va ria b le s using a sma lle r se t o f uno b se rve d implie d va ria b le s. F ro m a ma the ma tic a l sta ndpo int, it invo lve s tra nsfo rming a se t o f K c o rre la te d va ria b le s into a se t o f o rtho g o na l va ria b le s, o r implic it fa c to rs, whic h re pro duc e s the o rig ina l info rma tio n pre se nt in the c o rre la tio n struc ture . E a c h implic it fa c to r is de fine d a s a line a r c o mb ina tio n o f o rig ina l va ria b le s. T he E dhe c Alte rna tive I nde xe s a re a b le to c a pture a ve ry la rg e fra c tio n o f the info rma tio n. T he a ve ra g e pe rc e nta g e

f va ria nc e e xpla ine d b y the E

dhe c Alte rna tive I nde xe s is 79.12% (a nd the me dia n pe rc e nta g e o f va ria nc e is 81.12%) a c ro ss a ll sub -unive rse s. T he pe rc e nta g e o f va ria nc e e xpla ine d b y the E dhe c Alte rna tive I nde xe s is, o f c o urse , a ll the mo re sig nific a nt in tha t the c o rre la tio n b e twe e n the c o mpe ting inde xe s is hig h. F

r e xa mple , e me rg ing ma rke t style

inde xe s ha ve a pe rc e nta g e o f va ria nc e e xpla ine d tha t is g re a te r tha n 90% fro m a po pula tio n o f 7 c o mpe ting inde xe s. T he me a n c o rre la tio n wa s a lmo st 0.93 fo r e me rg ing ma rke t inde xe s. I n the sa me ve in, the E dhe c E ve nt Drive n a nd Me rg e r Arb itra g e I nde xe s c a pture mo re tha n 80% o f the info rma tio n o rig ina lly a va ila b le in a se t o f 8 a nd 4 c o mpe ting inde xe s, re spe c tive ly. T he E dhe c F und o f F unds I nde x a lso e njo ys ve ry lo w info rma tio n lo ss a s mo re tha n 91% o f the info rma tio n is c a pture d b y the o ne -dime nsio na l summa ry. On the o the r ha nd, the pe rc e nta g e o f info rma tio n lo ss is hig he r in the c a se o f e q uity ma rke t ne utra l (41.09% = 100% - 58.91% info rma tio n lo ss) a nd fixe d-inc o me a rb itra g e (35% = 100% - 65% info rma tio n lo ss). T his is b e c a use the se stra te g ie s we re the o ne s fo r whic h the he te ro g e ne ity o f info rma tio n pro vide d b y c o mpe ting inde x pro vide rs wa s the mo st e xtre me (se e Ame nc a nd Ma rte llini (2003), fo r furthe r e xpla na tio n)

E dhe c Alte rna tive I nde xe s g e ne ra te d a s the first c o mpo ne nt in a fa c to r a na lysis ha ve a n a ppe a ling b uilt-in e le me nt

f o ptima lity, sinc e the re is no o the r line a r c o mb ina tio n o f c o mpe ting inde xe s tha t implie s a lo we r info rma tio n lo ss.

E DHE C AL T E RNAT IVE INDE XE S CONST RUCT ION AND MANAGE ME NTPRINCIPL E S II

11

SLIDE 12

E dhe c Alte r native Inde xe s and the ir Constitue nts

I n o rde r to e sta b lish the list o f E dhe c Alte rna tive I nde xe s we first e limina te d a lte rna tive stra te g ie s fo r whic h fe we r tha n 4 c o mpe tito rs we re a va ila b le . F

r sta tistic a l purpo se s it is e sse ntia l to ha ve a t le a st 4 c o mpe ting inde xe s fo r it to ma ke

se nse to c o nstruc t a n inde x o f inde xe s. We the n e limina te d stra te g ie s with a na rro w fo c us (e .g ., se c to rs - he a lth c a re ) to c o nc e ntra te o n po pula r stra te g ie s. As a re sult o f tha t se le c tio n, we we re le ft with a list o f 13 inve stme nt style s with 4 to 9 inde x pro vide rs fo r e a c h style (se e ta b le b e lo w). T he se c o nd ste p invo lve s se le c ting the inde xe s to b e inc lude d in the E dhe c Alte rna tive I nde x fo r e a c h stra te g y. T he Ed he c Alte rna tive I nd e xe s a re re q u i re d to e njo y c o mp le te ly tra nsp a re nt c o nstruc tio n me tho d o lo g y a nd ma na g e me nt princ iple s. As a re sult, the se le c te d inde xe s must b e pub lic ly a va ila b le a nd ha ve tra nspa re nt style c la ssific a tio n (i.e ., we ll de fine d sub -unive rse s) a nd c o nstruc tio n me tho do lo g ie s, so tha t o ne c a n e a sily c he c k the pe rfo rma nc e s o f the E dhe c Alte rna tive I nde xe s. F ina lly, the inde xe s must b e b a se d o n a b ro a d da ta b a se (to e nsure a minimum de g re e o f re pre se nta tivity) a nd po st the ir pe rfo rma nc e s o n time (se e ne xt se c tio n fo r mo re de ta ils). We ha ve ide ntifie d the 10 inde x pro vide rs liste d in the fo llo wing ta b le a nd inc lude d tho se tha t fulfille d the a fo re me ntio ne d re q uire me nts in the c o mpo sitio n o f the E dhe c Alte rna tive I nde x. Xs indic a te tha t the inde x is inc lude d in the E dhe c Alte rna tive I nde x. I t sho uld b e no te d tha t the se c o mpo sitio ns a re tho se e sta b lishe d a s o f Ja nua ry 2004 a nd a re sub je c t to re visio n b y the E dhe c Alte rna tive I nde x a dviso ry b o a rd (se e ne xt se c tio n fo r mo re de ta ils). I t sho uld b e no te d tha t the list o f E dhe c Alte rna tive I nde xe s wa s a lso e sta b lishe d to e na b le diffe re nt le ve ls o f a na lysis. F

r e xa mple , b ro a d style inde xe s suc h a s the E

dhe c Re la tive Va lue I nde x o r the E dhe c E ve nt Drive n I nde x a re pe rfe c t fo r a na lysis a t a g lo b a l le ve l. On the o the r ha nd, the E dhe c Co nve rtib le Arb itra g e / E q uity Ma rke t Ne utra l/ F ixe d I nc o me Arb itra g e I nde xe s o r the E dhe c Distre sse d Se c uritie s/ Me rg e r Arb itra g e I nde xe s a re ide a l fo r mo re de ta ile d a na lysis o f tho se two b ro a d c a te g o rie s.

T able 6 : L ist of E dhe c Alte r native Inde xe s and the ir Constitue nts as of Januar y 2004

II

12

I nfo rmatio n availab le at www.e dhe c -risk.c o m

SLIDE 13

Bac kfilling Pr

c e ss

T he histo ric al data pe rio d fo r the E dhe c Alte rnative I nde xe s b e g an o ffic ially in January 2003. Ho we ve r, in o rde r to use a lo ng e r rang e o f data, “b ac kfilling ” was c arrie d o ut as fo llo ws: As a ll the c o mpe ting inde xe s did no t ha ve a suffic ie nt le ng th o f histo ric a l da ta , we o nly se le c te d tho se tha t ha ve pub lishe d mo nthly pe rfo rma nc e s sinc e Ja nua ry 1994. T his a llo ws us, ta king into a c c o unt the thre e ye a rs re q uire d fo r the c a lib ra tio n o f the princ ipa l c o mpo ne nt a na lysis, to use mo nthly pe rfo rma nc e s fro m Ja nua ry 1997 o nwa r ds. We the re fo re stric tly o b se rve the me tho d de sc rib e d in Ame nc a nd Ma rte llini (2003)

1 whilst a t the sa me time limiting

urse lve s to the fo llo wing inde xe s: Altve st, CSF

B/ T re mo nt, HF Ne t, HF R a nd MAR. F ro m Ja nua ry 2001 o nwa rds, the E dhe c Alte rna tive I nde xe s a re c o nstruc te d using the sa me me tho d, whilst this time inc o rpo ra ting fo ur a dditio na l inde xe s in the ir c o mpo sitio n: E ACM, He nne sse e , Va n He dg e a nd Zuric h.

Updating Pr

c e ss for

Alte r native Inde xe s

T he princ iple fo r updating the pe rfo rmanc e s o f the E dhe c Alte rnative I nde xe s fro m January 1997 to De c e mb e r 2002 is the fo llo wing : T he c o mpo sitio n o f diffe re nt E dhe c Alte rna tive I nde xe s is c a lc ula te d e ve ry thre e mo nths b a se d o n the histo ric a l pe rfo rma nc e da ta (thre e ye a rs) o f the se le c te d c o mpe ting inde xe s. E a c h mo nth the se suc c e ssive we ig hting s a re a pplie d to the pe rfo rma nc e histo ry o f the inde xe s c o nc e rne d. Ho we ve r, in o rde r to e nsure a c e rta in le ve l o f ho mo g e ne ity in the c a lc ula tio n o f the pe rfo rma nc e histo ry o f the diffe re nt Alte rna tive I nde xe s, the pe rfo rma nc e histo ry o f the c o mpe ting inde xe s pub lishe d in De c e mb e r 2002 will b e use d to c o nstitute the e ntire histo ric a l pe rfo rma nc e da ta o f the Alte rna tive I nde xe s. T his will a llo w us to a pply the sa me c a lc ula tio n rule to a ll the inde xe s, fo r a ll da te s inc lude d b e twe e n Ja nua ry 1997 a nd De c e mb e r 2002

T he princ iple fo r updating the pe rfo rmanc e o f the E dhe c Alte rnative I nde xe s fro m January 2003 o nwards is the fo llo wing : T he we ig hting c a lc ula te d thro ug h PCA a t the b e g inning o f e a c h q ua rte r will b e a pplie d to the pe rfo rma nc e o f the diffe re nt c o mpe ting inde xe s, a nd the re sult o b ta ine d will b e le ft a s suc h, whe the r the c o mpe ting inde xe s c ha ng e the ir pe rfo rma nc e histo ry o r no t

10.

T he upda ting o f the E dhe c Alte rna tive I nde xe s, to g e the r with the risk a nd pe rfo rma nc e a na lysis fo r the M(o nth), will b e pub lishe d o n the www.e dhe c -risk.c o m we b site o n the third wo rking da y o f M(o nth) + 2. T his time -sc a le will e na b le us to use , in no rma l o pe ra ting c o nditio ns, the la te st pe rfo rma nc e o f the diffe re nt c o mpe ting inde xe s.

II

13

9 Give n tha t the diffe re nt c o mpe ting inde xe s ha ve fund se le c tio n c rite ria whic h va ry little o ve r time , it is like ly tha t the funds inte g ra te d

b e twe e n Ja nua ry 1997 a nd De c e mb e r 2002 will ha ve a simila r pro file to tha t o f the funds whic h a re a lre a dy pre se nt in the da ta b a se in q ue stio n b e fo re this pe rio d. T his is e q ua lly true fo r a ll the inde xe s a nd the re fo re the e x-po st c a lc ula te d we ig hting s a re ide ntic a l to tho se whic h wo uld ha ve b e e n c a lc ula te d e x-a nte , sinc e the o ve ra ll info rma tio n c o nta ine d in the diffe re nt inde xe s re ma ins unc ha ng e d.

10 So me o f the diffe re nt c o mpe ting inde xe s inte g ra te the histo ric a l pe rfo rma nc e o f the funds, whic h the y a dd to the ir da ta b a se . T

he a dva nta g e o f suc h a pra c tic e is to impro ve the re pre se nta tivity o f the da ta b a se e x-po st, b ut this a lso pre se nts a ma jo r disa dva nta g e sinc e it implie s tha t the histo ric a l pe rfo rma nc e o f the inde xe s c o nc e rne d must b e c o ntinuo usly mo difie d. T he c o nstruc tio n o f E dhe c Alte rna tive I nde xe s b y PCA a llo ws us, pre c ise ly, to re spo nd e x-a nte , a nd no t e x-po st, to the pro b le m o f re pre se nta tivity, whic h a llo ws us to “fre e ze ” the histo ry.

SLIDE 14

I f, fo r a ny pa rtic ula r re a so n, o ne o f the c o mpe ting inde xe s usua lly use d to c o nstitute the E dhe c Alte rna tive I nde x do e s no t upda te its pe rfo rma nc e b e fo re the se c o nd wo rking da y o f M(o nth)+2, we will b e una b le to c a lc ula te the ne w we ig hting o f the pure style inde x in time fo r the third wo rking da y. I n this c a se , we will use a o ne -o ff me a sure e xc luding the inde x c o nc e rne d fro m the c o mpo sitio n o f the E dhe c Alte rna tive I nde x until the fo llo wing re b a la nc ing (i.e ., a t the b e g inning o f the fo llo wing q ua rte r). I f this situa tio n a ro se in the middle o f the q ua rte r, we wo uld the n b e o b lig e d to unde rta ke a o ne -o ff re -c a lc ula tio n o f the ne w c o mpo sitio n o f the pure style inde x fo r the e nd o f the c urre nt q ua rte r , e xc luding the missing inde x.

II

14

SLIDE 15

II

15 E dhe c Inde xe s’ Pr e liminar y Re tur ns

He dg e F unds usua lly o nly disc lo se the ir pe rfo rma nc e a fe w we e ks a fte r mo nth e nd. As a re sult, he dg e fund inde xe s te nd to pub lish the ir pe rfo rma nc e re la tive ly la te , whic h ma y le a ve pra c titio ne rs a t a lo ss. T

c o mpe nsa te fo r this la c k
f time line ss, so me inde x pro vide rs ha ve de c ide d to po st pre limina ry re turns o n the ir we b site . I

n re spo nse to p r a c t i t i

e r s’ ne e ds, the E dhe c Risk a nd Asse t Ma na g e me nt Re se a rc h Ce ntre will re le a se the pre l i m i n a r y pe rfo rma nc e o f its Alte rna tive I nde xe s o n the 17th c a le nda r da y o f M(o nth) + 1. I f a ll the inde xe s e nte ring into the c o mpo sitio n o f the E dhe c Alte rna tive I nde x pub lish pre limina ry re sults: > T he pre limina ry re turn o f the E dhe c Alte rna tive I nde x c a n b e o b ta ine d simply b y a pplying the c urre nt inde x we ig hting to the pre limina ry re sults o f its c o nstitue nts. I f o ne o r se ve ra l inde xe s e nte ring into the c o mpo sitio n o f the E dhe c Alte rna tive I nde x do no t pub lish pre limina ry re sults: > T he we ig hts o f the “missing ” inde xe s will b e a llo c a te d pro po rtio na lly to the o the r inde xe s, a c c o rding to the ir initia l we ig ht. T he pre limina ry re sults o f the E dhe c Alte rna tive I nde x c a n the n b e o b ta ine d b y a pplying the mo difie d inde x we ig hting to the pre limina ry re sults. I t sho uld b e no te d tha t pre limina ry re turns a re o nly me a nt to b e use d a s a n initia l e stima tio n o f the inde xe s’ pe rfo rma nc e . T he fina l re turns o f the E dhe c Alte rna tive I nde xe s will b e pub lishe d o n the we b site o n the 3rd wo rking da y o f M(o nth) + 2, o nc e a ll the inde xe s e nte ring into the ir c o mpo sitio n ha ve pub lishe d the ir fina l re sults. Only the se re turns will c o nstitute the o ffic ia l histo ric pe rfo rma nc e o f the E dhe c Alte rna tive I nde xe s.

E dhe c Inde x Public ation Date s

SLIDE 16

E dhe c Inde x Advisor y Boar d

T he E dhe c I nde x Adviso ry Bo a rd de te rmine s the inc lusio n o r e xc lusio n o f a n inde x in the c a lc ula tio n o f the E dhe c Alte rna tive I nde xe s. T he c rite ria upo n whic h the c o mmitte e b a se s its de c isio ns a re a s fo llo ws: > T he a va ila b le histo ry > T he c la rity o f its c o nstruc tio n me tho d > I ts re pre se nta tivity in te rms o f b e ing a re fe re nc e inde x fo r ma na g e rs a nd/ o r inve sto rs a s we ll a s whe the r it ta ke s e xisting funds into a c c o unt > T he c o mple te ne ss o f the pro vide r’ s inde xe s > T he sta b ility o f the c o mpo sitio n > T he re g ula rity with whic h the da ta / inde x is pub lishe d T his c o mmitte e is c o mpo se d o f 9 spe c ia list pro fe ssio na ls a nd re c o g nize d re se a rc he rs in the do ma in o f a lte rna tive inve stme nt.

II

16

SLIDE 17

We ha ve te ste d the q ua litie s o f the E dhe c Alte rna tive I nde xe s. We first a sse ss the de g re e o f re pre se nta tivity o f the E dhe c Alte rna tive I nde xe s to sho w the e xte nt to whic h the y a re a b le to c a pture the sub sta nc e o f the a lte rna tive inve stme nt stra te g ie s. We the n a sse ss the impa c t o f sho rt-te rm ma rke t o r e c o no mic tre nds o n the ir c o mpo sitio n.

T he Re pr e se ntativity Dime nsion

te st the re pre se nta tivity q ua litie s o f the E

dhe c Alte rna tive I nde xe s, we ha ve pro c e e de d in the fo llo wing ma nne r: we ha ve c o nstitute d a n e q ua lly-we ig hte d po rtfo lio fo r e a c h o f the stra te g ie s fro m a pro prie ta ry da ta b a se ma de up

f 7,422 funds (o f whic h 2,317 a re no t re c o rde d in a ny da ta b a se ). T

he po rtfo lio s the re fo re c o nta in mo re tha n 600 funds o n a ve ra g e , a nd a s a re sult we re c o nside re d to b e re la tive ly re pre se nta tive o f the ir ma na g e me nt unive rse . We the n c a lc ula te d the c o rre la tio n c o e ffic ie nt o f tho se po rtfo lio s with the E dhe c Alte rna tive I nde xe s o ve r the pe rio d fro m Ja nua ry 1998 thro ug h De c e mb e r 2000. T he hig he r the c o e ffic ie nt, the mo re re pre se nta tive the inde x is. We the n c o mpa re d this c o rre la tio n c o e ffic ie nt with the a ve ra g e c o rre la tio n c o e ffic ie nt o b ta ine d b y the inde xe s e nte ring into the c o mpo sitio n o f the E dhe c Alte rna tive I nde xe s. We c a n se e fro m the fo llo wing ta b le tha t the E dhe c Alte rna tive I nde xe s a re syste ma tic a lly mo re re pre se nta tive tha n the a ve ra g e o f the c o mpe ting inde xe s. T his c o nfirms tha t the PCA me tho d a llo ws the re pre se nta tivity dime nsio n to b e impro ve d sig nific a ntly.

T able 7: T he Re pr e se ntativity Dime nsion

T he PCA a ppro a c h e nsure s tha t the E dhe c Alte rna tive I nde xe s c a pture the la rg e st po ssib le fra c tio n o f the info rma tio n c o nta ine d in the da ta fro m the va rio us c o mpe ting inde xe s. T he first c o nse q ue nc e , a s c a n b e se e n in ta b le 7, is a ma ximiza tio n o f the re pre se nta tivity dime nsio n. T he se c o nd, sig nific a nt, c o nse q ue nc e is tha t the E dhe c Alte rna tive I nde xe s a re , b y c o nstruc tio n, syste ma tic a lly le ss b ia se d tha n the inde xe s the y c o nta in. Sinc e the c o mpe ting inde xe s a re a ffe c te d diffe re ntly b y the b ia se s me ntio ne d in se c tio n I , se a rc hing fo r the line a r c o mb ina tio n o f c o mpe ting inde xe s tha t implie s a ma ximiza tio n o f the va ria nc e e xpla ine d, le a ds to a minimiza tio n o f the b ia s. T his c ha ra c te ristic is o f g re a t inte re st fro m a pe rfo rma nc e e va lua tio n pe rspe c tive .

T HE PROPE RT IE S OF T HE E DHE C AL T E RNAT IVE INDE XE S III

17

Investment Styles Edhec Indexes Competing Indexes Convertible Arbitrage 0,84 0,77 Distressed Securities 0,94 0,88 Emerging Markets 0,98 0,95 Equity Market Neutral 0,41 0,35 Event Driven 0,96 0,93 Fixed Income Arbitrage 0,81 0,63 Funds of Funds 0,93 0,88 Global Macro 0,77 0,61 Long Short Equity 0,98 0,67 Merger Arbitrage 0,86 0,83 Relative Value 0,89 0,75 Short Selling 0,73 0,71 Average Correlation Coefficient 0,84 0,75

From January 1998 through December 2000

SLIDE 18

T he Stability of the E dhe c Alte r native Inde xe s

We the n a sse ss the impa c t o f sho rt-te rm ma rke t o r e c o no mic tre nds o n the c o mpo sitio n o f E dhe c Alte rna tive I nde xe s. T he first te st c o nsists o f c a lc ula ting the a ve ra g e we ig hting e vo lutio n c a use d b y the q ua rte rly inde x re b a la nc ing . F

e xa mple , if the E dhe c Style I nde x is c o mpo se d o f 40% o f I nde x 1 a nd 60% o f I nde x 2 in Q1 a nd 50% o f I nde x 1 a nd 50% o f I nde x 2 in Q2, we c o nside r a we ig hting e vo lutio n o f 1/ 2 * Ab s (40% - 50%) + 1/ 2 * Ab s (60% - 50%), i.e . 10%. We the n c a lc ula te d the a ve ra g e we ig hting e vo lutio n o ve r the pe rio d Ja nua ry 2001 thro ug h De c e mb e r 2002. T his te st sho ws tha t the inde x c o mpo sitio n is ve ry sta b le o ve r time (se e ta b le b e lo w fo r de ta ile d re sults), whic h c o nfirms tha t the c o nstruc tio n me tho do lo g y e na b le s E dhe c Alte rna tive I nde xe s to c a pture the sub sta nc e o f the inve stme nt stra te g y a nd no t sta tistic a l a rtifa c ts.

T able 8: Stability of the Composition of the E dhe c Alte r native Inde xe s

T he hig h le ve l o f sta b ility o f E dhe c Alte rna tive I nde xe s is o f g re a t inte re st fo r a sse t ma na g e me nt firms who inve st (o r a re willing to inve st) in he dg e funds. T he pra c tic a l a dva nta g e s a re nume ro us. F irst (b ut no t le a st) is the fa c t tha t the E dhe c Alte rna tive I nde xe s a re e a sie r to re plic a te .

III

18

Edhec Style Indexes Average weighting change Convertible Arbitrage 0.24% Emerging Markets 0.07% Equity Market Neutral 1.18% Event Driven 0.17% Fixed Income Arbitrage 2.03% Global Macro 0.61% Long/Short 0.57% Merger Arbitrage 0.05% Relative Value 0.34% Short Selling 0.07% Distressed Securities 0.27% Funds of Funds 0.08%

From January 2001 through December 2002

SLIDE 19

T he hig he r de g re e o f re pre se nta tivity a nd sta b ility o f the E dhe c Alte rna tive I nde xe s g ive the m no t o nly supe rio rity in te rms o f sta tistic a l pro pe rtie s b ut a lso in te rms o f po rtfo lio pro pe rtie s.

T he T r ac kability of the E dhe c Alte r native Inde xe s

de mo nstra te the a dva nta g e s o f the E

dhe c Alte rna tive I nde xe s in te rms o f tra c ka b ility, we ha ve c o nstruc te d po rtfo lio s ma de up o f sing le funds tha t re plic a te the E dhe c Alte rna tive I nde xe s. We the n did the sa me thing with the HF R a nd CSF B I nde xe s. Po rtfo lio re plic a tio n invo lve s a tra c king e rro r minimiza tio n pro b le m (se e the fo llo wing fo rmula ): whe re R

PF is the re turn o f the re plic ating po rtfo lio and RB the re turn o f the b e nc hmark. W re fe rs to the c o mpo sitio n o f

the re plic ating po rtfo lio . T he sing le funds use d fo r this e xpe rime nt we re dra wn ra ndo mly (witho ut re pla c e me nt) fro m the HF Ne t Da ta b a se . T he re plic a ting po rtfo lio s we re c o mpo se d o n a ve ra g e o f twe nty-five funds (whic h is a g o o d c o mpro mise b e twe e n industry pra c tic e – 20 to 40 funds - a nd a c a de mic e vide nc e – 5 to 10 funds11) with a t le a st 4 ye a rs o f histo ric pe rfo rma nc e s. We e xc lude d the Sho rt Se lling stra te g y sinc e we did no t dispo se o f 20 funds with a t le a st 4 ye a rs o f e xiste nc e . As a re sult, we c a lc ula te d the “in sa mple ” tra c king e rro r fo r Ja nua ry 1999 thro ug h Se pte mb e r 2002 fo r the 7 stra te g ie s c o mmo n to E dhe c , HF R a nd CSF B (the c ho ic e o f the HF R a nd CSF B inde xe s is mo tiva te d b y the fa c t tha t the y a re o fte n q uo te d b y pra c titio ne rs). T he re sults c o nfirm the supe rio rity o f E dhe c Alte rna tive I nde xe s, sinc e we o b ta in a lo we r “in sa mple ” tra c king e rro r in 6 o ut o f 7 stra te g ie s. T he tra c king e rro r is syste ma tic a lly lo we r tha n 2.5% a nd in 4 o ut o f 7 c a se s it is lo we r tha n 1.0%. T his is a ll the mo re inte re sting in tha t the tra c king e rro r o b ta ine d with E dhe c Alte rna tive I nde xe s a ppe a rs to b e so und (lo we r tha n 2.5% in 5 o ut o f 7 c a se s). T he lo w diffe re nc e o b se rve d b e twe e n the “in-sa mple ” a nd “o ut-o f-sa mple ” tra c king e rro rs indic a te s tha t the tra c king e rro r is sta b le o ve r time . I n so me c a se s the “o ut-o f-sa mple ” tra c king e rro r is e ve n lo we r tha n the “in-sa mple ” tra c king e rro r o b ta ine d with the c o mpe ting inde xe s.

T able 9: T r ac kability of E dhe c Alte r native Inde xe s

USE OF T HE E DHE C INDE XE S F OR PORT F OL IO MANAGE ME NT IV

19

11 L

e a rne d, M. a nd L ha b ita nt, F .S., 2002, He dg e F und Dive rsific a tio n: Ho w muc h is e no ug h? , Jo urnal o f Alte rnative I nve stme nts, Winte r 2002, Vo l.5, I ssue 3, p.23-49

( )

B PF w

R R TE − =σ min

CSFB HFR

In sample Out of sample In sample In sample

Convertible Arbitrage 0.73% 1.05% 2.23% 0.74% Emerging Markets 2.34% 3.39% 4.61% 3.19% Event Driven 0.95% 1.36% 2.40% 1.03% Fixed Income Arbitrage 1.11% 1.25% 0.83% 2.70% Global Macro 0.12% 2.23% 0.17% 0.13% Long/Short Equity 1.90% 3.25% 4.02% 2.07% Market Neutral 0.73% 0.86% 1.03% 2.28%

From January 1999 through September 2002

Edhec

SLIDE 20

E dhe c Alte r native Inde xe s in the Str ate gic Alloc ation Pr

c e ss

Sinc e E dhe c Alte rna tive I nde xe s a re e a sy to re plic a te , we ha ve te ste d the b e ne fits o f inc luding the m in the stra te g ic a llo c a tio n o f a typic a l institutio na l inve sto r. F

r tha t purpo se we ha ve te ste d the sta b ility o f the minimum va ria nc e

po rtfo lio . We c o nside r tha t this po rtfo lio is ma de up o f sto c ks (pro xie d b y the S&P 500), b o nds (pro xie d b y the L e hma n Glo b a l Bo nd I nde x) a nd a lte rna tive inve stme nts (pro xie d b y the E dhe c Alte rna tive I nde xe s, o r the c o mpe ting inde xe s). We impo se the fo llo wing c o nstra ints: a n a sse t c la ss c a nno t re pre se nt mo re tha n 60% o f the po rtfo lio ’ s ho lding s a nd he dg e funds must re pre se nt b e twe e n 10% a nd 30% o f the po rtfo lio ’ s ho lding s. I t ha s b e e n sho wn in se ve ra l re se a rc h pa pe rs

12 tha t a sig nific a nt pro po rtio n o f a po rtfo lio must b e de dic a te d to the a lte rna tive c la ss fo r it

to impa c t the po rtfo lio ’ s risk/ re turn c ha ra c te ristic s. T he la tte r c o nstra int is se t in this re spe c t. T he minimum va ria nc e po rtfo lio c o rre spo nds to the po rtfo lio tha t o ffe rs the lo we st va ria nc e o n the e ffic ie nt fro ntie r

13.

T he c o nstruc tio n o f this po rtfo lio invo lve s a va ria nc e minimiza tio n pro b le m: whe re RPF is the re turn o f the po rtfo lio , Ri the re turn o f the i

th asse t and xi the we ig ht o f the i th asse t.

T he te st a ims to a sse ss the sta b ility o f the stra te g ic a llo c a tio n (i.e ., minimum va ria nc e po rtfo lio ). We use d a thre e -ye a r c a lib ra tio n pe rio d (fro m Ja nua ry 1998 to De c e mb e r 2000) to c o mpute the first minimum va ria nc e po rtfo lio . We the n ro lle d o ve r a 6-mo nth c a lib ra tio n pe rio d to c o nstruc t the se c o nd minimum va ria nc e po rtfo lio , a nd so o n. T he la st ste p wa s to c a lc ula te the a ve ra g e po rtfo lio turno ve r implie d b y the se mi-a nnua l re b a la nc ing . We pe rfo rme d this te st with the E dhe c Alte rna tive I nde xe s, a s we ll a s with the HF R a nd CSF B I nde xe s, fo r a ll the stra te g ie s c o mmo n to the thre e pro vide rs. Onc e a g a in, the E dhe c Alte rna tive I nde xe s turn o ut to b e mo re e ffic ie nt tha n the c o mpe ting inde xe s, sinc e the y ma ke the stra te g ic a sse t a llo c a tio n mo re ro b ust, re g a rdle ss o f the inve stme nt style (se e ta b le b e lo w fo r mo re de ta ils). T he turno ve r is syste ma tic a lly lo we r with the E dhe c Alte rna tive I nde xe s tha n with the CSF B o r HF R I nde xe s.

T able 10: Ave r age Minimum Var ianc e Por tfolio T ur nove r

IV

20

12 Amin, G., K

a t H., 2002, Sto c ks, Bo nds a nd He dg e F unds, Jo urnal o f Po rtfo lio Manag e me nt, Summe r 2003, Vo l.29, I ssue 4, p.113-119.

13 Ame nc , N., Ma rte llini, L

., 2002, Po rtfo lio Optimiza tio n a nd He dg e F und Style Allo c a tio n De c isio ns, Jo urnal o f Alte rnative I nve stme nts, F a ll 2002, Vo l.5, N°2, p.7-20.

( ) ∑ ∑

= =

=

n i n j j i j i PF x

R R x x R

1 1

) , ( cov var min

Edhec CSFB HFR Convertible Arbitrage 0% 0% 0% Emerging Markets 6.10% 7.17% 7.34% Equity Market Neutral 0% 0% 0% Event Driven 0% 0% 0% Fixed Income Arbitrage 0% 0% 0% Global Macro 0% 3.70% 0% Long/Short Equity 0% 3.35% 0% Short Selling 0.80% 1.83% 1.36%

From January 2001 through December 2002

SLIDE 21

APPE NDICE S

21

SLIDE 22

Conve r tible Ar bitr age

I nve stme nt in c o nve rtib le b o nds. T he stra te g y is to b uy the c o nve rtib le b o nd a nd se ll sho rt the c o mmo n sto c k o f the sa me c o mpa ny.

CT A Global

CT A Glo b a l funds inve st in liste d fina nc ia l a nd c o mmo dity ma rke ts a s we ll a s in c urre nc y ma rke ts a ll o ve r the wo rld. T he y c a n fo llo w syste ma tic o r disc re tio na ry stra te g ie s a nd a re re fe rre d a s to Co mmo dity T ra ding Adviso rs.

Distr e sse d Se c ur itie s

I nvo lve s b uying b a c k, a t a lo w pric e , the se c uritie s o f c o mpa nie s tha t a re e xpe rie nc ing fina nc ia l diffic ultie s. T he se c uritie s ta rg e te d ma y c o ve r a wide ra ng e , fro m se nio r se c ure d de b t (lo we st risk) to c o mmo n sto c k (hig he st risk).

E me r ging Mar ke ts

I nve stme nt in e q uitie s o r b o nds fro m e me rg ing ma rke ts.

E ve nt Dr ive n

I nve stme nt stra te g y tha t e xplo its pric e mo ve me nts re la te d to the a ntic ipa tio n o f e ve nts a ffe c ting the life o f the c o mpa ny (me rg e r, a c q uisitio n, b a nkruptc y, e tc .).

F ixe d Inc ome Ar bitr age

T he inve stme nt re turn is b a se d o n e xplo iting pric e a no ma lie s re la te d to inte re st ra te instrume nts.

F unds of F unds

Co nsists o f inve sting in se ve ra l funds tha t ma y o r ma y no t fo llo w the sa me stra te g y.

Global Mac r

nve stme nt stra te g y with a stro ng le ve ra g e e ffe c t o n ma rke t e ve nts o r de ve lo pme nts.

ng Shor

t E quity

I nvo lve s inve sting ma inly in e q uitie s a nd de riva tive instrume nts. T he ma na g e r syste ma tic a lly use s sho rt se lling , b ut ta ke s c a re to ma inta in a pe rma ne nt o ve ra ll ne t po sitio n tha t is e ithe r lo ng o r ne utra l.

E quity Mar ke t Ne utr al

E xplo its ine ffic ie nc ie s in the ma rke t thro ug h b a la nc e d b uying o f unde rva lue d se c uritie s a nd se lling o f o ve rva lue d se c uritie s e na b ling e ithe r a b e ta o r a do lla r ne utra l a ppro a c h to b e o b ta ine d.

Me r ge r Ar bitr age

Me rg e r Arb itra g e funds inve st in c o mpa nie s invo lve d in a Me rg e r o r Ac q uisitio n pro c e ss. T he y typic a lly g o lo ng the ta rg e te d c o mpa ny a nd se ll sho rt the sto c k o f the a c q uiring c o mpa ny.

Re lative Value

T he o b je c tive o f this type o f stra te g y is to ta ke a dva nta g e o f the re la tive pric e diffe re ntia ls b e twe e n re la te d instrume nts.

Shor t Se lling

Ma inta ins a ne t o r simple sho rt e xpo sure re la tive to the ma rke t.

APPE NDIX 1 DE F INIT ION OF HE DGE F UND ST YL E S

22

SLIDE 23

T he PCA o f a time -se rie s invo lve s studying the c o rre la tio n ma trix o f suc c e ssive sho c ks. I ts purpo se is to e xpla in the b e ha vio r o f o b se rve d va ria b le s using a sma lle r se t o f uno b se rve d implie d va ria b le s. F ro m a ma the ma tic a l sta ndpo int, it invo lve s tra nsfo rming a se t o f K c o rre la te d va ria b le s into a se t o f o rtho g o na l va ria b le s, o r implic it fa c to rs, whic h re pro duc e s the o rig ina l info rma tio n pre se nt in the c o rre la tio n struc ture . E a c h implic it fa c to r is de fine d a s a line a r c o mb ina tio n o f o rig ina l va ria b le s. De fine R a s the fo llo wing ma trix: We ha ve n va ria b le s, i.e ., mo nthly re turns fo r n diffe re nt c o mpe ting indic e s, a nd T

b se rva tio ns o f the se va ria b le s.

(1) whe re is the ma trix o f the n e ig e nve c to rs o f R’ R. is U tra nspo se d. is the ma trix o f the n e ig e nve c to rs RR’ No te tha t the se n e ig e nve c to rs a re o rtho no rma l. λ{i} is the e ig e nva lue (o rde re d b y de g re e o f ma g nitude ) c o rre spo nding to the e ig e nve c to r U{i}. De no ting the princ ipa l c o mpo ne nt se nsitivity o f the kth va ria b le to the ith fa c to r, a nd V{ti}=F {ti}, o ne c a n e q uiva le ntly write e q ua tio n (1) whe re the n fa c to rs F {i} a re a se t o f o rtho g o na l va ria b le s. One ma y use the me tho d to de sc rib e e a c h va ria b le a s a line a r func tio n o f a re duc e d numb e r o f fa c to rs. T

tha t e nd, o ne ne e ds to se le c t a numb e r o f fa c to rs I

suc h tha t the first I fa c to rs c a pture a la rg e fra c tio n o f a sse t re turn va ria nc e , while the re ma ining pa rt c a n b e re g a rde d a s sta tistic a l no ise (2) whe re so me struc ture is impo se d b y a ssuming tha t the re sidua ls εtk a re unc o rre la te d o ne to a no the r. T he pe rc e nta g e o f va ria nc e e xpla ine d b y the first I fa c to rs is g ive n b y . By ta king I =1in e q ua tio n (2) this me tho d c a n b e use d to g e ne ra te "the b e st o ne -dime nsio na l" summa ry o f a se t o f c o mpe ting indic e s. F urthe rmo re , a simple no rma liza tio n a llo ws o ne to o b ta in a n inde x whic h c a n b e re g a rde d a s a po rtfo lio o f c o mpe ting indic e s, so tha t a n a c tua l de c o mpo sitio n in te rms o f a c tua l funds in the inde x c a n e a sily b e

b ta ine d a s lo ng a s info rma tio n is a va ila b le in e a c h c o mpe ting inde x c o mpo sitio n.

APPE NDIX 2 PRINCIP AL COMPONE NT ANAL YSIS

23

n k T t tk

R R

≤ ≤ ≤ ≤

=

1 1

) (

∑

=

n i ti ik i tk

V U R

λ

n k i ik

U U

≤ ≤

=

, 1

) ( ) (

n i k ki T

U U

≤ ≤

=

, 1

) ( ) (

n i T t ti

V V

≤ ≤ ≤ ≤

=

1 1

) ( ) (

∑

=

n i ti ik tk

F s R

1 tk I i ti ik tk I i ti ik i tk

F s V U R ε ε λ + = + =

∑ ∑

= = 1 1

∑ ∑

= = N i i I i i 1 1

λ λ

∑ ∑

= =

K i ti K k ik ik tk

F s s R

1 1 ' '

ik i ik

U S λ =

SLIDE 24

I t sho uld b e no te d tha t the E dhe c Alte rna tive I nde xe s a re me a nt to b e re pre se nta tive a nd sta b le a nd do no t a im to po st supe rio r risk-a djuste d pe rfo rma nc e . Adding va lue re ma ins the ta sk o f the fund ma na g e r who is in c ha rg e o f se le c ting funds a nd/ o r c ha ng ing the po rtfo lio ’ s ta c tic a l a llo c a tio n. Ne ve rthe le ss, it is inte re sting to g ive a sna psho t o f the E dhe c Alte rna tive I nde xe s’ risk a nd pe rf

a n c e c ha ra c te ristic s. F

r c o mpa riso n purpo se s we will do the sa me fo r the c o mpe ting inde xe s tha t e nte r into the ir

c o mpo sitio n

14. We will a lso c o mpa re the E

dhe c Alte rna tive I nde xe s’ risk a nd re turn c ha ra c te ristic s to the a ve ra g e risk a nd re turn c ha ra c te ristic s o f the c o mpe ting inde xe s (re fe rre d to a s Co mpe ting I nde xe s in ta b le 14). T he risk dime nsio n will b e c ha ra c te rize d b y the a nnua l vo la tility o f the inde xe s’ re turns (i.e . a nnua l sta nda rd de via tio n) a nd b y its Va lue - a t-Risk15. T he Sha rpe Ra tio 16 a nd the So rtino Ra tio 17 we re use d to a sse ss the inde xe s’ risk-a djuste d pe rfo rma nc e s. Ma ximizing the re pre se nta tivity o f the E dhe c Alte rna tive I nde xe s (i.e . the inde x o f inde xe s) c o nsists ro ug hly o f re duc ing the no ise o f the inde x po rtfo lio ’ s va ria nc e c o -va ria nc e ma trix. T his le a ds implic itly to a re duc tio n in vo la tility. Co nse q ue ntly, the E dhe c Alte rna tive I nde xe s a re no t o nly mo re re pre se nta tive a nd sta b le b ut the y a lso te nd to ha ve lo we r vo la tility. T he sa me re ma rk ho lds fo r the VAR, whic h indic a te s tha t the E dhe c Alte rna tive I nde xe s a re a lso le ss e xpo se d to e xtre me risks. T his la st pro pe rty is o f g re a t inte re st fo r institutio na l inve sto rs who a re ke e n to limit e xtre me lo sse s.

APPE NDIX 3 PE RF ORMANCE OF T HE E DHE C AL T E RNAT IVE INDE XE S

24

14 Only tho se inde xe s tha t fulfille d o ur re q uire me nts (se e se c tio n I

I fo r mo re de ta ils) we re inc lude d in the c o mpo sitio n o f the E dhe c Alte rna tive I nde xe s. As a re sult, o the r inde xe s tha t a re no t me ntio ne d in ta b le s 12 & 13 ma y b e a va ila b le o n the ma rke t.

15 We c a lc ula te d histo ric a l Va lue a t Risk with a 95% c o nfide nc e le ve l (i.e . α=5%):

whe re µ is the a ve ra g e re turn, σ the sta nda rd de via tio n a nd z the c ritic a l va lue fo r pro b a b ility (1-α). W re fe rs to the c urre nt va lue o f the po rtfo lio .

16 T

he Sha rpe Ra tio s we re c a lc ula te d with a risk-fre e ra te o f 3%. Sha rpe Ra tio = (µ - Rf) / σ, whe re µ is the a ve ra g e re turn, σ the sta nda rd de via tio n a nd Rf the risk-fre e ra te .

17 T

he So rtino Ra tio s we re c a lc ula te d with a Minimum Ac c e pte d Re turn e q ua l to the risk-fre e ra te (i.e . MAR=3%). So rtino Ra tio = (µ - Rf) / σMAR, whe re µ is the a ve ra g e re turn, σMAR the se mi sta nda rd de via tio n a nd Rf the risk-fre e ra te .

) * ( σ µ z W VAR − =

SLIDE 25

T able 12: Risk Adjuste d Pe r for manc e T able 11: Risk Ove r vie w

25

Edhec HFR CSFB EACM Altvest Volatility VAR* Volatility VAR* Volatility VAR* Volatility VAR* Volatility VAR* Convertible Arbitrage 3,46% 18,33% 6,30% 20,50% 4,03% 20,24% 6,30% 20,50% CTA Global 9,07% 22,55% 12,44% 27,44% Distressed Securities 4,91% 21,50% 5,40% 21,49% 5,30% 21,28% 4,50% 18,92% 6,00% 23,67% Emerging Markets 11,52% 35,16% 14,34% 40,62% 13,44% 36,88% 7,23% 25,78% Equity Market Neutral 1,93% 12,26% 3,44% 11,77% 2,10% 13,73% Event Driven 4,98% 19,45% 6,62% 22,73% 4,31% 18,48% 3,84% 17,33% 4,33% 18,30% Fixed Income Arbitrage 2,08% 11,51% 2,77% 11,28% 2,47% 11,80% Funds of Funds 5,91% 19,54% 5,71% 17,98% 10,08% 29,62% Global Macro 5,70% 19,26% 6,82% 21,85% 7,88% 26,05% 8,29% 24,16% Long Short Equity 7,72% 22,69% 10,64% 29,97% 11,64% 30,04% 4,29% 18,36% Merger Arbitrage 3,32% 14,15% 3,21% 13,14% 4,26% 14,45% 3,96% 13,38% 3,48% 15,61% Relative Value 3,33% 15,73% 2,26% 13,62% 2,63% 12,44% 2,45% 15,07% Short Selling 22,46% 39,62% 26,43% 49,07% 17,55% 24,87% 18,47% 36,90% 8,32% 29,36% Hennessee Van Hedge CISDM HF Net Barclay S&P Volatility VAR* Volatility VAR* Volatility VAR* Volatility VAR* Volatility VAR* Volatility VAR* Convertible Arbitrage 3,99% 17,59% 3,21% 18,26% 3,23% 17,83% CTA Global 8,24% 21,37% 9,50% 26,36% 8,25% 19,26% 16,77% 39,50% Distressed Securities 5,66% 21,37% 5,45% 21,33% 5,42% 24,30% 5,43% 22,85% Emerging Markets 11,71% 33,02% 18,25% 49,35% 10,46% 32,72% 11,87% 40,36% 14,66% 43,46% Equity Market Neutral 2,63% 7,49% 3,42% 17,12% 1,27% 11,37% 3,04% 15,92% 3,32% 13,01% Event Driven 5,37% 21,18% 3,22% 14,97% 6,31% 22,17% 6,01% 22,22% 4,03% 16,85% Fixed Income Arbitrage 3,21% 12,18% 3,33% 14,08% 2,24% 13,86% 2,37% 13,49% Funds of Funds 7,82% 24,16% 3,51% 13,43% 4,81% 17,86% 4,96% 18,20% Global Macro 6,82% 16,58% 11,45% 26,49% 4,03% 14,12% 5,40% 22,30% 5,91% 21,75% Long Short Equity 9,43% 31,59% 9,28% 29,95% Merger Arbitrage 3,65% 14,69% 3,28% 15,74% 3,23% 14,33% Relative Value 9,21% 26,56% 11,71% 34,35% 2,13% 14,55% 2,71% 13,92% Short Selling 16,50% 32,09% 24,23% 43,74% 18,87% 33,64% 21,13% 43,34% * computed at a 95% confidence level Analysis period: from January 1999 through December 2003 Edhec HFR CSFB EACM Altvest

Sharpe Ratio* Sortino Ratio* Sharpe Ratio Sortino Ratio Sharpe Ratio Sortino Ratio Sharpe Ratio Sortino Ratio Sharpe Ratio Sortino Ratio

Convertible Arbitrage 2,79 7,01 1,14 1,59 2,64 5,57 1,14 1,59 CTA Global 0,52 0,85 0,32 0,52 Distressed Securities 2,13 5,57 1,78 4,26 1,81 3,19 1,90 4,54 1,81 4,04 Emerging Markets 1,15 2,23 0,98 1,83 0,88 1,66 1,51 3,09 Equity Market Neutral 3,16 22,16 0,91 1,87 3,46 20,55 Event Driven 1,67 3,12 1,34 2,35 1,96 3,59 2,10 4,08 1,89 3,92 Fixed Income Arbitrage 2,46 5,70 1,35 2,55 1,92 3,16 Funds of Funds 1,16 2,79 0,98 2,15 1,00 2,45 Global Macro 1,21 2,93 1,12 2,33 1,29 2,35 0,91 1,90 Long Short Equity 0,91 1,74 0,89 1,83 0,68 1,39 1,94 4,02 Merger Arbitrage 1,72 2,91 1,52 2,49 1,05 1,75 0,98 1,42 1,98 3,80 Relative Value 2,18 4,22 3,06 13,32 1,95 3,57 3,29 12,07 Short Selling

0,01
0,01

0,10 0,15

0,39
0,53

0,20 0,32 1,53 4,08 Hennessee Van Hedge CISDM HF Net Barclay S&P

Sharpe Ratio* Sortino Ratio* Sharpe Ratio Sortino Ratio Sharpe Ratio Sortino Ratio Sharpe Ratio Sortino Ratio Sharpe Ratio Sortino Ratio Sharpe Ratio

Convertible Arbitrage 2,02 4,04 3,12 9,06 2,95 8,58 CTA Global 0,59 0,99 0,82 1,53 0,33 0,52 0,54 0,90 Distressed Securities 1,61 3,22 1,72 4,64 2,29 6,28 2,02 4,64 Emerging Markets 0,92 1,87 0,90 1,75 1,20 2,40 1,51 3,17 1,12 2,21 Equity Market Neutral 0,07 0,09 2,49 13,49 4,93 51,43 2,62 15,40 1,38 3,15 Event Driven 1,74 3,54 2,08 4,70 1,40 2,64 1,56 3,09 1,79 2,95 Fixed Income Arbitrage 1,22 2,41 1,69 4,01 3,20 12,37 2,79 8,67 Funds of Funds 1,07 2,33 1,33 2,76 1,45 4,07 1,42 4,06 Global Macro 0,35 0,63 0,41 0,69 1,12 2,66 1,94 5,37 1,53 4,34 Long Short Equity 1,39 3,85 1,27 3,46 Merger Arbitrage 1,57 2,63 2,24 4,84 1,86 3,28 Relative Value 0,92 1,60 1,04 1,95 3,78 16,01 2,39 8,77 Short Selling 0,12 0,19 0,04 0,06

0,02
0,02

0,27 0,39 * computed at a 95% confidence level Analysis period: from January 1999 through December 2003

SLIDE 26

26

18 T

he Ske wne ss a nd the E xc . K urto sis re fe r to the third a nd fo urth o rde r mo me nts o f a distrib utio n func tio n.

T able 13: E dhe c Alte r native Inde xe s ve r sus the ave r age of c ompe ting inde xe s

Convertible Arbitrage CTA Global Distressed Securities Emerging Markets Equity Market Neutral

Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes

Average Annual Return 12,66% 11,99% 7,67% 8,68% 13,45% 13,05% 16,26% 16,87% 9,10% 8,55% Min Monthly Return

1,59%
2,72%
5,43%
6,63%
2,09%
2,72%
5,41%
6,17%
0,13%
0,75%

Max Monthly Return 3,44% 3,63% 6,82% 8,45% 4,21% 4,95% 12,30% 13,48% 2,53% 3,07% % of winning months 86,67% 86,11% 56,67% 57,33% 76,67% 76,04% 73,33% 69,79% 95,00% 84,72% % of losing months 13,33% 13,89% 43,33% 42,67% 23,33% 23,96% 26,67% 30,21% 5,00% 15,28% Skewness

0,34
0,59

0,10 0,15 0,07

0,02

0,32 0,44 0,99 0,69

Exc. Kurtosis

0,35 1,98

0,21
0,10
0,42

0,51 0,87 1,19 0,99 1,57 Annual Std Deviation 3,46% 3,94% 9,07% 11,04% 4,91% 5,39% 11,52% 12,74% 1,93% 2,65% Annual Semi Std Deviation 1,38% 1,95% 5,47% 6,59% 1,87% 2,38% 5,95% 6,48% 0,28% 0,87% VAR 95% 18,33% 18,46% 22,55% 26,79% 21,50% 21,90% 35,16% 37,77% 12,26% 12,90% Sharpe Ratio 2,79 2,47 0,52 0,52 2,13 1,87 1,15 1,13 3,16 2,41 Sortino Ratio 7,01 6,12 0,85 0,89 5,57 4,35 2,23 2,25 22,16 17,14

Event Driven Fixed Income Arbitrage Funds of Funds Global Macr o Long/Short Equity

Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes

Average Annual Return 11,29% 11,33% 8,11% 8,30% 9,84% 10,13% 9,91% 10,06% 10,04% 13,13% Min Monthly Return

3,00%
3,06%
0,92%
1,36%
2,69%
3,32%
3,04%
4,17%
3,89%
4,15%

Max Monthly Return 4,29% 4,52% 2,08% 2,93% 6,66% 7,11% 6,12% 7,20% 7,45% 10,05% % of winning months 80,00% 81,48% 86,67% 85,28% 71,67% 73,33% 68,33% 68,54% 66,67% 69,33% % of losing months 20,00% 18,52% 13,33% 14,72% 28,33% 26,67% 31,67% 31,46% 33,33% 30,67% Skewness

0,49
0,42
0,26

0,06 1,26 1,19 0,87 0,70 0,48 0,90

Exc. Kurtosis

0,94 1,66 0,70 1,97 2,99 3,48 1,35 1,88 0,93 2,66 Annual Std Deviation 4,98% 4,89% 2,08% 2,73% 5,91% 6,15% 5,70% 7,07% 7,72% 9,05% Annual Semi Std Deviation 2,66% 2,56% 0,90% 1,22% 2,45% 2,58% 2,36% 3,50% 4,04% 3,96% VAR 95% 19,45% 19,36% 11,51% 12,78% 19,54% 20,21% 19,26% 21,66% 22,69% 27,98% Sharpe Ratio 1,67 1,76 2,46 2,03 1,16 1,21 1,21 1,08 0,91 1,24 Sortino Ratio 3,12 3,43 5,70 5,53 2,79 2,97 2,93 2,53 1,74 2,91

Merger Arbitrage Relative Value Short Selling

Edhec Index Competing Indexes Edhec Index Competing Indexes Edhec Index Competing Indexes

Average Annual Return 8,71% 8,60% 10,26% 10,89% 2,79% 4,76% Min Monthly Return

2,67%
2,78%
2,21%
2,10%
13,40%
16,58%

Max Monthly Return 2,39% 2,90% 3,33% 4,43% 16,57% 17,30% % of winning months 85,00% 82,86% 85,00% 82,38% 51,67% 51,25% % of losing months 15,00% 17,14% 15,00% 17,62% 48,33% 48,75% Skewness

1,09
0,88
0,77

0,17

0,01

0,01

Exc. Kurtosis

2,46 2,26 1,91 0,65

0,24

0,66 Annual Std Deviation 3,32% 3,58% 3,33% 4,73% 22,46% 22,00% Annual Semi Std Deviation 1,97% 2,08% 1,72% 2,20% 15,82% 15,11% VAR 95% 14,15% 14,48% 15,73% 18,64% 39,62% 40,84% Sharpe Ratio 1,72 1,60 2,18 2,35

0,01

0,08 Sortino Ratio 2,91 2,89 4,22 8,18

0,01

0,13 Analysis period: from January1999 through December 2003

SLIDE 27

I

4

SLIDE 28

I

1090, ro ute de s Crê te s So phia -Antipo lis 06560 Va lb o nne T é l. : +33 (0)4 92 96 89 50 F a x : +33 (0)4 92 96 93 22 r e se a rc h@ e dhe c -risk.c o m