Testing and Analyzing the Momentum Anomaly in the Egyptian
Stock market: MV versus SD Approach
Guidelines for the Methodology and Data Analysis Data Analysis of the work shall consist of two main parts as follows First : Analysis of the Momentum Anomaly (Returns) in the Egyptian Stock Market using the MV Approach. ! Testing the e"istence of the Momentum anomaly for the #hole period from $$%&&' till ($)$%&(! The Jagadeesh and Titman's (1993) approach is taken to identify the winner and the loser portfolios according to their past monthly returns !e in"estigate the profits of momentum across "arious lengths of formation and holding hori#ons Different hori#ons for the formation period ($%1& 3& '& 9& 1(&) and the holding period ()%1& 3& '& 9& 1() are chosen To increase the power of the tests& the e*amined strategies include portfolios with o"erlapping holding periods And to reduce price pressure and lagged reaction& the trading strategies are constructed +y skipping one month +etween the formation periods& and +etween formation period and holding period (Jagadeesh and Titman 1993, -hou& !ei& and -hung& (../) To a"oid costly monthly re+alancing& a +uy0and0hold method is used 1nce holding period returns are computed& they are transformed into a"erage monthly returns to facilitate direct comparison +etween strategies with different holding hori#ons The choice of the formation ($) and the holding period ()) is purely su+2ecti"e 3aired T0 Test is used to determine if the momentum profits in the holding periods are significantly different from that in the formation periods 4f there is e"idence of momentum profit in the stock market& the winner0loser portfolios will generate significant a+normal profits Value weighte !ortfolios shall +e formed using the 5arket -apitali#ation for each stock so that to ensure that if momentum e*ists it is not due to the small si#es stocks found in our sample !e shall use the methodology regarding this step used in the following article $i"e 6uintile portfolios are formed The top 6uintile portfolio is called the 7!inners7 6uintile and the +ottom 6uintile is called the 78osers7 6uintile !here the num+er of stocks is not perfectly di"isi+le +y 9& :e*tra; stocks are 1 " 3 a g e assigned to the middle portfolio (portfolio 3) to ensure the winner portfolio (1) and the loser portfolio (9) ha"e the same num+er of stocks <e*t& an "alue 0weighted a"erage return for each 6uintile portfolio o"er the ne*t )0month holding period is computed 4n each month t& the strategy +uys the winner portfolio and sells the loser portfolio& holding this position for ) months 4f the a"erage return of the winner portfolio during the )0month holding period is higher (lower) than that of the loser portfolio& the outcome is momentum Do Momentum and *ontrarian +rofits E"ist in the Egyptian Stock Market, Elsayeda A-del./attah 0smail Ta+le <o 9 3age 99 =se the same steps mentioned on 3age 99 +ut use the market -apitali#ation in sorting the 9 6uintiles or 3ortfolios ( !inners and losers ones ) and in order to calculate the 5omentum monthly return (computed in the e*cel sheets) shall +e used The a"erage monthly log returns for the >gyptian ?tock >*change firms are calculated to the e6uation @ At % 8og (3t B 3t01) from the $C) strategies where $ stands for the num+er of months included in the formation period and ) denotes the length of the holding period 3ortfolios are re+alanced e"ery ) period Dased upon total past returns during $0months& securities are ranked into 9 "alue0weighted portfolios (in descending order) with the top one called winners and the +ottom one losers The ta+le also shows the a"erage returns for winners minus losers T0 statistics are presented in parentheses& with one& two or three asterisks indicating significance at the 1E& 9E and 1.E le"els respecti"ely 5omentum profits are reported in the +old cells <ote ( our analysis will co"er only 1( months as contrarian anomaly is not under study) ( " 3 a g e #.$ Testing the 1isk 2ased source and E"planation for the Momentum Anomaly for the #hole period from $$%&&' till ($)$%&(! ( %A!M analysis). To see how well -A35 model e*plains the momentum returns& simple regression tests are performed The e*cess returns of the momentum strategies are regressed on market risk premium "aria+le in all formation and holding hori#ons& 4n addition to the regress of the e*cess returns of all winner and loser portfolios on market risk premium for all hori#ons Aegression tests are measured +y the Alpha coefficients and portfolios' sensiti"ities to the market risk represented +y the Deta coefficients 4f a pricing model is a good description of portfolio returns& one e*pects to see Fero or low Alphas and significant positi"e Detas !e shall use the methodology regarding this step used in the following article& +ut also instead of e6ually weighted 3ortfolios we shall use "alue weighted ones as pre"ious Do Momentum and *ontrarian +rofits E"ist in the Egyptian Stock Market, Elsayeda A-del./attah 0smail Ta+le / G 3age '( !( Testing the Effect of the /inancial *risis on the Momentum +rofits ?u+ period (A) shall +e constructed co"ering the period from 1B1B(..9 G 31B1(B(../( to test the momentum profit trading strategy in normal economic condtion) ?u+ period (D) shall +e constructed co"ering the period from 1B1B(..H G 31B1(B(.1. (after the financial crisis G to report if there is an effect of the >IJ due to the -risis on 5omentum returns across "arious $ormation and )olding periods =sing the same steps and methodology a+o"e to test the effect of the $inancial crisis on the 3rofita+ility of the 5omentum trading ?trategy +y e*amining the effect +efore and after the -risis ( using the Kalue weighted 3ortfolios too ) T0 Test is also re6uired for each ?u+ period !3 Testing Market State Effects on Momentum 1eturns 3 " 3 a g e !e e*amine whether conditioning on the state of the market is important to the profita+ility of momentum strategies !e define two states@ (1) L=3M is when the lagged three0year market return is non0negati"e& and (() LD1!<M is when the three0year lagged market return is negati"e ?u+ 3eriod (A) (D) and (-) shall report momentum returns following 3'0 (N0 1( months =p and down markets o"er the period from 1B1B(..9 G 31B9B(.13 ?u+ 3eriod (D) (>) ($) shall +e constructed co"ering the period from 1B1B(.11 G 31B9B(.13 to report momentum returns following 3'0 (N0 1( months =p and down markets o"er that period in particular All returns are reported in per cent and t0statistics also should +e pro"ided in each ta+le Note : You can refer to the following article as a general guideline (Momentum Returns, Market States and the 2007 Financial risis M!"#MM#$ #"M#$ "%%M# &'()%R* +, N#R*%#- in this ste. anal/sis , !' Analysing the 0ndustry Momentum The 4ndustry 5omentum is analy#ed with a +it different methodology used in the a+o"e points (11 01N) as the 4ndustry 5omentum portfolios are formed +ased on 80 months lagged returns and held for ) months !here 8 % 1 & '& 1(& month lagged returns and ) % 1& '& 1( & (N& 3' months holding period strategies $irst an e6ually weighted 3ortfolios method shall +e used ( for the whole panel data 0(..90(.13) !hile the !inners 3ortfolios is the e6ually Gweighted return of the highest 3 momentum industries& the losers 3ortfolios is the e6ually weighted return of the lowest three momentum industries !O8 presents the 4ndustry 5omentum retruns <o middle 3ortfolios is calculated 3anel (A) shall +e constructed to report the 4ndustry momentum profit& +ut a month shall +e skipped +etween the portfolio formation and holding periods to a"oid +id and +ounce effect ?econd a decomposition of the 4ndustry 5omentum profits into "arious components retlated to si#e and dollar trading "olume shall +e calculated for the period +etween 1B1B(..9 G 31B9B(.13 N " 3 a g e 3anel (D) reports profits for the largest and smallest (.E of stocks +ased on market capitali#ation (si#e) ?pecifically the 3 4ndustries that performed +est and the 3 three performed the worst at time are selected 3anel (-) reports the 5omentum returns according to the sort of portfolios +ased on Kalue Traded stocks 5ethodology used in the following article shall +e used for the analysis of 5omentum 4ndustry ( 2ust refer to the following pages only ) Do 0ndustries E"plain Momentum, To-ias 4! Mosko#itz5 Mark Grin-latt Ta-le 000 6 +age %7& Ta-le V 6 +age %77 Ta-le 0V 6 +age %7' &ypotheses of the Stuy (!art ') ",0, *he momentum1under reaction h/.othesis: %2treme mo3ements in .ortfolio return will 4e followed 4/ su4se5uent return mo3ements in the same direction, ",2 *he 4eta risk cannot e2.lain the momentum .rofits in %&67 there is a4normal return and no risk differential 4etween the winners and losers .ortfolios, ",8, Momentum strategies generate .ositi3e .rofits following u. states and negati3e following down states market, ",9, Momentum rashes has no effect on momentum returns in the %&6, ",:, 'ndustries e2.lain Momentum Returns and ;rofits in the %&6, 0mportant 8otes The closing prices of indi"idual stocks are downloaded from 5etastock data+ase The returns at time (t) were computed as the natural logarithm of the indi"idual stock prices (3)& that is@ At % log (3t B 3t01)P1.. This study uses log returns rather than raw returns& +ecause the results using raw return in estimating portfolio performance suffer from measurement errors ()on Q Tonks& (..() 9 " 3 a g e 4n addition& the monthly treasury +ill rate is downloaded from the -entral Dank 4n sum& positi"e and negati"e monthly returns are used to form different winners& losers& Fero momentum (winners0 losers) and Fero contrarian (losers0winners) portfolio All data files are prepared and uploaded for the data analysis purpose ?ample for the study is found in the e*cel sheets 5onthly returns ha"e +een calculated so no need to redo this step 5arket Aisk 3remium ( 5arket Aeturn G Aisk $ree Aates) are also calculated for the whole panel data and found in the e*cel sheets 8agged 5arket Aeturns are (3'& (N& 1() are also calculated and shall +e uploaded to you too All 5arket -apitali#ation data is found under the files called Dulletin(+etween 33N 09') Dulletin files include so many stocks data that shall not +e included in our study e*cept for the stocks only listed in the e*cel sheets ( ranging from 19. to (.. stocks during the whole period Secon : Analysis of the Momentum Anomaly (Returns) in the Egyptian Stock market using the S( Approach =nder preparation To +e deli"ered soon ' " 3 a g e