HAL Id: hal-02434392
https://hal-ifp.archives-ouvertes.fr/hal-02434392
Preprint submitted on 10 Jan 2020
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Analysis of the oil futures market
Karim Faid
To cite this version:
Centre Economie et Gestion
Analysis of the oil futures market
Karim FAID
Décembre 1992
Cahiers du CEG
~
n° 13
ENSPM - Centre Econom.ie et Gestion
228·232, a.venue Napoléon Bonaparte, Boîte postale 311, 92506 RUEIL MALMAISON
CEDEX.
La collection "Cahiers du CEG" est un recueil des travaux réalisés au Centre d'Economie
et Gestion de l'ENSPM, Institut Français du Pétrole. Elle a été mise en place pour permettre la
diffusion de ces travaux, parfois sous une forme encore provisoire, afin de susciter des échanges
de points de vue sur les sujets abordés.
Les opinions émis(!S dans les textes publiés dans cette collection doivent être considerées
comme propres à leurs auteurs et ne reJlèt.ent pas nécessairement le point de vue de l'IFP ou
de l'ENSPM.
Pour toute inforrna.tion complémentain', prière de contacter :
Saïd N ACHET ( Responsable de la publication ) tél. ( 1) 4 7 52 64 08
"Cahiers du CEG" is a collection of researchs realized within the Center for Economies
and Management. of the ENSPM, Instit11t Français du Pétrole. The goal of such collection is
to allow views exchange about the subjects treated of.
The opinions defended in the papers published are the author(s) sole responsability and
don't uecessarily rellect the views of the IFP or ENSPM.
For any additional information, please contact :
Abstract
The aim of this work was to gain some insight into the American crude oil
futures market using the means of multivariate data analysis. Using these
techniques, the first step was to find out
if
there is some identifiable behavior of
the crude oil price and to whkh fondamental market factors such behavior is
most related. Two models are developed, a descriptive mode! to explain the
time behavior of the crue oil futures price, and a forecasting mode! to predict
the prke changes whitout calling on the speculative factors but using solely the
fundamental physical factors of the market.
Résumé
Ce travail vise
à
apporter un éclairage sur le fonctioru1ement des marchés
à
terme pétroliers américains en utilisant l'analyse de données. La première étape
vise
à
explorer l'existence d'un comportement identifiable du prix du pétrole
brut et de rechercher les facteurs fondamentaux du marché auxquels
ce
comportement peut être relié. Deux modèles ont été développés; un modèle
descriptif pour expliquer le comportement dans le temps des prix
à
terme du
ANALYSIS ()l<'
nm
OIL nrrUlŒS MARKETKnrin1 FAID
I. INTRODUCTION
ln ordcr to formulate a geneml mode! of futures pdcc bchavior, oue must lïrst study the market dc!ennina:nt, of prke. Studies of other speculaûvc markeLs such a.1 ùic stock market have shown ùHlt llcmiu1d theory fails
Lo offer an atlcquatc explanation of ptîcc behavior. 1-Iowever, with coJnrnodity futures, a grc.at.cr variety of
quantity relatives are available. Pricc lluctuations can be rclated to both speculaûve and physiciù variables.
The former entmnpass volume of trading and open intcrcst, whilc the latter include supply, dcniand, and
stock levels. The tom! number of detcnninanLS of commodity prices can be dassificd into four groups as
follows:
Il Acting u,rough Demanù
- Consomption. - Exports.
• J)cdved demand for final prnducts.
- Govcrnn>cnt policies (long run).
· Demand relatives such as the prices or substitut.1ble coinmodities, or consumption/stock ratios.
21 Acting through Supply
• Production. - Stocks. - Weather.
• Govenunent pofü,ies (long run).
• Supply relatives such
as
tlie production or substitutable commodities, or production/stock ratios.31 Acling thmugh Economie Conditions
• Business conditions
as
rctlected in industrial production, unemployment,and the gencral pricc levcL
- Credit con<litions which defü>c the availability
or
loans for speculation or commodity storage.41 Acting through Market Composition
• Speculation. - Hcdging.
ln this stuùy, we wil! be concenied with a description of ùie intricate relationships bctween futures crudc on price and the physical variables of lhe energy market. lt is clcar that lhis will cert.tinly not give
Uie wholc pkturc,
and
we evcn e,pect to get a veryfuzzy
picture duc to Uie fact that the futures market isin essence a speculativc market in which speculative variables play a fun(latnental rolc. ln fact, two
.speculalive parametcrs, short hcdging and long speculating, reprcscnt tlie effective suppl y and demand for
futures contracts, respectivcly. However, because of the fact that a cormnodity futures market is far from
independent from the co1rnsponding physical market, the physical variables of tlie market do undoubtedly
play a significant mie in Ule fllmrcs pricc behavior. lt is this rnlationship we will concenume on and try to
deciphcr.
Whi,t we are interestcd in is tlie short
nm
bdmvior of Uie futures crude oil price. From the 011~1et, westar! with the following hypothcscs on the futures price ,·esponse to the physical suppl y and demand:
-a/ 111erc is a weak relationsllip bctwcen futures prices and supply or dcmand itl Uie shon run, but prices
may
be
better relatcd to the compm1cnts of supply and dcmand :Supply ~ Production+ Stocks Dcmand"' Consumption + Exports
~b/ For energy futures, shortwtenn price trends arc extremely rcsponsive to variations in inveotory
-cl The pressure of demand on supply is a more immediate influence (m pricc than arc the influences taken individually. Diffel'ent parnmetcrs eau be used to measure that pressure. 'lbesc are the various pressure rnlios that could be uscd : Dcmand/Supply. Dcma!Hl/Stocks, Expom/Stocks, Consumption/Supply, Consumplion/Stocks, Stocks/Consumption, Stocks/Prnduction.
-d/ Priccs of one comm,xlity vary with the priccs or suppl y and dcmand of other commodities. One can relate the priccs of a commodity tlm.mgh a pricc index to tl!e prices of domestic commodity substitutes, foreign commodity substitutcs, or to the prices of its convcrtcd products. For crude oil, wc would look at
its converted products
gince
il is not a CoJnlnodily that can be casily substîtutc<l, a1.least
not in the shorlnm we arc interested in. This is why we choostl to examine the priccs, supply, and dcmand for oil products. From Uic New York Mercanlile Exchange (NYMEX), we gatliercd data conceming the monthly priœs of West Texas Intcnncdiate Crude Oil (WTI) contracts, Unlcaded Regular Gasoline (HU) contrnc1s, and No2. Hcating Oil (HO) contracts. The publications of the Amcrican Petrolcum lnstitute (API) provi(!Cd us with monthly (lata conceming rundamemal market factors such as imports of crndc oil and oil products. crmlc oil stocks, pn.xlucüon and stœks or oil prodm:1s, etc. The data set runs from Fcbruary 1987 to Dccember 1991. The table below lists t11e variables used in this swdy.
We shall base our :malysis on thrcc approaches of multivariate data analysis. First, the Principal
Components Analysis wiU provide a gcnctal pîcture of the in1.cractions betwccn the futures prices and the
fund,u11cntal market factors considercd. Second, the Factorial Discriminant Analysis will allow us to
idcntify Uu)se variables that can best discOtninate betwecn increases and <lecreases in the crude oil price.
Finally, the Canonicat Corrclalion approach will providc the !inca,· combination of the variables tlmt is U1e most correlated with the crude oil fulUl'CS price a11d wlüch allows us to run tl1e optirrnJ multiple regression.
VARIABLE Deflnlllon
WI1MP West Tex.as lntennedh:1tc Moothlv Price HUMPR tJnlcadcd Rèl~ulo.r <1asoline Mt,nlhlv Pricc HOMPR No2. Hcatirl1.t ()jl Monthlv Pricc
CRIMl'O Ctude Oil Inir-v-.rL.'t CRS'ltJC Crude ()jl Stocks PRIM PO Produtit Irntx)l"ts
HtJPROl) Unleaded Gas~,iinc Ptoduc.tion HUSK>C llnleaded (Jasolinc. S~ock~ GAPROD Tot:,! (lasoline Produçt.ion GABTC:C 'rotal G~i:;oline Stücks HOPROD Hcating Oil Produclinn HOSTtX.' }Icntir1i::! Oil S1ocks
REPROD Rc:;Jdoal Fuel Oil Produç.lion RES TOC: Resîdual Fnel Oil Stoc.ks KEPROD Jet Fuel Kenrnine Production KllSTCX:: Jet Puel Kerosinc. Stoçks
ClNPlrr lnput 10 Crude Oil l)istillati~'Jn Onits
PJ;OPER Percc.nta.ec of Rcfinerv C:anacîtv Onerated
cs·nc:lN
Crude Stncks/C'.rude Inout CST/CIM (~rude Stock:;ICrude hnports CIM/PIM Crude. lrnnons/Product l1nrv,rtsm.1snnw
llnlcadcd Ciasoline Stocks/lh1leacled G.asolinc PtùductionGAS/GAP Total Gasollne Stocks(I'otal Gt1.sùline Product.ion l-!OS/HOP lleatinl! Oil St.ock:-./l-leatin~ Oil Production
RES/RIW Residunl Fuel Oil Stoçks/Residual Fuel Oil Pn)duction Kfü/KEP Keros.cnç Stocks{Ke!'osenc Production
.t!m.1..
1/ Stock.s of unlea.de.d ànd lei'!ded 1no(or gasüline repre!.cnt only finish,id Ill01or gasolin('. localcd at rcfineries, pipeline and
bulk terniinah. A large portion Qf motor 1:;;1s(,1ine $loi::ks at refintries consisls of cornponc.nu which rnay be blended or
con1pot.111ded to prOO.uce eith~r finfahed leadcd or unleaded n1olor gasoline. lltercJorc, these stocks cannoi hc classifièd ,,s ci1hcr
fin.ished kadcd or unleaded 1nütor ga.'>ollni::. 111e ~u,n of finl!ihed lcaded and unlc.adcd n101or ga._<;olinc. .stock..~ plus the componcnts
.!Uocks equals H)lal n1otor gas(,line. st(l(;:k.~.
2/ We will 1nak.c. ui.e 11.lso ùf oLher vadables d~rivr.d fronl the ba.~ic vatiables .i.tatcd 11.bove. Starting frorn a v,u-iabh~ Y ni
DY ~Y -Y ,nd Yl~Y m lii 111-l 1n-l
correSl)(.)nding to ù1~ 1iH)mhly çhangc and the one·moJUh lag!J.ed vari"bh!, fl?;<:pcctivcly.
We shall look at the <lllta !rom two diffcrcnt angles. From the Ors~ hercafter rcfcrred to as the descriptive framework, wc shall seek the cxplanatory power for tlie crnde oil futures price in the physical dcterrnin:mts of the mal'ket and the oil products fmurcs prices. We will thcrefore be using the val'iables described in tl1e
table above with no timc lag. Our goal will be to find out how much of the monlJ)ly cmde oil futures price
(WflMP) can be explained by the othel' variables uscd in this analysis.
Fro1n the second angle, hcrcan.cr rcferrcd to a.s the prcdictivc fr~unework, we shall seek U1e forecasting
power for the monthly crude oil futures pric.:e variation in the physical detenniuants of U1e market and the
oil products futures pl'iccs. W<J Will then be using u,e samc varinblcs but lagging the monUlly crnde oil futures price change (DWTIMP) by one pcriod. This is donc in order to reach our final goal, namely to find out how wcll this price va!'iation can be forccasted fürough m:ünly fundame1\lal physical factors of the oil
market.
Il. DATA DlsSCRIPTION
Since the variable or greatest imcrest. is the crudc oil futures price (Wl1MP), we chose m dcpkl its behavior over the time pcriod considered, along with the priccs of two oil products futures contracls u·a(led on the
NYMEX, namely Unlcadc,J Gasoline (HUMPR) and llcating Oil (HOMPR).
40 S,11,l,I
"'
0ê
i
~
,.
~~
0~
[![
'1.
~ ~WTI Monthly Pricc (WTIMI'}
'°
>ll~I.
..,
100 HOMl'R""
JO..
•
'
-
•
"'
"'
40'"
~·
0 0ê
~
œ~
i
~ <ê
!
~~
WTI •nd HO Monthly Prices (WTIMI' and HOM PR)
40 .$11::,h] t/i:t:~l ](~ 80 IIUMPR
•
.
\'
'"'
•
..
•
'
"'
40 ]() ,0 Q ()~
WTI aod HU Monthly l'rlces (WTIMP ,nd HUMl'R)
In these figures, it is easy to spot the peak in priccs from April 1990 to March 1991 due to the Gulf war. The smaller pricc increase at the end of 1991 was due to the crisis in the Soviet Union. Whm thcsc figures show very clearly is U1c high positive rnrrelation betwecn the prices of the crudc and products futures contracts, especially between crudc oil and lleating oil. From the following table, showing the characterization of the monthly crudc oil price WTJMP, it eau be seen that the correlation between this variable and the oil products priccs is cqual to 0.94 for HOMPR and 0.88 for HUMPR. lt is most negal.ively correlated with the variation in crude oil stocks (lX:RSTO), a.s could be expccted.
[~T!MI' CIIARACTE.~}.l,ATl(>aj
~---·---·· -·-
~-·
", ... --- --- ----··~-.
, .. ,,----f ICl-11'1.RN.~Usrrc VARIAfil.f.S I (:i:AA. 1 V ,'IES'l' 1 Pl'«-'.11':'A 1~---·--· ··-·-·--·---·· --". ""~---.• 1 WtlMP 1,Q(X) 1 6.;.8'.!8 1 .0000 1 1 fK.~B _944 1 l,3.198 1 .0000 1 1 H\"J!•(Pf{ .ea~, 1 10. 465 1 .0'..100 1 1 l<ESICC .582 1 4.9$3 1 .0000 1 1 IŒF~1<:0 ,580 1 4-si~;a 1 .0000 1 1 r.ES/IÙ' .57) 1 4.88~ 1 .0000 1 1 HF..b""'IT.C .5'..I~ 1 ,1, ~,09 1 .0000 1 1 ]·!UFf-tC[I .463 1 3. 753 1 .0001 1 1 !~""'ID.: . 444 1 ).571 1 .0002 1 1 GlM/PIM .:'!OO 1 2.994 1 .0014 1 1 J,r:GKC .290 1 2 . .2)2 1 ,Ql2H 1 1 H-r:F~ .221 1 1. (,.85 1 .Olj(J) 1 1 !(.tj/f*)P .2Hï 1 l.644 1 .0~,01 1 1 aUMl{) .214 1 l.. 623 1 .0523 1 1 HJP!Ul , 15) 1 ] . l$l 1 .1248 1 1 [ ~ .l'.')1 1 .988 1 . .1.6H, 1 1 (~'ID . l'.tl 1 .957 1 .Hi9J 1 1 (:!.Nr-UI' .11.l 1 .8)6 1 .201") 1 1 G'<l'l<:t, .10) 1 .r12 1 .2200 1 1 aŒ3'lU .096 1 :nti 1 .2360 1 1 a\S'ro: .075 1 -~it;,J 1 .2868 1 1 HI.JS/Hl,1P .OSl 1 .379 1 , 352.4 1 1 [tU\.Sl"O .01~', 1 .338 1 , )678 1 1 U-IU.S'R) .OJ~', 1 .112 1 .45~,4 1 1 F'El:1«°.û ·-.000 1 - .0$'1 1 .4774 1 1 CST/Cl'.N ... 022 1 - . lfi'.j 1 .IJ353 1 1 GM,/CJi',f' - .163 1 ··l-7.J;;! 1 .1089 1 1 G\.9Tl'.1::: - . rll 1 -1.292 1 .0982 1 1 <"1:rl.'/Cl.N -.;'n,9 1 -1. 662 1 ,048;] 1 1 Kl:-S/KE',P , .. :;!$1 1 -1.919 1 .0275 1 1 FIUMK> - ,237 1 -2.;n.l 1 .0135 1 1 [L"'"Rt)'"JD - . 324 1 ··2. '.)19 1 .0059 1 ;,.---·- -·-~-· .•.. '., -·-··- ---· -- ---·- -- ~-. ·'·" --,-- --·-
---1-From the following table, showing the charactcrization of the monthly change in the crude oil pricc
changes in the priccs of the oil producLS commets. Howcvcr, lhe cormlation is lower than lhlll found earlicr between U1e monthly prices of cruùc and proùucts.
··---··· .. ··-"""·' , ... _ ---· --· -.. '" '., ,._ ··---···-··--... ,"
··-
--..,
!C'l·ll'IFJ\C'If.l\lf."l'lÇ VMIAI\l..E5l ~J.~R- 1 V,11:81' 1 ))ftll;'A I
1 [WI'Uarr' I a.KJ1P1 1 [HtMP1 I fft:;,Pl:;"R l I ClNf'LTr1 1 <:R'i-:nr.:c1 I GÀ.P!-lè.(_!l I ll<.l.MJX11 I ClUMl·î)l l PEFK'Dl I H.)FR:Dl I nt:SIDl I Hi.JPP[.ül I r:cr~,"IDI I IT~/IŒPl I l]-ll,t:iI\.")l 1 \J=IP-,;!1)1 1 ["i,l';,9ID1 I G:il'/CINI I CA..91(.('.] i jj\J..C:'1("(:] I IB!.":.~10.t I ClMlf'IMl 1 !ilJ'Ml:l\tl I CST'/Cl.Ml 1 ~r.:6I1X'1 I IŒSIC:Cl I C..iA:?/GJ\.Pl 1 }(}::.IB(f)l I Hf::~9.n:",Cl I n::-1:::/l-ûPl I M!JS/HOl'l I tJ:t1FR.J I wrIMPl l RES/rE.1·11 ,·,···-·---·!" .1..0.}:.) 1 61.6.86 l ,t.).)(X.l I • 1:;l(, 1 .j' .\43 1 '(.()l)'.J I .,)JI ! '.L<l.20 1 ,({IQJ I .4Cû 1 ::i.i,n 1 .CûOl:I I ,:)8,1 1 :).004 I .001J I ,341 1 7,.r;:l<l 1 .OC!ol2 1 .".i3:, 1 2.~;aa 1 .00<1e I
.Jll 1 2,}A2 1 .Ol)86 1 .1/J.S l 1.85~ 1 .ô:\19 1 , Ul4 1 1.380 l .QBJ/ 1 .1:1..;1 1 .!l28 1 • J"/6.!l I .J.J.';I 1 .886 1 -18')8 1 .10;;: 1 .756 1 ,22'1"7 1 .Q74 \ .551 ! ,2fJ09 1 .l"):',9 1 .4)6 1 .]]1$ 1 .01_;:~ 1 .<101:\ 1 -3417 1 .Q.J) 1 .2"14 1 -40J6 t .o;:;6 1 .192 1 .4;;40 1 ,(106 1 .0'1.S l . .J.fQ.l I -.ûl.1 1 -.()7Jl 1 ,4681:, 1 - .041 : -·.30~, 1 ,)804 ! -.012 ! ,·.3)4 1 ,3769 1 ... oi;:i 1 --.s·.1:1 1 ,3oi1.2 i -.012 1 --=~:r1 1 .291:,7 1 ·-.0':16 1 -·.71.3 1 .2379 1 -·.rt-1 1 --.910 1 .1,!'n3 1 ··,J::W ! -·.973 l .J6S<1. 1 -,1~,:,; 1 -Llll'1 I ,126J I -.163 1 -L,221 I ,111.l I -.189 1 -l.4l5 1 ,078'.", l -.198 ! ···1.4.f!"l ! .0684 1 -.~01 1 "1,:")QB 1 ,0657 1 -.20S 1 ,,J..5.;2 1 .0616 1 -· .;;1:;1 1 -1.90) 1 .0285 1 ·· .:i:.:.o ! -2.708 1 .00:.\<1 1 1·---···- '·"·" .. ---- ---·- -··" .,. - -- - --- -- -~" .,.,. ··-
---;-In facl., the change in U1c crudc oil price (DW11MP) of one mon th wilh mspcct to the prece(ling mon!b is most correlatccl with a grnup of variables including ilot only the changes in the prices of lhc products registered in the preceding month (DHOMl'l and DHUMPl) but also lhc variables dcscribing the rcfinery opcrations in tl1at samc period (CRSTOC!, CINl'UTl and PEOPERI), the correlation ranging from 0.34 to o.54. 1bis is duc to the fact that in the rase, for example, where U1c two former variables (DHOMPR ,mù
DHUMPR) are positive, this will inùuce a very prompt responsc from the rcfiners thrnugh an increasc in (1)
the crude oil stocks (CRSTOC), (2) the amoum of m1dc input to t11e refineries (ClNPUT): and themfore, (3)
the percentage of refmery capacity opcrated (PEOPER). 11,ese steps arc ~,ken in ordcr to ensure the higher refining margin induccd by the higher prices of producLs. The induced highcr demand for crude woulù then
la tel' be reflectcd in an incrcase of the pricc.
The monthly variation in the crude oil price is also positively con·clated witl> a group of laggeù variables which includes the variables describing the production of rcfined proùucts and the amounLS of crnde and products imported, the correlation ranging from 0.1 to 0.3. At the other end, it is negatively correlated witl1 the prices of crude oil ami heating oil rcgistered in lhc p,·eceding month and to the pressure of refined products suppl y, as measured by the stocks/proùuct.ion ratios, registered in that same period, the correlation ranging frnm -0.20 to ·0.35.
ln the table below are givcn some basic statistics of lhe variables.
···-
--· -.··~
---~" ,. -·--- -. "'" ,..---·-., ~" ~----;-!.:l1I'-œl. 1 1'11NIMl,l-l WIXIH.« 1
~.-· - - - -··- ~,. '--·-·- - - -·· .• ,., .•. - - - · · · · - " - - - "• ',. 1
1-~JMPR 1~8~·,.cm 112.62 44(, . .8,J 9S4. ~i5
l·i::N!iR :i(i).20 11€, .(j'..\ 408. ~7 9:;:..sa <:RlMH) :;43"! .S\) 713, 9~ :,~9;';.oo 6/61),0(1
:; CR":-m:::'C 342840. ,;i;i 1:-008. (\!:\ '.i22 tV:, 1-QO :;,.1.14~;4g_(XJ
6 [C}'l,::,TRl lHl.f/;, &)63. '/6 ,·2:wr,:i.O'.) l.f)~36.(X)
.,
Pi~IMH:.1 20\~}. 4G ~!82 ,•17 JJ7).C'O ;:eo.!:\.oO"
l-R:rffi::ü 6090.97 f,if.:.l? 4709. OO 714::.cxJ'
HUS'Tl.:r: 155092 .9() 146~;·: .4~; 130864. OO lB336l.. O'..l .J.(! J;lfU:.~) 580.14 ~li!0.7.? -·10237. OO l.787]..()() 11 (fl.ff,.\..1) 693t\. 47 ,~,!3. 98 ltltl.'3.00 ?379, (X)"
(;A.9I'Ci; 223556. :"10 1079:!. )~i :mr11S.(:(J 2~;372'.J. OO l) [l,';/1...S'ID -555.22 r,.,:r;;i .(',/. -· 1()0')2. U) 21828.(û"
lfJl~ü :wi:n . 49 1'76.70 ;nJ7,()() '.l'.U0.0.:115 !Œ~ro:· 1 l.J.(;IJ.~)9 .90 1315'1.44 93'.',a;,.oo 141848.lXI 16 LHC6"'ID 1 -9? .92 7~)16.~;6 ,.;!117?,.0Q 11892.CXl
.17 Rl.';l='R'.:"D 1 937 .20 ·h:,.:to 799. CIO l.167 ,(X)
Hl Fil:~ 1 4440fi, 74 :J,Hn.;1~i 36388.(.() L;l3S\'i.Cü 19 œ.IiS)l.:) 1 119.59 :n-;J.6\l -4994. (X) 4;14;;.oo 1
w
l<:J:'l~ü 1 l215.9S 90.41 1C(}9. OO 1427. OO 1 21 ~!:"'S'Kt'. 1 '10951. 81 238G.4l 35607. OO 45531. (K) 22 fl<:ES'ID 1 24.24 1()()6 .1 ·1 -~):_\18.(X) )39Q.CX) 2'.:l C!Nf'IJ!' 1 l.:1'121. 76 497. ~,4 (1994.(X) 14461. 0(1 20 î,B'".A'l:f\ 1 s:, .48 3. 30 76..9'3 93.23 15 c.-::.·t/CIN 1 1S.:,6 1.10 23.22. 28.23 ;.;6 C.'S'T/Cll~ 1 64.()'/ !:\. '.\(, !;,]..JI:! 94..~6 ;o CTl-!/PlM 1 2,(f) ,11\ l..~n <1.04~/~ lllJS/llUP 1 2:7,.$:,i 1.~;2 .?.2. ffJ :!i).88
29 <~9/(',AP 1 32,29 ::: . .l~) 28.4ü J9. "/:', )0 !-t"S/l·üP 1 40.)(i •1.19 32.30 :"1().;:·1 )) [-'.f::S/1",EP 1 47, ~,(, 4 .21 )9. )(, (i;::.·u; 32 :KES/Kl~P 1 33.7~ :::.2;! 1·1.8:; 38.Tl ·t·---·-- -'• '"'
··-
··-·-·-·---·- ·--- ·-·~-,·-· 1 CWI'lMf' 1 .JI 10,63 -.$2.6;; 85.27 1 CHIJ!@! 1 2. J.} sa.11:1. 1 -,1s9.·n
256 • .1:~ 1 [H.":MF'l 1 il.3;?. .:",6.52 1 -10:J. t)'.l Li6.94 -1---,., ... , --,-·---·- ,···--·---·-·---- ...•For OODlputnlior1al t\'&~QNi, !~ prlre of crude I~ oe:Xpt't8!1ç() 11'1 $l&hllrrel nnd !llll pril'I;)~ ùf l)f(l(l~ct~ nre cxpres~tl il'l 10 1~lllMR~ll<;ifl. 'lbe p)ry~it.1\ r.l:tilll:11':.~ ~.te e~~ned la tl10ll~t1.0d~ .;ij l'ü.ffl:l~ per d.lly exœpl ~liX:b Wl:litll ~ft; !11 lhon~amls of b:itioi;lt, lliçfl:"/OI"<::, t~ prenure- ntlœ< (~H.W~~OO~c!hln) ;ue-e-1\preS!llld itl i;hy~ 1u1i;:t Ulf rullo CIM/PIM iff
dl.n~Jllàcn~.
III. l'RlNCIPAL COMl'ONENTS ANALYSIS
Ill·!. DESClUl'TlVE FRAMlcWORK
G]
32 Active VariablesWc swrted the desc.Tiptivc analysis of tl1e <>il futures market with a principal componems analysis wilh all the variables for which we have gatliercd data activatcd. In this descriplive framework, the goal is to seek the rclationship belwecn the tlme behavior of the monthly crudc oil futures price and the physical detenninants
or
1he oil market. As can be seen from tl1e resulrn below, tlic quality of the rcpresentation achicved with these 3Z variables, as measured by the perœntage of variance cxplained by thefirst
two principal componenLs, is somcwhat poor. The quality of the representation of the indivi<luals on the phme is also, in general, quitc poor.... -·---1-- -- -- ·- ... ,. '"" ... ---+-·· - -·· ""· ' ''". Nl..M I ElGENIJAIJJ'E I .;.- - - -1"---" -···· •• • ,., . j - - - --+-··- -·· ·•- '"" w,t 2 3 4 [, l{), l,9]l 4_9f,(,6 J. 797.S :)..$527 1. 9711 31. 85 15. 52 11.87 11.1.0 6.16 )l .SS R.Xl ~/).2:4 "J0.3/l "16.50 6 7 8 0 lD 1.6699 1.264.? .9089 , 754<1 . 6041:1 ~.27. 4.01 2.84 2.36 1.89 81. 72 a;;.;3 88.$'7 90.9] 92.82 ,;.- -- ---1--·-- -··· .. ~~----4.--- ---+-·'· ,·,- ---··
1 \lt',RlAf~J:.'."I 1---··-! Wl'.IM!? ! HtNF~ HGIIB C'!UMJXJ CRS"Il::C r;:cR.S'II) PIUMFO r!Of'îil'.";,l) ti\k"1It)::: (iil.l$'K.1 Gr"• .. :N·ûf) t GAS'R):.;· !'.J::1",,HJO HC.T-'!'DD HC.sll)'..~ [HX:-11:J RE'.PF\'.:.il) ·-.1:,:1 ,(,(\
-.:w
.. -~,,, , 33 ... 49 .. _:-;s .o.; -.:n 81 .. ):/ -.;;io · .")!\ ·-.:i4 ... 40 .Jt, -,JO ··.115 .O!l ··,31 .44 .89 .0() ·-.2Q ,(,0 . '.ll . 3:1 , .Hi .28 .73 ,71 41 ,)4 . 70 . ·.10 .2B .26 .'.lO ,74 ... 67 ',0(1 -.'..-~! .~:, . 67 ,. ,)8 ·. :.l4 - , 1.2 .1.6 . 47 .::!~ . .0·1 ! lï.E::~l'T.X:: ! r:r,'E.S'I.tJ ! Kf::P~Xl ! }.'T~-::I\ • .C ! ["N::~~:io I C1Nl:'Vr ! Pll.:lll':.1{ ! GST/ClN ! CST/ClM CJ.l,\/F'IH ! f,lU:i/!.fUP GJ>S/C'-.AP : HU:.;/HJP f(f.S/fŒiP ! Kl=sm,:r,
.. -.67 .5(; -.(Jl -.41 ,,().] .~\8 -.68 ,'.',9 .11 ·-.46 .~,] -.21 - .26 .Q;-1 .01 -.77 ·--~o .1e .Hl .. , 11 . l.l .'.18 , FI ·· .. S6 .ÎÎ3 .)0 .O"~ ,(,7 .t'l:1 ··,44 ,(;J .::.ci ... 19 .in .40 .O) ,(JO .6S .43 ,<\() .6'.l ..• 09 .:m -.20 -,29 ·---'' --- ··---~ --- -·+1 Al..191 1 11.w_; ·-\.JO - . (;.8 3. '.< .1 ,i, 1 .·n .(lol .Ci'.! 1 SE~l 1 .,4. 8) 1,;:l l./(, .\. 9 .(i .7 1 .7() .04 . 0~) 1 CC91 1 .. ;;,28 2-41 ·· l-~:8 .9 2.0 . 7 ,;)(, _;?.:', .07 1 N")91 1 -2. 38 1. 6(1 ,· l . ;~(i . ~) .9 .'"/ .20 .():! .()Ri 1 [f..'91 1 -4. 30 1.66 ;:.61 3.] .'J '..l.Û ,48 .07 . lB ~ ,. ,-·,··--- ---··· -·· .. -.. -·---·--·· .. -· -- -
..
,., .. ,·!Despitc the fact that the qua!ity of U1c œpresentat.ion achitvc,1 is quite poor, sm.nc inteœst.ing features of the oil market can be dcduced, and thus smnc insiglll gaineù, from the rcprcsentation of variables and imliviùuals whicb follows.
·t--·--·- .... , '~ ., ... - - -- -- --- -·-· .. - - ··- .. ,. , " . - - - ··-- -· ---1
I VARIABlJ.: 1 Mli:i\N I QJ'f. [EV I CT.l)RDl.W\'lE l
.6 ~ ... , ,----·--.. ··---·-·-.,
.
" t c.~:n'/CTN I KFS/1.:J:J'• I lfüS/Hi:.JP I f,;.\.~:!t(: I c:~n'/CIH I Gl'I.S/Gl\.l'' f···-·· 1~,, S6 .3). 7'l 2:,. 5'.) 1::1?3~}:",6.SO i,A .07 :l'.),;'!f! ·I···--··-··--·--· ... I VAH.lM!,E I MF.AN I CINFlrl' I FB.::t'.!:}! I G-\J:t,'{'.() 1 [H:.n'ID 1 (RTI1FC! I PRThffi) l.'.3(t,~.l. 76 J:5.18 6038,47 --97 .02 1 '.i4:rî. ~9 I Jû5}.lj6 ... ,' -.. ··-···--·---···-· .. c E N T"
A ··-·--·----·-- ... '' ···! J . .10 ;;.;.:.i 1- 5~: 1(17:1;; .. 17.i ('\. 36 .'.;' . .)9 . :w , 61 1 ,7(1 1 , 7H I .~1 1 $l.Jf'. r:1:.v I o::c~<111NA'l'E 497 .54 3. ]0 24}.98 7916.:")fi 713, 91 21.(),•17 I'., z 0 N E -.so - , 41 ., ,4.1 - ,38 - .3S ., - .. 11 1 . ···-···-·-·!~··
.-·--···----·--·-··. . . ··-·-····---··· ,·-·-····--1 1 !K.:Clt'.t:; 11.16199. 90 1 l}JS4-44 1 !{FJÎ:'h{:(') 1 121~).9(i 1 ()Ô.<ll 1 W!'J)1P 1 201. 56 1 4J. 66 1 1~1:,~;/1'1:l' 1 4?.56 1 4.2] 1 IK)5/H)l:' 1 4()_'.\6 1 1,.19 1 l'f.::.MPf' 1 ~,63. 20 1 .1..lf,.Q} 1·---···'' . ···-··---··-·"-•• ---.)MPH ~;/HEP KEPR.Ri~S1WI'tMHOS'1'()C KE~>Tcx; !JOS:i'I'ù(: ,S'.i .~,9 ,fi(.! . 6) ,6::, ,1;5 HtJ8/HlJP GAI;'!'()(: GAS/GAf'HI.JMPR ):X.';ASDl:ll1S'I'(1 C::S'I'/CIH
ç:,;'l'/C.1.N . 0 HllPHOD -€-IM/PIM ~ S T \ ) HOi?P.OP CR.l.MOHùS'l\'.:J P.E})F'GAPHOD ~.'INPUT ·- . .!:i DRESTO REPROD CRS'T'OC ,Ü
A close study
<Jf
this figure allows one to definc four differnnt directions :• North-East : Direction of Pœssure Ratios
HUSIHUP Unleaded Gasoline Stocks/Unlcaded Gasoline Production
GAS/GAP TotM Gasoline Stocks(fotal Gasolinc Production
HOSIHOI' Heating Oil Stocks/Heating Oil Production
CSTIC!M Crude Stocks/Crude lmports CS11ClN
Crude Stocks/Crude
Input • North-West : Direction of l'rices and StocksHOMPR No2. Heating Oil Monthly Pricc WllMI' West Texas lnte,mcdiate Monthly Price HUMPR Unlcaded Regular GiMline Monthly l"tice
HŒ1DC Unleaded Gasoline Stocks
H0~'1tX:: Hcating Oil Stocks
RESTO<:: Residual Fuel Oil Stocks
lŒSIDC Kcmscnc Stocks
• West : Direction of Rcfincd Products Production Levcls
HUPROD Unlcaded Gasolinc l'roduction
HOPROD Heating Oil Production
IO
*
South-West : Direction of Rcfinery OperationsCINPUT Input to Crudc Oil Disûllaûon UnilI
PEO!'ER Pcrccntagc of Refinery Capacity Opcratcd
One can tl1en sec Illat prices arc more correlated witll stoçks than witll pro<lucûnn levcls, rcfincry
operations variables, or pressu,·c ratios. The 180' angle bctwcen the directions of pressure ratios and refincry
operations rctlects tlrn high ncgativc cmrelation betwecn the.se two groups of variables duc to the fact that whcn stocks, comparcd to daily production levels,
bccome «Jo
high and represcnt many days of production,tl1c refincry opcrations will tend to slow (lown U1rough a decrcasc of the crude input 10 the refincries and the
perœntage of rcfmery capacity opcrated. The graph shows that, as a consequencc, the mnounts of crude imponcd (CRIMPO) a11d total gasoline produccd (GAPROD) will also tem! to fall.
We now tum to the represental.ion of the 59 mo111!1s uscd in this analysis. Four points, mo far away from the centre. weœ hrought back to the border of Ulc grnpb. Two of lhcse. OC:90 and N090, arc from the Gulf war pcriod. In the graph, the Gulf crisis period distinguishes itself vcry wcll with the corresponding months far away from 1h,i mean.
AXE '.l AU!lO DE:91 Sl::91 ·191 Jt;'"ll 1 . - - . . ~ ~ ~ ~ ~ ... ;,,e9 -6 -~~ ()(:91 F'f:90 N\)9l () dl,88tlY89 FE:!37 MA87
n:o.:,,
AP . ~ 8 7 "' "· ,., ···MA(l9-- - -·-0 AXE 1A close study of the graph abovc shows that it can be divi(lcd into four quadrants corrcsponding to the four seasons :
FALL WINTER
SUMMER SPRING
These quadrants are, in genentl, crossed clockwisc by a line joining tl1c various montlls of the ycar, as the exatnple of tlle lincjoining JA89 to FE91 shows. Let us rceall tllat we found four directions dcfincd by four groups of variables :
Prices & Stocks
Rerlnery
Operatlons
Pressur(!:
Ratios
11 - In summer, the refineries opcrate al high capadty 10 smisfy the high demand for gasolinc in the American market The summer quad111nt is crosscù by the rcflncry operatio11s direct.ion characterised by the crudt\ input (C!NPUT), pcrccntage of reflncry capacity opcrat.ed (PEOPER), aud total gasolinc production
(GAPROD). Thls high lcvel of rcfiucry operat.ions also bclps build stocks of heating oil in exp<.,clation of the
conling cold season.
- Bctwecn sommer am! fall, wc sec high levcls or heacing oil and unlcadc(l gasoline proùuction as shown by the westward direction of !lie refined proùucts pto(luction levels. ln fall, the stocks are high and priccs climb in response to the increa.sing demand of U1c cold seas()n. The fall quadrant is crossed by the prices and stocks direction defincd mainly by the ptiœ of hcating oil.
• Bc1wcen fall and wimcr, the pnxluclion of reflncd product, starts to fall and ti1ereforo Ille stocks, in tcrms of daily production, increasc. We end up in the wintcr quadrant which is crossed hy the direction of pressure ratios. Going from wintcr to spring, thesc ratios start to dccrcase when production picks up in spring. The production of refined pnxtucts continues to increasc, especially that of gasolinc in expectation of the coming increase in dcmand, until we reach again the high level of rctïncry operated capacity in
su.ramer.
Œ}
6 Active VariablesRecause the quality of the rcprcscntation, as mc:isured by the perccncage of variance cxplained by U1c fim
two principal co1nponcnt.s, achicvcd in the an~tlysis ahovc wilh 32 variables wa.s soinewhat poor, we chose
to restrain the numbcr or active variables to six : WTIMP, HUMPR, l-IOMPR, CRlMPO, CRSTOC, :md
CINl'UT.
As the rcsul!s bdow show, U1c quality of U1e rnprcscntation was very much improvc(l. 1-lowevcr, rnany parts of the picture drawn in the prcceding analysis arc now missing.
11,e
active variables now dearly dcfinc only two of the direct.ions round earlier; namely the direction of priccs and stocks, and the direction ofrefioery operatiOU$. As a conscquencc, U1e seasonal discrimination in lhe reprcscntation of individuals,
especially betwecn wintcr and spring, is quite poor. Nol much insight into the problem was gained through ùlis analysis. t·---••• " • ••+" •••• "•• "" ""'"• ••+•• ,•" "•• "" " " " T " " " ' " " ,• S " • • • • t NllM t ElGEhVID..JE: 1 ,r., Ct.N.% 1 -~-··--- -+-' ,·-···-·· ---- - .,. .. , .. , "" -- ---1---- -· - ... ",·-·-r ) 4 5 6 :i.l,l9!.i :,:;.oo 52.00 l.\!{.r.:7 30.4S ~ .629:: .218-8 .178~\ .0269 10.49 ) .65 :! .98 ..:15 9ï. .93 96.SS 99.SS 100.00 ·1·----·-·" ., ·~ --- -- - ... " ' "-i-··· --- - ---····--!- ,. ~ -- - ----T
[ C()ORDI~ATËS ON. AXRS l
!.!?::21
I\F,\F,:IAIIE.EI ) 2 1 ·--- -~---···-,,---·r
'
wrIUP ' - .!H:'. .49 -.OJ'
r!UNH~ ' ·-,91 .25 .05 /f'.'.'.NF}\ ' • .84 .47 .01 CRlMKi'
_, ,(,4 - .65 ,·,24'
CF.t~tr.X::'
-.44 , .. $9 .68 ' ClNJVr'
- . 51 .7'.l - . 33 +· "·" w~··- --- ·-··" • .. , , .. ,"" "" --- ""'"' • ., , ... -;,+---N---~~---w,·--- ,.~---~---,,---~w---+
COORDINA n;s, CONTIUllUT!ONS AND cos2 OF INDIYilJUALS ON AX~;s l TO l I
+--- • _______ ,,~---~---•~---~w---~
,,---•·-+ 1 ]Nl)IV I c:n;JR. 1 l[ffl 1 2 :; ;; l 1 '.! ) 1 "I"""""••• M " ' ··---···--" ,---··· ''"""'•··---··--··.
··-·---··· 1 FEB7 1 2.6:l 1. fü\ . f.). 1 J,f\ :1.:l 1.0 1 -64 -~·- .04 1 1 W.B'l 1 2.82;~.r,
1.:n 1 4.:) '7 .1 4. j 1 .46·"
. (1..1 11 'i'PB'l 1 2,)7 l..96 .so 1 "'). l ),(;
.-,
1 .56 ,)A .03 1 1""'"'
1 l.Sli l .. füJ.
"
1 ). \) 3.) -l 1 .48 '5() .(X) 1 1 J!JP,7 1 1.0~ l..Ol. -.68 1 .l'i .1) 1.J 1 .4.:'s .JG '1(1 1 1 ~Jt..,87 1 .39 .52 - ] -~il 1 .1 . J 4.1 1 .Q'J . 13 , 71 1 1 .N..187 1 .S9 . u -J .û4 1 .J _() 2.9 1 .,;n . Ùl '7:l 1 1 f:F.87 1 1.0:?. .38 .. , 14 1 .6 ,1 .l 1 . 75 . 10 .01 1 1 (;Cil? 1 .9f) .97 ,'.-{) 1 .'., .9 ."! 1 ,41 ,l\:l .11 1 1 1.zy\'7 • 71 - .03 1. 12 1 . l .a ) . 4 1 ,,:,3 • (J(J .5fl 1 1 11.=.:wr 1.37 ,. ,29 J..'.lO 1 1. 0 .1 4. ~' 1 .i;\'.·! 0'' .:w 1 1 JI>.88 1.81 ,)Q .s:1 1 1. H .1 .1 1 .8l .02 .07 1 1 l'Ee.8 ;'!. .2:. .t'.i7 .:'.il)'
2,8 . 4.,
1 -8!~ .08 .04 1 1""""
2.o;i - .10 -49 1 2.2.o
_., 1 , 91 .()0 .o:. 1 1""""
.97 - .48·""
1 .5.,
l.,.(l 1 . 46 .l,;i .35 11 MY&\ .(il;) -1.20 . 93 1 .3 1. 3 ?,.:l 1 .16 .Mi .29 1
1
Juan
.97 '1.4!'1 .86 1 _i:, :~ .o 2.0 1 .20 .47 .Hi 11 éJ1.8H 1.20 ··1-42 .2f, 1 Ji 1. ~l
.
;~ 1 .'.14 .48 .Oi: 11 Al.J&?, J. .41 ··1.39 - . 64 1 1.1 }.8 1.1 1 .'.19 .33 .os 1
/\XE .0 -,6 1
'"""'
l , LJ ~.i "/6 -.02 2. 1 .1.i"'
1 .!l:.1 .13 .o, 1 Œ:Afl .l..)7 ~;g -· .22 1. (1 . 3.,
1 ."i<l .1.:3 .OJ1 ,JA89 l.?l .14 - . ~l'.l .s .o 2. 3 1 -~;4 .01 . 32
1 FF.1:19 1 l. ~,) .39 -.62 1. J .1 1.1 1 .BO .o;·, . l:J
1 !1l\.39 1 .94. 'a;:: ·-.()'1 s .6 .0 1 , f1) .4) .(X)
; Ae'fül 1 -· . .23 , / . l ··.Vi .0 _(; l ,., 1 .OJ .:io . 1~
fü[l\) 1 -·. "..\l -/~~\ - .18 .) .:\ .1 1 .07
-~,,
.o:: .)i,IB9 1 - .2'.5 ·- .96 ·,H;; .0 _() L(l 1 .02 . )() .;ir:,JlJ\9 1 -·.26 .. l, IJ7 , 1,7(i .ll :!.(J 8.'.5 1 .01 _,jQ . ~i7
1 A\JB~l 1 .0:1 -· 1.·l·l · l, )8 .o 1-9 ~;. ;i 1 .OO . 51 , 47 1 t1.-::.1l9 1 - .SD -.96 . l. :w .l
.,
~:' •. 2 1 .08 .29 . (i~! 1 (1..':!l9 1 - . 37 -.56 . .(;{}.,
.3 1.0 1 .09 ,J:;! .,..., ;< 1 1-I:189 1 -.H; - -~.9 ... 1,J .o . ) .0 1 .02 .28 .01 1 tt89 1 - .\J'.l. .26 ... 13 .) .0 1 .28 ,O;J .01 1 JA.90 1 - l .JS -.2) ... 32 1 1..0 .o .'.l 1 .76 o·, .04. 1 FEl)O 1 ... 66 ·-.;:N - . ;!./. 1 .2 .1 .l 1 .6~', .).l. .O'l1 Ml'a.90 1 ... 19 ',)8 .ile 1
.o
.1 1. 7 1 .o~ .08 .371 11.?90 1 - .42 ·,1,Q)
1.·n
1 .1 1.0 7 .9 1 .o, .21 .6:'.l1 M'i'90 1 -.81 -1.83 1. 7r:1 1 . 4 ] .1 8.2 1 .09 .44 -40 1 ,'JU90 1 ··.tlll -2 .93 1..58 1
_,
8.Û 6.8 1 .06 .71 .7,l1 ,n,90 1 -1.80 -3. )8 .94 1 ] .B lü.6 :L-1 1 ,21 . 7) .06
1 M-190 1 , 4 ,li4 - i. l.+1 .6S 1 11. 7 1 ") ].] 1 .91 .o:, .02 1 ~è."90 1 -ri .a7 . 7S . 3~, 1 18. 7 _f; .:) 1 .9' .02 .OO
1 CdO 1 ··1.76 ~.<1J . 3~, 1 12. 3 10-9 .:l 1 .65 • 34 .OO
1 t,.r:f.<l) 1 -3 .51 3. ll .;;:i.'.1 1 f.. 7 B,9
..
, 1 . r,r,-"
.OO1 r.P'.JO 1 -1. 58 :LcJJ; .. l$ 1 '-' :i.6 . l 1 . ;;~!l .i'ifl .OO
~---.:90 N090 3 SE90 DF:~JO JA9.l
FE:ll JUB'" Afl9
oC:9 .l.N091MA9:î OCl:l'"!
i\Ptr7
MY8-~FEf:\7
AP91AP!~9,J1.J87 r,•ggg
Dl•'.89 AlJ~7 SE87FEG9.JA88
0
1---l'<'.'l+-.
N 0 ~ 9 MA88(~iJA90HY9:lr'E90Mi\90 AP88DEH? ~!E88
s:r9 lC:! t.S"I .l.OCH 9NOfJ 9 DE(l HN()88
1'fYfl 9J\P9 0Jl18 9
A\J90 A\)91 Sf~fl9 AlJ89MYfH)SC86AllB8
~iU9:hH,89 ~ ,188
MY9ù
... 3 ,'JU90
, .. (i -4 ()
IIl"l. PRElHCTIVE FRAl'vŒWORK
G]
35 Active VariablesJ3
In t11is predictivc framework, U1c goal is to seek the relationship between the variation in U1e crude oil price
of one month (DWTIMP) and the physical detcrminams of the oil market in the preceding month. In a füst stage, ail the variabkls available wcrc activatcd in the analysis. The rcsults below show that tlle quality or
the representation achievcd with these 35 vartablcs,
a.,
measun:d by the pcrcentage of variance explained by the first two principal components, is somcwhat poor. The qualily of the reprcsentation of tl1e individuals on the plane is also, in general, quite poor.1 VJ\RIABU:: 1 :l i·---'" "~---·-~-·~---·---+ 1 [".Wl.'JMP . Ll - . S3 .oe
'
WJ'IMPl ··.7,'>2 • 61 - .J~i'
C11UMP1 .Ôl ... 33 -.2? Hl,NPRl .. ,$9 . 33 • _4f;'
CH.:t>!Pl ··.;:14. - .24 .06'
fl(l,!ffil .. ,•;6 .65 -.i$'
C'l:!.fW>;"Jl , .at- -- 31 .•. ;Jl'
C?f,'ICC'l ,·.4i -.:::s - .3[!'
CŒSIDl .2.J - , 28 -.4S'
PR1W"{"J1.o,
- . 37 .Ji1 -t- -·- -., ., .. -.+- --- .. , .,. ··~ --- -- ... ,. -.---- ----;-'
lilJPfC01 ,, .. SH .03 .,._;nNlJM I F.JCf}lilAWE I O.lM.~ l ' HtlS"Il.).;'.l -· .58 .32 "·,39
,,---+· "' ~--- ---- -;--··· .. ---1·---· _._,. "---,1
'
PlUS'Itl1 - -l~ .25 .72 10.1848 2Y .10 ;)9.lù'
GAPR.Xl1 -· '1~~ - .41 .J62 5. '.!<1~14 i;),2'? ii.Jl 1 GA.91t:Cl , 71 .2il .26
) 3.8702 11.06 ~,~:-.. 4'.l ' CŒ.S'!l:ll -.20 .28 . 73 4 :L fE46 1.0.S'.! füt96 ' HŒ'R:Vl 1 -.66 - ,(fi . l l s ;::.4028 f',.Sî 72.a:.i
'
H<.:mo:1 - .:)4 .55 .SC"
l. 134[',;J S.27 78.0~ ! œ.)f-Jll:)l - . 67 - . '.î2 .32-,
1. 717F:i 4.9.l 83.00 ' R!'PA")[Jl -.22 -. l6 -·. 06"
1.0378 ~.97 tlS.97'
RES'"IO:~.l -.63 .$'1 -.04 9 .87'..19 ~:.so 81,1.<)t:)'
~ D l 1 -.40 .OO .4:'..\ 10 .'}917 ~?.26 90. '1'.3'
"""°"'
·- .64 '(il .08 T- - - --· "·--+- --- ·· - ·· ,+---····+ w·,· - ---+'
KE:5,KCl - .<l3 • ~,4 - .l.J.'
[ , ~ l ",2$ .OO .10'
cu.n:uri --.80 -.'1.8 .:;:o 1 Pl:PFf.Rl ., ,A] -. 39 . l.4'
c.~1'/CJ.Nl ! .27 .\6 - .56 ! C5'l'/CIMJ ! .78 .%$ .09'
CIN/PlMl ! - .65 .09 -.42 ' 1-{ll.',/HUPli .61 . 47 -.22'
C-:l\.S/GAPl I .82 .)6 .OO'
\l .. "l,/HC;>l! .O~, ,(i)-"
' îŒS/~F.Pl! ., . 38 .67 .OO'
:KfS/KFPl! .JG -.19 - .20+··---
••---M~-·.·---4
COORr,INATES, CONTRlalJTIONS AND cos2 OF INJ)JVIIIUAl,S ON AXES 1 TO 3 1
+·· · · · · -I INDIV I 1 -··-·--- ---·-··- -· - .•• I J.[:EH I 1"AB7 All.S-.1 MYS7 JU8"/ ,Jiif.l AlJ87 t'-f:"27 r:t:H7 N)87
ms,
J.~Be Fr"~"'""
APM I Mif!fl 1 ~IUf!B 1 ,JL,8.!J I AU8$,
-,._·1a ::-.19 1.49 1 9.1(, ~.14 ·-.97 1 li-O~i .'rl . 6S I t,.?.\l -.61 -.86 t 2.i!:/l -1.93 .)9 .96 -1.56 L88 .94 -.71 1.5) l.]5 .80 1.6..'I 2.J.4 1.14 -.02 1.44 .68 -.70 ,fl,1 r:;·, :::.OO 2.40 .74 ;1.-.14 4.1'.1 1-43 J,.6l 3.86 ,,7S - ."1:i ;1.61 -'.1.53 -1.K; ê.6 -l..80 ··,91 16 -).11 --.6~, .S3 1 ~E8B ! ccas I N:)88 ."37 -;1.9J ;:;.;:;. -·l. .92 2.?1 .. 17 --.lS '.1.89 ,(,!=\ ··.94 l.?6 1.4$ -J.7,!l l,il.8 ·,'.ï7 -l..!lS ~1.74 .21 .22 4,.0'1 Eli:88 JA89 FF}!9 Ml\8~ APl\9 M'fW:1 ,Jl.l!J~\ ,~L.89 Aù89 I ta):) I CC89 t J-Œ9 l me0 1 ,J?\.90 ff.:90 Wl.90 .l\,P9Ô M'.1'90 ~1090 .JL90 AUOIJ ~'90 cdû NJ':'.lO (l)-)Q iJA91 I FE91 I M'.91 I APJ1 ! t,r.(~)1 l ,;JJ.J(Jl I JI.91 I NJ91 I t:l:.""91 I CC91 I u;B1 I tl:1Jl 2.49 1.)3 2,7(! 4.)4 ",47 -·1.4;; 1.SJ ... 2.03 ·'2,44 1.19 -~.81 --1.49 -1.67 -2.62 .70 ··2,7$ -2..49 I.77 1 ··?..JI -l.64 1.11 1 ··"3,73 -·J-~;fJ .6~, I -1.81 .79 ,(;") l -2.45 .4) ,-.Qf, l -1.43 ., .82. -J..7,0 1 - ."13 ·-1.45 -.67 1 -2.6'1 2.52 3,)4 1 l.92 1.63 -·),)"/ 1 .9f) -LOS -5.18 I --.;n -i.1:; -'.i.25 1 ,,;i.;:9 -4.37 -1.16 ··4.92 ·,J.23 -.$1 .,6,2 .. 1 -J.Hl -1.16 ··6.90 i.06 .43 -4,13 7 .61 ,-.:A -2.83 6.38 -1.39 -l.'71 5.24 .39 .·,23 4.f:io -:;.os ".14 ;?.."Il -;~_34 .t.21 -.al -4.JS ,16 ··1,07 u2,t)!) -2.11 1 ~,,, ,. ,81 -4.18 ··1.0) .31 1 -.:,1.r1 -.a~·. -.82 ! -<).(:18 -.85 -.61 ! ··4.lM l.5~1 l.'?5 l -2.38 2.51 ",9!'i 1 -:1.JS 2.09 ,-1.H) .j.a ~ · - - - · · · · O:NlR,'
3 1 10,} "J.3 1.0 14. (; 1- ~, .4 6 ,, -2 .2 4, 7 . 1. . :3 l.. J t .;i .1 .2 .8 _f) ,J .;?, J .O.
) .;;: ),.:} .8 1. 0 . 3 .1 ,0,,
.o LO _,"'
.2 3. 3 ~.9 .7 1.2 2. 5 .2 ,] 1.;1 ;~. J I.L ,;; J. .Cl .4 l. 4 .2 ,(1 '2.7 - l ! ,() 1.;; 3 .1 1 .o ,() f,.7 1 . 5 . 3 1-4 1 . 4 1.. 6 1,Q I .1 .9 (,.i I .û .o 7 ,] 1 l.l .6 3,1 I 3.'.2 , l ,g I .<) !.] 2.7 1 . :;:: ;;. ~; l.Ô I . ~; ;:.:; .2 1 1.3 2.0 l-4 1 .8 .9 .G ! 1,) .5 .2 l .6 .2 .2 1 1.0 .1 .o I . '.! .2 . 6 I . l . "1 .1 1 . _, ~;.1 s.o I . fi . 9 5.1 1 .2 .,), 12.0 1 ,0 ,4 4-1 1 , 9 1.a .,, 1 •LI 3.4 ,1 ! 6.S 1.5 .G ! 8.1 1.4 .J I ,~.$! 18.7 .1 I 1.4 U.l .9 I .S S.9 -l I .0 6.8 L9 1 .o 2.4 2.4 1 .2 .2 B-4 1 .0 .4 ).1 I l .CJ . s . ) 1 j.O .3 .0 1 l. 7 .?. . :,l j 4. 2 .2 .;~ ! 4.0 .7 1.4 1 J.ù ;!.O A I .9 l .11 .5 1.,,,,---+
n:t:·;.:::-1,J1. 1 ,,,.,.j ."ll .12 .03 1.r .ùi1 .01 .7) .01 .01 .6Z .01 ù'·' .L9 .16 .01 .os , 1'1 .21 .06 .01 .17 .11 .04 .18 .;;:?. .. ir.i .(X) . l.O 0"' .O:l .Ql .àl .15 ,17 .o;.: .2.:1 .54 .Ô6 _{)(\ .55 0'1 .û2 .32 .)0 . .12 .18 . ;:1 . O':i .o7 .24 .o;: .Ô() .3'., .01 .OO .1<1 _;:;g .01 .Où .57 .. J7 .OS . .1\J .09 .22 .09 .0.1 .08 .13 .OO .Q{l -44 .16 .0$ .;~o .44 .O.l .0$ .07 ,12 .lB .06 .)5 .10 .13 .32 .02 .J.'"J .22 .11 .~8 .1.7 .08 .4] .10 .02 .29 .o~; .ô4 .29 .01 .OO .O~, .02 .01 .01 .04 .O.\ .05 , 11 . l.9 .11 .os _3;J .03 .03 .75 .OO .06 .44 .là .LO .OS .4J .. 1::i .OO .4,i .05 .01 .61 .û:.; .OO .19 .64 .OO .13 .G7 .03 .07 . 6~, .OO .OO .6:1 . .l'.:! .oo .18 .13 .04 .o;~ .56 .OO .06 .:w • .3(, .09 -~l<l..~,O .o:~ .OO .66 .05 .04 .78 .02 .01 .65 .06 .09 .2) .25 .04 .18 . .14 .()4 ··-··--- --- -- ---... ,. "'" ··-·,·~
Œ)iï;iù:s•;
NT A
1:_I<:>~3iIYAl!IA Ill, ES A
Ni:> ..
!.~-~ 1vi'iii:w,s]
i\.XF.. 1
+--- -- -- - - ··- ., .. , .. , ., --- --- ~+
I VARIJŒJ . .E I ME!:l\!'-l 1 !i-lJî'. Pf:11 1 a:.œDJN!'.'lE I
AXF.: 2 . 6 + . li:[,;'l'{X:::l I ID:"IPERl I CRll1I·l.]1 I CINIUl".l I C'>11.P!:;J:Ol 1 [I.JC.8'.l!J1 1 !·Df.~.01 a:: .. 47 ;,4.::i;.).Cr9 J},4;:;0.31 6931. S4 -181.98 2877. :i] :!.Y3 ·11H.îl'J ~,ô:i..70 7.40."}9 79HE.~.,'î 1.7:;_ 7~:, ... 8) .. 8.:; ,,BO ... 72 -.67 - .66 .... ···t CEN'T'RALi.~ t..r)N.E ',·- ···- - - - -·-- - - -·- ··- • •• I" I [('l~i'"DJl ;r,9 _ 9i! I Œl'/ClN1 2'.i.':,7 1 1ŒS/KE!'1 33.Hl J )'1U9/IRJPl 2~,. !:,6 l GflS'lO.--:l 1:tD''/16.70 ! L~'T'/CTM1 1 61. :16 I c;:;,.s;c;;1PJ. J:~.'.H l'.i.."J:W.'17 1.10 ;Lï.11 1.::,:?. 1.0814, 9~; f).40 ;i. )8 .:n .37 . 36 .61 . ?l .78 .B:i ·!·· -···-·· .. -·· - - - -·- ... ···-···-·-·- - - --- -T" .\-··----·---····-·-··' • - - - -.j.
1 \r'i\RJJ'.IiI..E I l·IEAN I ft.lr .u=.v I CtXFDil{1\.'lI:: 1
s • ()(~90 0 0C9J. DE90 ~1A9.l N091.MA90 Dt;:9:\ MA9i
Nl1S7
. Af'!}OMA.89MA(H3 N089 (' 7B7 DE89 r·i;:s9JA8f:!DEB7Fr·;88 ~ .. .!,JU91 ,JU90AP91SE:H7 Af"'H9JUîf7
sr:;9 J,JL91AU9 l,J.l\90,JA89MY91W.90NOÛÜ APBB
1.)Cf39SEH 9Fh:9 O,JL88AfJB7 JUt.18
SE(jü JL90 SEB8DElH3MYH9,:n ... 87 A?87 MY87
I
AUS(),}L89 1 MY8ù .... --jbJJ~îl) • ,J089 - - - ' -4 0 1 AXE 1This new principal components analysis actually doos not add much to the basic picture we draw rrnm the first analysis. lt simply defines one more direction in addition to the four dcfined cartier. The ncw direction is ùie souùiward direction characterized mainly by ùie variation in Uic crnde oil pricc l)W'l1MP. The graph represenling Uie individmtls is still comprised of four quadrants corresponding to Uic four seasons, which arc crossed clockwise by ùie linc joining the varions months of ùie ycar.
FALL WINTER
SUMMER SPR!NG
Prices & Stocks
Rcfinery
<.>perat.ions DWTIMP
i:1
ressurc Ratios
Along this n~'W direction, which fulls bctween spring and summcr, wc also find the variation in the price
of unleadcd ga.soline DHUMPl registcred ill U1c prcceding month. 1bc picturc of 1he oil market sttucture is the same as the one drawn cartier and lhis new direclion entcrs simply to rcflect the jump in the gasolinc price in rcsponse to the incrcase in demand in sommer. This is followed by the jump in Uie crude oil pdcc when demand for crnde incrcases duc to the fact that input to refincries is increa.sed in Ortler to meet Uiat demand for ga.soline and to
build
up stocksof
hcating oil in expectation of the coming cold scason.llij
6 Active VariablesAs bcforc, becausc the quality of the rcpresentation, as measurcd by the pereentage of variance explained hy the first Iwo pdncipal cc,mponenLs, achievcd in Uie arn~ysis above with 35 variables was somcwhat poor, we chose to rcstrain u,e numbcr or active variables to u,c six following val'iables : J)WTIMI', WTIMPl, DH\JMPl, HUMPRl, l)HOMP!, HOMPRt.
the problem is gained thwugh this aoalysis. This suggests tllat prkcs alonc arc far from hcing able to
provide a è'Ompletc picture of the oil future~ market.
f 6 FIRS'l' lilGE!'cfVAtUESJ
,1.--- -- ···"·"-!··· -- ---- •· · •·" '"·I·- - ··----· ·• ·"·i··-··----· ···-!- J CO!lRÏHNAHi,S ON AXRS
i":fo}]
rn.JM 1 f:1(~:WA!JJE 1 ,,., (IJM.11.., 1 .j.,·----·---~·· . ,---·----+• _. .. ,.·---1- 1 VAfl.V1F!l.El l ;~ 1 a. iii'.!-16 4R .;~(\ 41.,.;~<1 .HS
'
1. 9F09 33.0"l l!Ll.l [Wfllil' .10 .31 3 .::s:11 9.22 1)0.5)'
wr:rt-!Pl .. \)6 - . 22 .06'
, 39~3 6 .. <;9 97 -lJ'
[:+-!tl«='l ... ~:[, ,76 .\)9'
. l.40"i' J.:)•I 99,4"! JR.l1JJ:iRl - .9J - . 03 ... l l"
,()3?.ô . ~;:.\ l()Q.(li,J t:n··}1J'l .. . .:,a .?8 .2a' l·K'.'M.PRl ,, .94 - .18 .l.4
•..
-·--·---- ~, ... ---·•--· ·;·--·-·-' ···--.,...
.... ----· ---··-·-~' ···----T
+ - ... " .. -.... -· ·-,., , ... -···- ··-... . ·---·---· " ... __ ----·. ,, •.... -.+
COORO!NATES, CONTRlllUT!ONS AND
cos
2 <)F ll',l)JVll)UALS ON AXES 1 Tl) 3 1+···---··-' ... ·--·-- ... ·" ,. ,, ... --··---... --·-·---··\·
1 INDI\/ 1 cr:cJR .. (!,:('-;'IR. o:s.~~.1, 1
;.---· .. ··'"' .
'
1 111"~1 1 :;; .l 2 +~··-·---·-- ... ---·----., 1 W\!\'! 1 1.4G - .29 , ... 10 L _1, .) .0 . fi~~ .O'.l ,(:<'.) 1 Mf'i.\'1" 1 ,T) .65 , .. )4 ,,l . 4 ,,1 .49 . )S '(1'.,\ 1 H'if!.1 1 .99 .Hi , 11 .6.o
_() 1 .88 .0:.1 .01 1 ,:ioaî .4J. .-1~; ,03 . 1 -S .ô 1 .20 . (:) ,(i(J1 Jli~'J .38 .:)6 ,1? .1
.
) .0 1 .23 .Il~ .OJ 1 Aüe.7 - .07 , .. 09 . 0:) ,0 .o .() 1 .01 .Ol .L'I..) 1 ~87 .61 · .70 .18 ,/ .1 .1 1 .'37 . 4~1 .o:i1 Œ.87 .7:l -· .Ofi .'.l.<:1 .3 .o .4 1 .t?.l .OO .Hi 1 n:.iO ,l.] .20 , l.4 .o -0 .1 1 .Ol .ù-1 .02 1 (F;.:l7 ,(:,!) - .7:?. ,Of, . 3 .;'; .o 1 , 31 . ):", ,($) 1 1 ,Th8t\ l.f,() -.70 .:·,g 1.3 -4 1.0 1 .(,{\ .lS .10 1 1 Fl:-:i~fl 1. 36 - .10 ·, 18 ) .. ] .o .) 1 .82 .Oô .02 1
1 Ni\8B 1.62 ·-.i:S . ~>9 l. G ,1 l..] 1 .83 .o;i .11 1
1 flf'!)-8 1.81 .~il. ,2] l. 9 ,2 .i 1 .86 ,()"'/ .01 1 1 1-fif\S .'ll ,$7 61 .:: :7 1.2 1 -~12 ·p . lCl 1
1 J1J88 l.17 ., • .?.1 4'
.,
.o ,(: 1 .SO .03 . 1.1 1 1 ,n..98 1.60 ". f,l . )fi l ' ~--, ,2 'fi 1 .8·1 .os _()5 1 1 AfJBB .l,.66 ,,._, •,c :\b 1. fi .].,
1 .91 0'' .Q,) 1 1 SF...'!8 ,l.87 ... )7 . .'!1 J, l .1 .4 1 .71) .0) .ô:.l 11 m<B 1.19 m ,'2,"l .17 J. l . ] .1 1 .96 .01 .Ôl 1
1 N:&\ ;:.1.e .51 :i~. :i.8 -i= .4 1 .90 . o:, .Q~ 1 1 ŒBB L 91 1. 39 . ~\'J :J,:;; i. 7 1. 0 1 .61 , 32 _()6 1
1 ,JA89 1.2.6 1.2.0 .89 .9 1, '.! 2 '[, 1 .:17 .34 .H) 1
1 FEM ."li) .61 ., . .;;i .3 .3 .6 1 . :-J4 .24 , 12 1 1 Mll.89 l.~:·1 .27 . Hl 1 1.0 .l .1 1 . ")l . 03 .Ql 1 1 AP89 .12 1,40 31 ,0 !. 7 .4 1 .01 . 91 .os 1 1 l-!'i8~1 •.92 1.02 "2-~,ri , 0
,.
-9 ::!Ü.8 1 .}Û 1·· ,Tl 1 1 .JlJ89 . )l - .8J . \l .1 .(;: .0 1 .ü,) .30 , 01 1 1 ~îlJl9 .:.J9 .53.
"
.l .;l .5 1 . l l .20 l"' 1 1 AIJ89 .61 "'.7'S . 74 .;·; .s 1. "/ 1 .19 ,,. .. , . ~~~; 1 1 $;9 .86 .2) . 61 • 4 .o 1.2 1 .6) .OS ';-i:i 1 1 CC89 1.z~,
1.20 . ,(-0,
.o
].:;.!.
) 1 . ~)'.! . 76.o,
1 1 t-t)S9 1 .03 - . li:! , 69 .0 ,(1 l ,. _, 1 .OO .1)4 .60 1 1 ti:;fN 1 .35 .()4 ,BB .l.o
2..4 1 .10 ,()() .tf, 1 1 .JA90 1 ··1.47 1. 99 1.27 :,l 3 .4 ~>-0 1 .20 • )(i .1~1 1 1 FE90 1 - .'1;': - . 69 ·· l. ·1:1 .3.
'
9.2 ; .08 .08 .4"1 1 1 l-lh90 1 - .06 -1.58 - . S2.o
;::.].,
1 _ô(J 8') .ü\J 1 1 J\P90 1 .:.:2 -1.08 -·.02 .o ).ô .o 1 .ô4 .85 .ô(J 1 1 K:>.'8~) 1 .13 .28 .. ,89 .0 -1 ;i -~ 1 .02 .()il . 78 1 1 JU'.}û 1 .JO - •• ,.! ,·r -,83 .1 ,3 ;;:. i 1 .06 .JI .48 11 ,.'Jl.,\:lô 1 1.0~, .ù7
.,o
. 7.o
_,, 1 .43 .OO .OG 1 1 N.J'JO 1 .18 3.~I, 1. 81 _() 9.0 10.2 1 .OO ,57 . 18 11 Ec.'90 1 -tl.62 5,90 - . 57 12. ·1 30.). 1.0 1 .38 , 6,1 .01 1 1 0:"90 1 -~j .9.2 1.$2 . ~i3 20.9 i,9
.,
1 .90 .09 .01 1 1 N:190 1 -5. "/6 -J.00.
:~·;
19. 7 )..
~ 1.0 1 ,a, .11 .O.l 1 1 r.E-YO 1 •4.Q~ -J .86 ", 0"7 9.8 12.s .o 1 .,:;i .'8 ,1)() 11 -~91 1 "l. î,; -) .69 l.4l 1. 8 11.fl f;.2 1 .1~· .70 , 10 1
1 rt:,~n 1 -1. 70 .,,-~ .),6 -,99 l. 7 <i\,Jl
-,.o
1 . )1 , I,S .11 11 Wl.9] 1 ... 10 ,,1 _-14 .23 .û 2.,S ,?, 1 .OO
.n
.01 1 1 AI~l 1 -.43 ,2!~ -) 72 .! ,1 9.2 1 .03 .01 . ~).l 1 1 t.f\.'S:11 1 ., .. ci• ,,,.,, - .08 ··-1.J .2 -0 .1 .24 .OO .Ol 11 ,Jl.J9l 1 , .. tiJ :)o , .. J.fj .2 -2 .1 ,3() .1.9 .02 1
1 .. ,vn 1 .16 - .42 ,94 .o ,?. ?..l\ 1 .01 _(li) . 46 1
1 Al.191 1 . 73 1 ~:,~)1 ,1.07 1 o:~H -.4.:=i 1 N]91 1 ·-1.Sù 1 Œ:91 1 ·-L01 •·---··--.. -· ... -·. AXE 2 .S + .o J::f!Jl•fPRl HOMPHJ W'l'IMPl ". 9 ~---"•''90 OC90 .'.i(l . lJ -44 .J?. ·-.:r:i 1. ~:o ,21 . (14 --J .30 ,. , 1;,0 1 . ,) .! 1 . :°)t; .26 .02 1 1 ."l .0 1 ,80 .l'.\ . 01 1 1 . l _\) 4.:, 1 .1~: .03 . 73 1 1 1.:-, .0 .0 1 .'J\) .û'.1 . (X.l 1 1 ,f; J.. :~ .<l 1 .·:m _ l~ü .05 1 ... ·-··· ---·-·; DHOHP1 Dl-li.1MI'1 .. j .0 AXE 1 AU90 ,JA90 A.PB\! DE(HJ l'f(89 (X:89,JL87MYtlBJA8~lN089
S!"~9"J.Al)91 f'!~89JU8Î' AP8"J JL90AP(l8
0
,!---"-;,,·
K:: ~91AP91NOfl7oçH7M\:87MAB9AfJB8N09ù Ff~9 l
,Tll(l lMY91NOf! 9r.fY9 0,JU8 !3MA87 ~:;~:8 80C8 8
FE90AUH7 AUfl9MAR8,1A881:-i'.·:88
ùE91W~90AP90JUB9~TU90
MA91
,._l _ _ _ _ _ - . ..e;;J?O
··6 -4 --2 0
IV. l•ACTORIAL DISCRIMINANT ANALYSIS
With the factorial discriminant analysis, our aim was to find out which or the physical dctenninants of the oil market and of
the
prices of futures conlracts in one montheau
btJSI foœcast the sign ofthe
crudeoil
pricc
change in the following momh. Inothcr
words,the
problem is to seek those onc-month laggedvariables thm
can
bcsl discriminate bctwecnthe
two grnupsof
observations dcfinedas
DW'J1MPP " { Observations I DWTIMP?: 0 )
DWTIMPN = ( Observations / DW11MP < 0 )
Due to a practical problem of insuflicient memory, we wcrc constrained to the use of a maximum of
nine variables in a factorial discriminant analysis. (Jiven Ulis coustra.int, we chose to run three discriminant
analyses with three difforcnl small
groups of
variables to gct as much insighlinto the
prol>lcmas
possible under thecircwnstances.
IV-1. GROUP OF l'RllSSURE RA.TIOS
Die first group contains Ille eight pressure ratios wc have tlcfined. The power of discriminati()n as mcasured
by the percentage of observalions well cla,sifietl rcachcs 60%, This suggcsts Ùlat physical factors by
ùiemselvcs, as expe,:te(I, can harùly be suflkicnt for bnilding a pcrfonning forecasting tooL
DESCRWl'ION t"J'ri.TI'~; 27
"
~I'lMJ','P l:WfIMi~ ---···-·-· ... ----··-·· ... "· MEAN 2S .191 ;:';i. Î139 CB1'/CJN1 .:'Ul'.tlt:V {.923) l l..7,38) MAXl ;t;_'JS/0 2(1.;Uo M1NI ;o. "î~,ù 23 ,JJ() Mr.A"'l (,4 .ti).3 63, 774 CSll'/CIMl sm. rF.V ( 10 .4)4) (6.û82) MAXI 94. 2(i() 'rl,220 ~l.îNî 51. )80 51 . .380 1-ŒAN 2.638 J."l'..l:l. ClM/f>fl\ll S[Jf,1'.";(V ( .S7.0J 1- •1JB) MAXI J. '!(J.) 4_(140 Mn:JI [.flJO 1, 840 ME1\N 2~,.45() 2f,. 6::;o HUS/111.IPl sr,r.L".ii.V (1. 73l\) (L)Ol) MAXI 29.fl.80 29.0lQ Mil~I 2;;.B'/ô 22. 9)0"""'
32.'.·l~{J :t~ ,411 Gf!B/GA:i'l t:;fJf .[@/ (2, 98(>) Jl.67)) MIIXJ 39. 7~><.) 16.070 MINI 23. 400 :W.~50 Mi:")l,N 40. 387 ,)0.164 H(JS'/1-DPl tnl'.Œ:V r,:t.043) {4.267) MAX! Sù.270 4f1.à?O MIN.1 )4.0l.Û 3:1.)0Q,_
.;·1.0:u 1r1.91J;?. [<E:1/Rl:-::Pl SDI'.ŒV (3.080) (~i.001)wm 54. J.11.0 fJ. 760 MINJ 4(l.SS0 39. 360
"'"N
34 .204 3:1.476 ~!rŒI:'l sor.œ1 (2.)17) (~.116) W\Xl :'l8."nO ;li).900 Ml.NI' 1!). 95(1 7.·1. 850 CLASSIFICATION CWI'IMPI' r.WI:'lMl-'N C;l\.'1.JPS Œ' CfUGlN,,,,,,---·---' --·---·-'" r:wrrnr,r, 1s 12 [WI'nm,.I .1 l 20 (~XJD (iFCllPS C:F O!UùIN--·---· ,.·,·e---·(60 . .;!4} ()9.&6) c100.00}
:r:ŒNI1.F'ICA'l'll'H (.F MISGl"AS.':llFJ~1) lNDIVIr:tNB
(~:-t.JP o,ll'I;MPP
'
A.r~a Mî\89 AP~)9 F:F}J9 CC89 ,JL.()0
r;:i~J.lf' [W]'lM!~
NJ87 SEB7
msa
fl:";AA C(.'.(l.g ,JU!tlWI.RJJIB.r.IB NU,t; . Nl'\Ml:: 28 r..·:::rr/cINl 2f< ŒPI'/C!Ml 30 CIM/PlMl '.n HtJS/HTJPl 3''1 C'>tl.S/GAP1 '.l3 H:B/HCP1 :14 Fi.F,'.E/REP1 :)~:i 1'.'ES/lŒ.Pl O.l'l""~'mNI' R2 -[.() -, ŒRRE.l.A'l'!(:N.'7: VAI-HAf'.'.tf.::; . WI'TI! F'.L,D ('Ih!'.'1::i:J:JOl.d:.26), .o7;~ . ()~\~) -.O!J\i · , 07:.~ ., .,n~, .029 ..• '1.2) .177 . O!ll~;o F .'.M<l:14 Tl .. ... o:.:~.10 AP\:ll ,Jl.8'.i Al.189
MY9l ,TIBl ;.:IT~i:l. f\E:"Jù ,JA9l ('f:-9) l"<B:r!tl:::t;lcN C'Cfl'f'JCltN.\~: . o:."'91 S'Ol' Œvr ~n1J11:m· f)l!l;:RIMTNh.NT , F\.tK.'rICN
ci::1:rrr(::JJ:Nl}> (rl'J:JJUl::? ()f RJJ;[{F:S.'SlCN 'I'iPE)
-· .228661 ... ).ü!)'.':,04 .16'.lJ .6644 .0337S8 .GJt)Ol9 .0351 . <)~i!jl,l ... ~38296 ... 4.;~:-;;1:u:i .4820 , 9238 .9tl<l.S38 .467HiO .llll5L 1.04~4 -.827768 , "3'."(!790 .3394 1.1~,74 . l)8l 7~·,~.i .û:)1:\7')4 . ()IJ~);~ . 7879
- . ). l."lltlti - ,O~,S91t, .o~.04 1.1m1s
.l.9HIJ7'.; .09-1179 .co:n 1.~!881 }.Jif<):;1r..1 1. 61.6):)1 4.i:'.:_M:.i . 3-188 '(;4}'.11 Pf..\'.UA
..
.81783 .11,96:119 rri::w.•
.tn78'.l , PFCT:1\EILITY .'.;,('!'}(~ .G:;01:, .360] . 2s19;1 .;i.::;;p . 4346 .2726 .;;()}"/ , 7:187 -1.6 -1..1· -·l.2 -,.L,(J ··.8 -.6 -.4 .2 ··.t -~ .'.l ,5 .7 .9 .1.1 -L'J -1.5 -·1.3 ,1,1 -.9 -.7 .S ··,.1 ·,l .0 .2 .4 .6 .8 1.l'l 1.2 < ---.----1·---1·---1--·-+--··1"- · ··+·· ·' - +•· " ,·+··· - -·1 ·• ··-<·--···! - --·1·---·+·-··-1·- ·--.;.- ..• 4-, ,. ·,.j.,, ~·, I·- - - I··· - -t·---·t·---1-·- .. ·-+--· -1-- - .. -,. .. , ,., +•· "··.1- - -,1---·; DWI1MPP A!.t=JO ,TI.,9 L."<l~l 10"_"":.) 1CCfl9MY9 ltf:89 APll9 JL9('0'.."90APMtiE89Af'9l.JA.89 A!..191 ~)0 ,JA.90At1M ("J.;:8? :rtfJ7AP8'/ [;fr!ij M'{[fl ~V\67 .JU!J?+··-".j. a"~+-·•-,I·- ··-,,1 - - -t- --·+----+ ---·1· ---1"----l-· , ".- ··, ,.-..,- ,--1. - - - ;.---·t---1----.j. ---+- -"+ - , ""°" , "·' ---.i.---.1· - - - ; · - --·r---+···- .. ·+" , .... ···
--1-(J O O O o O O O O 0 Q ~ J 1 2 2 3 0 0 0 0 0
l>\VflMPN
~ l
[1:1°!0 ,lU9 ~!-W)O
Al':JO .:1Ufl.81·t)f!9.,JU9ÜTI1W F1:fl0L"f..'9.1 ,TI,139.JA.91
FE8l\M'i90[E87tl:'B9,JAf!!:ll:,t'.J<!"/lc[f) .l.M'i69MY8B..SE:88CC38AU89 S'f.:R7,/US$'1 AU87 1----.j.~-' ... " , .. .j.·" • "+··--.1 -- -·1·---··---·t ---1-- -·-1-- --·;--·• ·,.j. " , ... , .. ····---.;.- -- •. ---;----+---1--·- -1-·-·•-·.j.·,. ., --1.-- - ~-- -~ ---;----;----1-- ... , .• ,,.,,
0 0 O O 0 0 0 2 J 2 ~ 0 •· 0 0 0 0
IV-2. GROU!' OF l'RICES
The second group consists of the lagged prices of Um U1rce oil contracts (WTIMPl, HUMl'lH, and 1-IOMPRI) and U1c lagged variations in the priccs of the oil products contracts (f>HUMPI, and DHOMPl).
·nie
power of discrimination is around 64%. The rcsulls of the analysis show that Ute variables considered can be ordcred by dccreasing power of forecasting the sign of the crude oil price change: DI-IOMPI, DHIJMl'l, W'llMPI, I-IOMPI, HUMl'I. Oue ought to note the importaut role played by U.,c changes in the prices of the refincd producLs futures contracts. The ovcrall power of discrimination of 64% is only slightly bene, t11an the one obtained with only supply pressure ratios, suggesting that p.riccs by thcmselves cannot make up a pcrfonning forecasting tool eithcl', the oil futures market uot bcing a purely speculalivc market but one that is !inkcd to a physical markcLl
uN~:Aï11. n,sc11.,M1N~!'] ANALvsis oF a11.ouP!!J?.wT1Mri:-AND nwT1MrNIIŒSCIUl"l'ION
:;:ï '..\l
111.•rrMf'l"' twnr--o:-N
MEMJ~ 1.9:",.:.i()() :r~o'J .031 Wl'\'.Hfl.1 f;J;f,!'.IT:Y (3S.31S) !41j .?B;-;)
Mi>Xl ?d6.880 J~,:=i. IEO
MIN! 138.'.J61) 144 .H"}(J ~ rJ.:no -10, "/13 aRJ!~l t';;./l' .Vf::V l\~1,1.f,4]) (~,4. 260) MAXI 356. 12.0 i~:J .640 MINI 64.6~-0 -I~W-'f'lO Mff.V ~.tl:L .68\"J ~,89.319
tf, .. MHU ~;'.l'.Jl'.[.el {i.16.27$) ( 108. 974) HN<l 9S4. SSO 9Hl.()Œl MlNl 4~.i'.L Jj(\ 41!.i.810 Ml:N-J :N. (:01 ·· 17 .Q79 r:11:t1Pl SIJl', ffiV l (,) .J'JJ.) (40.864) MA.'.':I ?t.~1. ::;.i.o 53 .840 MIN1 ··l!CJ. "}::/J -lOJ. 930
ME"l ... "I fif-il -1I~, ~/)~~.986 !·K11PR1 S:'Jl'.L:if)} O(r'î.œi,:J) i L:4. 074 l l.ffiY.I 906. )9(1 \l~·l,.~~~ MINI 408, "f"Î() 11}.dCO .. ····-··---· CLASS.IFICATION GJ-î.JJFS Ci!-' 1\FFFJ.:'.'l},'l'JCN Dri'TIHPP t.W'l.'.tNrn cn.::uI:s œ e1uGil'k· ··---··---···· tUI.'XMPf' 11.1 9 [:wtlMFN 12 1~1 GC(ID BAD T 1, 034 . )ü~, 1 .. !~4·2 . 071 .:i.~,l .803 GK.Uf.'S CF crur.::!.!1,-,·~-·· ---·· '"'", .,,,,·---·-···"""' t:W'l'').l,l:PP Hl. OO (66.67) l,\l.(X) uil.:mi 'K!i'AL 37.00 {ÈiJ.7!1) '.,1.00 ('.n.33) 12,(,}0 (38.7l) 27, OO {:lQD.00) 31.00 (1(10.00) 21.CKl sa.OO (36.21} {100.00) ,rn •• - ••--- -- - • ~· "'"" M •• -• - - - " ' " ' ' " ' ~ 0 •,
Iœ".n';Œ'ICl'\.TlCN CI'' Mn;i('.U:1-~î.PJll"J lNDNiùJn,I...S GIUIP i;:wrIMPP ;
W\.87 MY'87 ,Jl.$7 OCJl7 APS8 Wl.89 !Ji'l,!9 APY1 cdl GK1.JI1 LWt'.\'W-N :
n:::i!7 M':{88 b'f-:.a.1.1 C.C8S FE89 HY89 ,HJ:19 N:"J.39 MY90 ~lU90 Jl~l tv..11
[im,cRIMINANT UNEÂR f'UNCTI~8) VAfU!IEJ..feS , CŒR!!."l.A'l'l(l,l:,~ DISŒIMm\Nl' . f!J:l;K!i~:.STCN ~;t7'1,' T
'vMI.ABI.E.5 !-1,t,t;'f.l".Œ o:Iir1:"ICll:'NlS [EVT S'Il!ŒNT NU!; f'Wlë , Wro-l F.L.D Ctfl·FJGlm-.'lJi ( ~JJI..."IS CF RS:.:t,t,S...:,J(:N 'l':(f'E)
l,. ùl'/18 Ff'\!.:EA :c·
[!f!STOGRAMMiè~j
3 ... <!-8 ··2.'.i -1.9 ,]..'.', -LO .. .(, .? .·l ,7 l.:i J.ô ;.!.O 2.:::, ;).9
-3.0 --2.6 --~~-1 -l.7 .1,;1, -.!\ ··,4 . .l -~·) .9 1..4 ].8 2.3 2.7 :i.6
"f"•••• ,. .... , ...•.•. _ - - •. - - - "!"- ---~ ••.. -•···. . . -~ ---1·--- •·- ,. "< · · ···! ·- •· -·t - --·•·--···'!···-··+ ···· , ~ , .... •·I· · · · !··· - - ; ---1·-····+·-·· ··+·· , "·+ • ,. "+·· - ·· < -·- •• +·---·t
UW't:1.MPP
.Jl/q
Cî~:I -S'fJl\) NJ911:E)1
MA89Ml91APB8 ,)UR"i'MY'YhJI..90:t.."~90 MY}O
MA87 (;(,;FJîWH"i',TI.il 70:~.) 1AP87 At'B9i\\.!3!;t.l.";il&'..'.C80.JAû9 [:e.'l8 ,J~lô
j·---·;----·;-·· ""'-! • ··-· \.• •• -i - - -.. 1·- •• -+ • - '· +' ., ., !-- - -,!,•• --1·--""t -· ••"f"·"·'"'".j.' '•• 1- ···-,! - - -;----1"-- ••1"• ··"+' • .,,j.' "" <· - ••• <·-·· -·1-- -·•- -··1-••••1"• •• - .j. •" ,·.j. 0 0 0 0 0 O O O 1 Q ~ 2 3 4 l J O 2 0 ô C O O O DWTIMPN fJ.')":I] l.f'.91 AfJ8\f ,JlJ:)1 ft";Bf~Jl)(l;l,JL8~1FF.il~\.JU;)0 FE9 i AP90JAR1l,.JJ..8(12f(tl9.X.'8(~)0!:EM [."f:."')0JA91l"f·1:lQ ~ ) I J N.:fJO $1::"3"'ft,1/l.9.lW'l.88AIJ87tE87~!1.lS('N.;'.'f!.ï'l'Ul\~Mi8S
+-··'· --+ .. , ., "+·- - -~·- ··- ;----·1 •' ., .. ,_ .•• , ... - - · · - · · - - - ; · - - - .. - - · ... ,. "1•"" -··.j •• - -,1---1---·-+·-·- ... ", ... ·-·-.;.--··-·t---·;--- --t -···· ··+ · .. ··+ '., "• +·-·- -~ - -··-;·---1--··-·+
o
o
1 2 4a
s
2o o o o o o a o o o o
0
IV-3. GROUP OF PRICES & PHYSICAL VARIABLES
0
Finally, we considcrcd a Ulird group consisling of Ule pricc variables considered abovc (W!IMl'l, HUMPR l.
HOMPRI, DHUMP 1. DHOMPI) and physical variables relate<! only to crudc and rcfinery operations
(CRIMPOl, CRSTOCI. CINl'UTl, PEOl'ERl). With Ulis mixe() group of variables. the power of
discrimination jumps to morn than 72%. If wc ordcr again Ule variables according to their dccrcasing fot"eca.sting power, we obtain in this case U1c following scqucnce: DHOMPl. DHUMl'l, W'ITMPl. PEOPERI,
HOMPRl, CINPUTI, HOMl'RI, CRJMPOI, CRSTOCI. This shows Ulc central mie playcd by the variables
dcs,nbing refincry opemtions.
[iÏNEAIR u1sc1u MINA.NT· AN A 1: vs 1s _OF.GR.ours· DWTI MPPAN_D.DWTiM-P~
1n;sCll.lPTION 1:rrr.:t::rrvFs 27 )l rwI'IMPP tW'l'.tMFN 'r 1:-it:e ,. ,., "-··· -·-·---- --- ---·----·· •.... , .. '·--- -··-·----·--- -~~ l\~;,.~,00 207.031 l .()JIJ. .30S wrIMPl si:11'.l'."!.V (J!L:n5) {41.282) W,l(I J)('i.88() 351 .~ .180 MlNI 13B. 360 144 .8'}0 MOY 17 .3.10 - IO.'ïl:J 1. 812 .07'), t:-tll.J'MPl
_,
::IJT'.ŒY (59.l:A3) (;;.4.;~(,Ô) J~i6.1;!0 1~17 .640 MIN! -·64. ffj) -159.770"'""
581. 689 ~·89.]19 .2Sl .80) IWf'fG $)'1'.ŒV (118.2781 1108.974)"""'
95,:1. s~.o 918.GOO M!Nl 4~/l. '.l::!0 146,810 MFNl 24 ,{.,Ol -l'J .079 2,9f.,Q .L\Oil [:#'...'MPl sur.[&!" (63. 371) (4().f\64) l<'l(I 226 .940 :",).~10 Mil-H -80. 750 ··103.9)0_,
5:",1-41;1 ; 573.986 , 72) .,ru HCMF\U s.:1r.ŒV ( 107 .()~/l) (124.074) Mii.XI 906. 390 9$i;.S8ô MINI 408. 970 4J.i! .i?.(X)cF •. ~nt:cl. sr..rr.u::.1 (*.., ... -..·) (''**"'~ ... ) t-'AXI :!!l4S17. 0l)iH7H99().000 lfl'Nl :t~'.1617. OOOJ~;JC\~;1 _ 000 MF/IN ClN!Ul'l ::lH'. [!EV MAX! ~lllil L\4(1;:,,. ')'fü) 13:lfG, )?(,1 (62).]()9) ('.-\6ù.323) 1.'116] ,0()(.) 1426),(.\),') 1191)4,(X)() J'.?."]'.!1,0XJ M!:PN 1VS.!)r15 $1;,.\4:) P.Ü)l1JU CT"Jl',Df;V (4.201) (2.;rJ~} t®:1 9."J.;?30 9l .. ;.',HO MlNI ï(,,. 930 81, )JO CLASSIFICATION (.~~JI~.> CF (RIGIN"·-·-Gr{XJPE] c~;- Of{!Gl.N··. r.1;,n·1H!-'P 1'..w'l'lMPP 20 7 l:.:w'l'IMFN 9 ;;2 '.:!O.(~) (71. 07) J,!..0::1 ("70.97) I!n.D 7 ,!:X:I (:Yi.:J,1) 9. (() ('.:l9 .0)) , 7(i9 'TUrAI, ~:'! .Ci.\ (.\.(!(). i)'.)) )1,()0 i:I.GX),CX)) '\'l'..i!'hL 4J. O(l (12.41) 16.00 (27 . .59) 58.00 (l.00.00)
U:.f:NTIFICATf(~ (:f' MJB::::l'.A.:{:-:J.PTJ:"1) INDIVJ.U.""l,[-;
CuRCJJP r.wrJMPP :
(X.:87 Mi\89 A!·'S9 l~::1.)9 ,JI...9() ,J!,S.<l CCJ]
(~UP CWl'm't.l
»'.i!'t ::E:38 \.'.r.:)-)8 ,TI..B9 Al189 t'J.)."l\'.:I MY90 ,.'Jl.11).l r,.r..·1:t1
