Aleast:- Foy abmne — we coniinute Tb, Maintate apiny, somne congusnpiten, pla - THe: 4% "Log" i o —™ Coe Prt Pave + ee Ve : Ga Ofats G Ww she presence of “lige — a" ite Yous Yy oye a er ee to 10" Hoot aoe tro | oad fel na W test wen fos | om] 29. ald M12 SA 35 ma a. a Th a : : abs oe : RoW OZ 10 705 ri ‘ : - Yee + Rist - Oxder facet: ‘aka, Markov Ast ‘order Putoconelation — wuliple Linco Regreoaon Model aw oka i Ye = pit Ba lat + Ba tor bo it Pa Xie t Me ‘whee Ay = Salta # Eg ==b cAR(1) © sPrevvts ave auto cw ridatedl* f Wecon aloo haves» We = JS, Mea + a Sect ve ovder It = SMe + S'Una as ‘e+ —> f. condatnn togfiaent? & (ay), 7 + GRLK) F Sed, Wee, No oule-conelation F Sri, tiger degree of outd-condlation (ove ov conden) yo SoA, Ve ailig- covrelatton- — 1. ‘ora ‘yon omdy Go followed by another ve em % "& 0 -ve envy fo filo wed, by arnother eye ent a Le dovdatnh sO NC Value to fotugd by -ve value & pose value - 49,, followed by $ye. valueee Consequene of auto-covelatin « = i ts > Soreness & 49 over -cshirhated » > Fotimators are, 00 longer QUE, they ave LE —> y= Xpru W% te assed that e(« ts Wes) | ou Ff 920° i 0.4 oo: — auiocancdation at logs can, we ‘he Eiiaae Y, = E(u, Us), where = 0, ee f+ y= E (Ue) = find Y, (at sto) = 0 > fiito-cniclation = ‘ % = E (Ut, Urs) =~ Me “{vantuy) Vor(ug) Alsousnneg ts & Y av suena} 0 cet — Aiuto Xegeeave process ar (P) where Faure g lag Ube gies by AR() —_ Hoving atalily a Mp (9) Nh = + Obst Batee + gee ten —=> (peniined, ARMA (9.4) N= Guys + Uy Ua t- + OpUrpt & + BGat Ghégtt hey Y= Pot Bi Xe Ur where sup filows Us Purse E, ARIA) procero Jor cowect eoitmation, ff] <1 “for E(&) -0 and, €(E, Go) = [% Y 2-0 o f Sto F you define ya > b Ve then, 3 "Ve Putt in > wfltfl):a& - we (c)4 ov Up ILY eyFapanaion F a 2 FOF OF. : e vw Thacfors Wy = (+f + (tL) EYL Pe ) & eat fat+ Parr far + Wee et fee fréva + Seve t Squavieg 00 oth aides & ‘akong. expechon, E(uy = € (et Sev a fers Pear J © Elwp = vol) = oe ~P E(w = \

> ~~ " » y we Awe y Teshng for Autocowclation : 0) Quito- Watson d- Test > ‘ nth Higtage j ei Pyare gf de z ( &, - Aya) = Rabo of difference 6 ar ey terms ty R5S- ae tI e . — We can, cally dtain the vesiduale on euler” SUCCESSIVE we we egress the snode} > Resuangons: | 1: the vegueeston ended cnust’ Awilude the ~Antercept ei 2 the seid ependentt Variables (ic) ‘cnust be - von-otochastle [yandom) 3+ Als taust follow AR[ 1) proceso (% ur= Suite) See 4+ Applicable only Sov AR(1) series Not for higher order veges. scherne: 2 Nok applicable YF, Ube Ov sot Corset distnbuted 3» Regression ceudel duce. not, Wndlude the fogged ° dependent variable ag one of the 4eqresso"- , Ex Pot Bi%e* Pea Tr = yy > Wye No fret order senal conelatron, (0%) auto covrelatron Wie oute-wovdlaiton te present i i ob “Opplicable5 we ave wovleng on lavge Sosnple, the aiecesave. R55 cam be consrdeved. 0 be equal ey 0 2 Me Ba a hah 2 Bu, a) a HR’ i © wo 9, 2? i Pd- aeW-aeintn 5 tat tea = a E yt 2 ay bles eS * 7 tH ~ |a- eU-H Doe, [= for (Me Ue) ade meer st a Vav (ut) = & derarsinator by 9 Je (ul) =e d6(o4) — ¥ feo _ dee @ No eae (es ore‘ conc f } Sy lap da Wt _ pve outa covclatton io presen, 7 ‘ feo =p Are * mene autoconelatton © i. steps 3 to wnduct Dw dedaticht dah shoo’: : Eshenate the seqregon crodel wang ols & ddan tl Glalate the: d- statistics & “the table “wilh DW d- stables é dis dh ¥ on ye %) a) constitu, and ands thy ov. the cattical ‘valueo- ” du: upper Wonk 5 dus Yaer me ww) Gooclude? — : ; —> f dg, accel the’ oui hy pothesis ;(raw back + Did ee ) Bao The value oy Soltow under Atvonelvsve zone naa 7 peu Godfrey Tesh > aka. Geneval Test for autvcowdatton > ‘This © 4 Langrangian Mutliplier Test thet. ovev comes the drawback of, D-W, d- Test § ARI1) process: — Yee Bot BX + Pr Mat + Uy where, Uys Putt Vea thu, i Cy 45) eam 2010 & const: vovance ve — Sheps: yen 4 4) Run the send wag OLS & “obkann ae a §) Regios fh, = cei + oe Xy + re his hi tyr + hpagtts ‘and obtain 1k ‘call tb ausdtiany : &* 0 F ‘il No-f Obsewatuns tn this. sepde) = M-pr,, wheve ©: Sotal no-of Observatrans P: total co: of lag “berms ‘y) Obtain (Fp) Re which ” fotos Sg! FY ve xp] hee Xp to the cMeal "Value 7 — If (Op) exceeds the exited) value, wer veect Ho Duos +h test > in the proene § | 0) fee varnable as ne fj § ‘the nde pendent varia bles — Ye = Bit Pa an + Po Yar + Pa Ys +N where ("follows “ARI+) a] neg where, ns noe FF eboserrahons 7 DW devalue” » estimated tovorlance of the apes nye dependent variable: « > za tay then we week The Hy ( 2-teat for lage sample dala) 4,2 ‘P='0 Wy: [#0+ Gouching joy Ai : ™ 4 Generalised, Last Least Suave tabod : ed we he keow fot bas = Prt Bok Fry. WHOS ARLE) Mth = tute ae) We km, tela’ — Cava — [% enouel “Prioleen of auto wondaton Lm be cat +O» se I potod “4”, tt should ‘aloo “We ue for cae bate Wes i ' a ee = Prt py kta + M4 + thutnply : “oh the atdeo + i . ; kha = lee Peete + fy —=@ 40-@ eve Pg fll 0) teats) + + [We-Purs) > yf: ptt pht+ & —@ oe ® + D & © ove stoner, © Rae-@, wing Os A hte 1 ie “for avto-audatton fe Arareformaton.. 19, called ois Helbod- — (aser2 # ¥ ft cot Frown ‘but hos 4p beahrta- We woe,” First - Difference Method ” : * Of fet (teseely -duto- comdlahon) aubghiheg fy “40 @: (Ye- abs Bs (Xt-Ma) + _ (th Ma), : DY, = Pa OG + es here Oy = Aieh- Dyernee F hun Oy = pe Oke + Ce using OLS. Free fm ovlo- coal Applicable only for a bigh- deguee f auto-condation (1, Jest & But how to know F fxs? — if dR, then tigh ato conclafton qo’ prevent & Fivot— diffeence to appteabie- “4 CochYane ~ Ovcutt Procedune ; ‘) €atonate the soode. usteg OLS § “obtain t Run: the ve ") : LS ee a 2 uto wrdation —«) Now, follrw yt - pty pt XE t Oy & continue the G15. Method: ANB > Hutticoline arly + “> crptt assummption: No “Meeay” courclatn (exack comclatton ) oud _ be present: Y= Prt Bako t Paks +U Xa & Xy ave mot distinct. Thee ' | ts a peyfect covvdlation ectween tp & Xa; _ Muth cally neantty: 49 pregent- i ten sa we Ye Bt BN 5 PY Br (Pytaps ha tM oS We compot calculate & § ps * nde pendent ly- Gan be. only eatimatce “agather-, : ' fon eed FT Gh t Gyo fr 40. & & +0 olds, ther sul alticasty ‘9 preset : Tt eng ‘af aX, = Xz “esp ak wx — Ys Xpt & whee Er~ N (0 sz) (vee distr ath ¢n0an=0, rater ) are present - “Rink (x) = . +k explanatory, yaw ables ave centered g scaled fo uct then, £3f obswvatons of X xx. kek nati KY ker motrne € ld Ye (x. G- te) where, "Kg "19 the fh column of X the column vector ave “fnearly dependent: If there emis 2 xt of constants 6, 6, Syy---r6 that aie apt all 270- K & &Xy =O tee an y sas equate holds “cua dly” then Sank (x'x) < k, conequety fer)? docs nok Cdet: then snulticallicavty’ exists- ee *) Estierato ave sell LUE ‘ort jpu may not find fable to eshmate the exact esfirmator. -Y — HO 0 %5 opps tue _fov eosne subse! ¥ ‘a Yas Yoon, , then these will be a siay-linéar’ sepedeny ‘in ts = i Yq & X5 axe “non-linearly” cowdlated, , Ne prt Paka + Pala t V aad PY Bit taker Pot [hee » & When do we suspect Math chi 2 4) Gooharnt on the Hudel oy 40 thé | copulatory * tg. ‘wore Ely = i (y Mh) 8785 LEGS Aechnaly consusnpton, VT ars gy ca V ssneame of the faseil fos oo tegen te Yoes goveenimes Sf YA Abehl Hy Pek wn Naf nusober of vovtobles (5) 7 v0. obeenaons tr) ’ — eoealy io snedieal’ fiddo- wl) SF, srodel. te Weg the crves ~ Prue —> F i ia : 7 r > a> Tm unos product wie wf have” Ate wh, fers * — The ys > fe oP salalas koa te oe ei * osepere of utficoltacarity.: — Ff, *% Ms ove hocaly a then Wal 6 me ve gh srngulay sabi: —> We know, B= (xx) xy : » (ey ara en] “ty & in case of pafeck va, [xx] 20- Hence fey) gon —> Ve pit Poke + Pyto + Ve ae Sf §-t) then yank [ye Hani of Tf Jax then Nor(B) ~ Var (B,) <> 00 6a Tf the vamable ave Peypectly covdaton Then the variance of conmnaters, become vay lavqe---Mowkvey, OLS coMmators ufll atl be BLUE thie maker the precise estimatton very difficult. —> Ff vanance of catimotors to large then the confidence Interval may be sauch wider ve, chance of Mgecling “Hy 16 very leas = %y hts, we mai get OMe OF MOVE yaniables to be stahottcall lasigeaficant even Aeaugh “Re value seems In be tigh — > Therefore, 015 sheath and se ave wey senenive, even +0 a srnall change oecuny ei the data. eX ye ¢ Reteation _of tit cottneaviy — f Ras high & some of the t-stahighes volute ave swt ota daily sigeficart, “then there 1 a chan of snuthicoll neavtty. — > When |y¥x) -0 Ernqulax sain) : ‘ fis |vx\ tw declisnng, the degiee sith otneaity swacases seo, ie gee # obey he find ont pot wise — Ff 2 oF ove variables. ave Woearly ated ondatnn coefflaent- on ' ; Ky Y= Bit PaXot Po Xo HV oo Me ae off-- Voviehon Teylotan Factor (MOF) + Y= Bit Ba Xa + Pyty tY (a i. Z(me-%F (IR) | Co; ))= ee ee gues £(4-%) 1) « (r Ress (4 ee glee Su ( QP: to cot the R2of athe 3) Qs ti Obtained by vegrcosin: Ky on Ny yey vor ( 6) = Cae . ‘ % ganolorg.” wall) = EE Caloof S : , \ =p Variation’ Inte, 2ly-B) Ur » Therefore, Vor (6) c oe v (VF) VIFj oop ui / If geo me vary = L (towwot vale of wr) a Tf VIF <10, then snuthicofeearity, 1 nb Very’ Thavrefal g YH VIF 70. then some von we te degue ¥ et present » 2 Re VIF 0 L viheen ae wee fovea’ od 19. 5).4- Ws a0 oat8 4900 : % Solving tater problem eI ay —» H thee to a data deflaesey probitesn; ben nating, can ‘be done == at F you fave some a- prio’ Agformation , - "Ye Bit Bete t Poet : ising + Weatth Bf Py OND Py ae Se, donsusophton Hake, wet Bia no 107+ that wet Wir Hence, we con wrke, Y= Pit Pala + Opal + V' Ye Bit Pepe mow) v TP meena) : We cos wr Y= prt PaXtU whee X= ht Ka Hew we, \eoow, oe tat By ba py ; Nw. revtonincaly ne fe oot vey haifa)! _a _> Tf we combine fimne-seres data with cwas~ sechonal data, then we are compressing the vaniattonis and hence snulfootnnnily deerases > Ff aw woe vawables ave ‘horarly vdlated hnatly cond ated), then dropping one J the variables, decteases ynuth calhneanty. © Aloo wok at adjuoted R varab ; value. gore Hence ue. Tay A¥op. some, MMOD! ngEHoRE vaviable leading 4o “apedfication envy" ; “ dropping a Vawlable — dnent f * ee ao ‘eave Je Aven" ohgrificarly Sed adjusted RE Aheer we can = Yuihalt neaty. 49 decveased by tng Pnayal Corponest tral (tea) — Badge Requessis ” tof the dass of cohmators ave VWoeor but dot unbiased ( +e, chey on bad] Aber, south athovortly to present a 4 tre can be colved by hdge Regresstun ! ss la ¥ ecinal / Qualidanve orialds : Se ¥ 2 Darnmy vaiables con be coptured Aienply y asalgming a quawittat ve value: ' — definition of poor > snay be Dummy variables can be mode in The fme-eevies data as well. buat Duro variate 4” ‘ohd “duiwwriy variable 0% should lee properly aselanl bu: logore COVED & fifter COVED : ~ Vovally 4+ Yepresente’ the: presence of abtnbute Qs Vepresents the olpsence of atte bute = aka dvchotarrous variables | Thdependent ‘Vanabtes wage a (Ea. Expenence, Gender) pelts if the peroon qo female eae fo a) [> other whee avertra pe _ Noe: there obo a pooetbilily F dependent vorrable bang 0 dusnm Variable 1 N = ee —an aw mudd containe a QUOH Tre NOG del ts hKown 09 ‘Pegyeesion, Model” ative | Nawables Athen 49. Any (nova) | Mate & quashative- vanes, the (-ANcova) « —> Ff a veges then ths $00 =, If a wooded canloine only “gyal x9 knot) Aralyste. of Vanance, — Fa sande) contains bil “G0 Rye soodel 19 bnd@on’ 08 Aralyors ‘f (ovanance — Rewon, Woge Edvention . fxpenence Gender Howl Stas ; 0 5 g V(famaS 5° 0. fom) gO OE lll 0 ra 3 250 \0 3 O (eral) © (oanea)* Oa gt Ba iy Guaale) romney s 400 + Pee uct Misa IS 6 500 4 iphod gt ‘ i art Qn s Were 1,0 do sok hove a quacilive value. ‘They ave qualita 16 vet De LY fae pe Ca yf sated: | , 0, iF Oberutoe 10, af da weaned: © $f Saje Dovey Yovobles a — wage - be + 6o(qerder) + ‘pr (Educaltoa) +! u ' =. Pot 60D +. .Pr (Goveaton). + 8 ‘ - ” [we con also vwitke wage = Pot PuD+ Po du) + u) ; we defines = 2 (45 Hf ifesnale , , 0, Otherwise —> Aotapctcton of “by : GVERctiene education @ envy term thd Mae YF Serene .& tale gre different 6: dMfeence of sean wwoge & Jerrale avd rrale “(wn at) — sp. [4 ‘ale ‘ “or _ . tise =H 8705 female caring move <0 5 vedle eaveing’ moebe it Yowale 670s deole carsing more 0, otherwise Geo 5 Sesvale “ny tmore:, —7 We htpDdtEW ¥ Deo, We eee ¥ D1, We Pot Bit U Elw}o. ©) + Po E(w)os1) = pth —e0 frm & @, Py = CF €(ul)p-») Hence G Yo he hffevence’. rr beweeh De\ & D0 > Sn Prt 60+ pilEai) ru» oor if Devs w= Pat p(y) t¥ ye L we peak platen E Mae)» por pli). E(apo) © Be 6 pe lea @ a 028. €(.)\ Aona a Here ¢ 9 fe diffeence of eM yoge Toe Del (fesnale) ond:.D-o (ale), — @ : straight tne wtih intercept” Be! and: slope. “By @ + Straight, hoe with santercept "po" ond per ‘Br £ (al oer tiv) — E(ulpo edu) 67, Ge €(wlo- yee § = Vemee, “Go 49. known ao “wabercept bah variable” oince only Abe anteveept ' Crosgrg but aut the slope-+ Tuo Durrny Vowiatles + ; t of which # Oe duro —> Considering 2 explanatory vorrables OU! —> Ja toose two dumminy vaviobles ve Bond Ds given Wy ‘)Q-0, ‘pro ¥ the peso 1 Jr Grau A a) gir, teat the pon ie Jom ie a) Dads, yt Ye por %» fom Group ¢ — Regression cquattan + Ys pot padi tig dt Oru viene Varn (0, 27) “6 “oean=g ooh vatlance gf o* —> Ff yo ove sorcluding the antevctph >) > oe Ago for lende,, Wee el). dureeny voriables te ( Vf, Woe Case , Aheve ove 3 levelo (Group A: B,C) we Use 2 dusesny votables i BAD Ft oe Tf yu ave'including voth rotercept ond" dummy variables ae then we encoustey 2 vanable hap" where exe viight wpe a problem. of pexfech, mmuttalhncanly. > However, YF rterapt tam ts set wodlided, then we Can Woe “ dumsey,Vortables fp ‘0 levels — he prpbru of ey giohote E[vlpeo) * & t+ [Ge if 80 : E(y| oat) = pat Pr —> Jo, > & (Yn) ~ €(Yloee) fo = sneasune sngon salary of 0 von- groduate pr cyffevence i sorean calavy fo graduate & a’ 800 Heve, grove — > Now, consider two usreny variables, Vee Pot BO AP DFU : mf Vf poduate mf if vob godt Oy vier uiee ' 6, atherulse We ull hove 4 diferent cuses1) €(¥]Das0, Beet) = Pot fr “andicates ‘the overage valory of 0 fon gvciduate peveon W)€(¥] O12! Das 0) = Pot pr - andicates the average alary fa guoduate peison wt) €(y]Du0, Dio) = gues on definition , his does nol enst Contiadiclina. our w) E (Vout, Del) = Bet prt Pe uw derstarding Tio also comnot cust iG yee York of the snatne of exploratory vaviables “fall chow by “4° this © aMled as the “Dusneny variable Tiap’ —> Now, let uo dip the tatevcept ‘term ov : We PO PTY Hn emen ) € (Now, Dy <0) = fr + 0%9- salary of a graduate 1) € (Y]ow20,, Dazl) = Bat gr cology 9, 900- grodunte ay Rules When the explaratory voriables tead 40 ‘hn coutudll ~— exclugive cotegotes, then dee “m-1" dueneny . vartables- ANerntely tise Tn” duennay variables ‘out drop ‘the ‘wtewept term. # Slope _Dusovy Variable : i ioe —> slupe dueiny vonadble can ‘oe snduced, into the wedel, just by antevading wih the caplaratowy variable: : == he pt Pot BX + Py XD + vu is i - sigh ‘ gn Un oonseep = @ In Con Exp) ile ngt_equation lope _dureeny Vantoble) « Ys = By BiXi t+ BaD + Pak +U wheres’ Da. ef le y the 4 peoon loelongs, #0 Group f J} 9, otherwise : |hee, €(¥lo-0) © pret | > oo , E(wor=1) + Pet BX tt Bhi > (AYP be Boe E(eloro) & E(wia=1) nas oth veg slope: aot + Example: Wage equotton ’ + ing Wage = F (Experiences Educatroal fant) . dd ts F person obtained only priteary educatton >. ‘hind / ls. otherwise 4 as 2 Desi [o. 3 nal pe} if obaained. stam . Qs 4 5 obtonnel Ua, 0, where 2 0, vibevatoe - por bi + 050, + 4305 + det (Aéupped Dj" 40 cold top) oy cam dop any one of dhe” diieneny vortables, the dropped dvtemy yoriovle (Di) * called the * veference vawiatle” : Nore” AY Di Dy; Pe, Dy ‘oxlong 4o aly one agicony vale +2 1" educattonal attean anes ore Now, Ai Del, D,-0-Dg” " i =D Ve Pot Bx + 02B, “he = (porta) } Phi “prt ag” % the ratereept for The cave of wecandavy educatton Now, Dye Par O~ Oe - a => yy = (pot da) + BX: ” putty amtevcep Joy VG Cducahor Now, Dy 1 Dye O> Ds . , ~ (Pots) Pith: “Pot ay” sotercept for: PG . educat™- sates nhereept for pritrany educattov whl) the nsercept of the eqression randel Wedf- 46 Antercept for pray educatwn’ = po 4 Durarny variable tn Time-sévics data: — =— ~ . .— * 5 Th 49 dested Jo vemnve the Seasonal voriahon an data. ‘Thre Yo called * smeonally- adjusted “Firme- veneo data" —=> Example, quarterly Arne 66° data> Bul we conarder’ only DS duswny vartables fv 4 quarters Yes pit Rake + PrPe > Ba Poe + Po Pas +N Qy of Le WP oale Tel OF Lg Vw Ors, Dye hye 6, ofberwise “e 5 PN) 6, obetubse - Dhukenabe —> ov crampe, iy = Waaaig + 2geSTDy 4 Bat 62 Op — 12 Dw ; Bow gestanetcs > (ana7) [RON (4: 04) Lope) . Anterpuctation: “How sawch the ong. value of} ¥ 4 the guavier that Necelve a. dusesny vowiable 4 a differ fom thot of the Yenchemark value ° [Benchsnaye value + “Qs (po) J: What 12 the snean oale 4a the gd Quavter ? = lave # 945-37 = OT # Dynamic Evonometirc Model: - 5 TiN Now, state (ot one yume) We xe Bke + PaXat + Po Xt, des ome or move Wee Nags oF an explanatory frown a6 “Dighibuted. Log Model” per + Rx Xtek 1 Ue peccee + Pe Nee FUt —F the yeqresdion ancy variables then te —> Con cleo contain dent variable’ ” Rutoveg reser Con a ain depen anable's log 7 Som ag Yee et Poke Prter eM Mba tle J. Model” (Arr) > TF a regression daly contoing ‘og dependent vaviable 1-¢, 90 ‘ag aadependent vaviable, t 19. calléd * avlovegreesive Mrdde” a Uy =< Pt) atu% Tomgortioce of Log? — Corsumpon dodostt change swrrvediatél ; —> Production also takes hme to “change Eu 0+ flu) | boon oe . 7 : ke copftal 5 Ls Labyy —> Inohtuttonal - Reason: . ¥ Preco.to the soenu card of a veotawont fv dishes fill be Say * gileagt for come time, ever afley there 6 pice’ change ‘nde angredients used. fo make ‘Ihe! dish % Dushiuted Log. Mode = * aT eg “yy Sot Bok +: By Xe F/R R Mba ty Pe Neg te Be Med Ue —> fy will give the short tev Aenpact ai sum of pe will gue the long. tern Aenpact — pi cha eo An One period.» Pos Br > changes (of yy) An Nett peiod, a x Afi k perad, « g pee Pot Pit Ppt t Px = Bt gues tote) Yorpics ed \ + rt acy, + * PSot ie rasa tps ; — pie Pig BL uit gives he! adwal! change wm tat pati a ae Pooge pds # Eshmatton of Dictabubed: -Lag Node! = - . Ye= ocr Pol + By Xty + Paktato + Be Ken tle since Xp % Mon Stochaetic, all Me, Kiga-nss Krk ONE, Mon-stockastlc —> All (iaga) &_Tiohegen (1ega) ee, . Requete yy once on Xy, then of Ky) then.on Xr and’cd om 16 , Ye= At poke + Ve = CP Poke + By Ke FUE Ye = a+ Pore + Pits + Pala + 4 Stop whert ‘opposrte sing ceity >No option guide about tow geany “Lage fo be chaser (tag leah) 5) Grosumer a lo of “df” decrcasea the gfideny ) thgh chance OF petfect tovthcaltineaviy, ao aUccesstie lago haghly cord ated: i night beNN _i—iUCOUSR Gh yh yck Distibuted Lag Model Node * Papesed a (geometrically debned echemne {up Ve = <4 PoXe + B Ny FAs, + PrXv-p 7 Us _ _ Rssurnphons Lot 1) AI the po have same ston ‘) the p> decline Joometuically = Piz Poa! where t= 012... 1 & 0eAet % EEA” 19° Reown as Speed y adjustsoent | ¥ by asoutming Dede) (noo- aie: % we ules out po rom changing signe: #: by Geousning aati, gives by Fre Wie rr +0 the vecesth past (B, ps...) and teaser weight to distact past (Pro, Bp Pr) 16) We enoure Task dum of pd which roe {ong wus) asrpadk 15 dione EP » Ta) a) a Eos perp es mee PPro | = Pot Bolt pyat t ---+ + Bod | ~ Polis as dtso4 at) = (Ce) ° — Sine pe ave geosmet ically pS Pal» pa > pra? That's, > Pa? Ba > Pe ~~ Ys xe Pron” X4 + Ped Yu + pea? Xieg tee $y -_@ Cone\denng the above Infinite chotnbuted * ke Model... Then, Vas a +. Po Xt Pod’ Xb + Ped” Rts t+ > Mea, —@ Nultiply ooth sides by a, AY = Act pad teat Pod “hat PO tha Hd AV —@ Subba & fom O, ' Ve aM = « (a) + Bo “0G - Waa) + Bay vt) + pe (Ate Px) + (M-d a1) Ye Avbs =< UKA)+ pos, Sere + (ws duns) , 8 WATE, = long wo rout pliew 4 = + ; To the tog Wun, We O60UM | Anat We We Yt Was \ex foi te oo Ready Sai >, Ve ect Rory t pity + Prtha don? il, |: 1 Diobbuted ly Hoty to > Ye AM = (EA) # pode wee AM)” 2 Ruste vegvesve ode) = Hn > «lr a): + pou tata t" ve, ae "iy yore the adiual ivonable . & Yet 6 Abe ophal ded, Void, *) Pavinal Adjustment Mode) > Change to eG Yeo Me oy ait be cqual to a proportion of optimal chawge hes yfeny oe, =o yet )el fh ty) where A> adjust sent coffident &k oer Speed of rae vi | # Thedgh Then Mis vs 2 adjusbenert: to the ophnal levd ty ols * Tf d=o, then VE = Wat “The 6 90 adjustthent of Yt ye dose the At 4, foster fo the agus — Noe fortial Adjustsnent Mode Hlode) : ye Suppe VIG dared. level oF. governs for ‘a puri ard thio 7 on gales’ xy ‘Thect He’ srodel can be? = Pt Pak + Me 0 a From wil adjustment chodd, © Yet Vege = ayn). —P Vy Ma bd, bf: ¥) +t TRS (ayers a+ ve Font. ‘e Hane deen) —@ co Y > Nat AB > Ap, - A Ya +4 )

Pw ean No veviston 40 the cvpectaton is © ye enade- : / ¥. pert Pe —o@ , b =e We sna not have the Gorsuenptton Expected Income pial see tote ~ Damar formoney Expected. vate of ntereot + yz Xia agents veri thor expectation upward. .~ # fe uboo a Pao ner ad he ‘ = Tx, i cy 1h v sas @' a ‘a ond, : = prt opal ty ( (-¥) yi) + ue = Prt palit * Polk 7) ty + Me * Now, FO % cone fov “4 ib snust te comect ov i uel. | Yi B+ prttiraie y mas) a a walle + ple 4) Mi + (A) Ue —@ ¥ Subtvact @ fom @ a weGajoa «Tpit ltet [u-( w) | (3 bh |ld) a = ple pete + (WO ua) NS > ye ple pint (4) th > puto vegreanve. toude whae te = Ue (EY) Uta : ‘ Pal :.,dlope coeffident for Xt ye phe pier pf tte < whee ph- pds pte pats phe (+4) W) Rational 2 Expedation Mode Hodel: People ‘Nea bok ony ros ther past experience out aleo 20 fiom the data’ that 4 ‘avatlable 49 the preeens 2a, +anieee she . wee ld) Uy a “Equation * Mtodes TW sow, ony ag egstoy, model. MES ce BX + Pa Mab 1M —+ Endo geneous “ase When Ke ‘become dependent variable —> A system dean bing the fort « dependence valiables 4s » called a ao ¥ grmultaneous equation. oo, here EK, uy #o, hence we get a tose | wiefpoent eolimator Frangencous Varieles => Exdogéren us Variables , 4 Eriglitined, 69> the, Jarcton Fie ako. pre- determined sanatles system § value of which ly Vanatte thet cn determined by the stmouttaneous | “pronde cxplanatro t antaradion of the vation tr | - feieae rare b vals which dre delevrntned Jr” the model ( jpratly determined ) an athe wooded, — gfe FRB | 9 guaniy - desmarded, 4 : Re pice of the case good Qe ft Pil + ps Ros Pov 5 pel rdated good (roo ) ee Met oubete : anarme, ¥ Gut P' cab aloo depend on ge “then, + : #. f= =F. 4,9 +O + W: weatha od adion ~~ P vera (BIRPs Bef) GW Fl" I com clealy se, (Pu) eo > ee as