Auto corel atcon.

Antocomelten 1s speial Case

not betoeen tew o(or
reters to the elatenetwp (lomelateon)
but b eteen the suecesive
are) ditterent varables
Dn elateon beteen
Var able. Dt is the covr
velues c the time
Aeey obsevateny odered in
membes o
<pe
spae (2n the Case
data). hen the
Case s - seton as spatiay
act et s caled
obse ed i hen t is
Correlatetn nspaea.
ato coYelaton Call ad ay
t teme semy data,
obseed i the case
anayns, the clasmcal
Covvelateon. The Yegesion
sen a elte n the
descws the covr
Yegresin nodel
lineas
ten Ui
distbane
atcon.
Examp le1 : Spatal Auto covvel
to one househojd o
tesr petaining another ousehod o
wcth e r tem 66
cerr eat ed

Exanple' 2
Quartes teme seney data.
labos Strike atteteng owtpt in one
to believe thal o
qusrtes there S
the dus teon we cay Oves the net quaster.
reason
tws quarto, thee is no
to be l o s y the nent quates. But
awto CoVete , ogr
deund enu, here
 CLRM ass meg that there S no asto
The
ôn the distb n i

-E(u)J[y- ECu))
iti)
E ( u ) =0(br
Destmbaue tem rel cten g to any obseaoon
othor
ntluened by the distebanu tem relating
obserateuy spa
es yel at ed
ut, et the de stlbane tem(i)
we call that ato co vreton.

E Cuiuj) +o (be iti)

atocowelaten
cov CUt,lt) = 0
Assump tons t
Teme
Ut-) +o es auto ovrelateou
but wvCut
dada 4)
No
autocovvelatdu
auto covrelaten.
selttnal i4) - spattal
data
Aeto correlatcon
Reasons tes
A
Salient beatee
tene sejes varables
netia e t a og sluggishnes
turme senes data is covvelat ed. Fa nstene
er es
BUeeesire paniod aN, Pnu
dees, proeute ,
y
Bees r h buui nes yles
teme exhibit
errplayrnot and nenplomn
u,u

Tene
(o) (d).
 Bioy: excluded Vamable Caso ()
Speutiateon

Onome Related gros

Quantty
Dernanded ongina model,bt
IS th e
the eun (5)
Let's YUN
ave
due to sme
(6)
B4X4t + ut reloten,
Vt tue
whee covrt rmo ded
OF eqvate
(5) s the tdeet aa systenhe
will eead to
ttocorrelton.
p cttenn creatngy ncorr eet
Fumete nal fo pm
outpt + ui - )
spee Pa
MC B+ pa Outpt;t
is emgmal and
oututi t vi
Ba
MCi= B+
Supplt t nart agnetmd
cobweb Phen¡mehn
phenomenon. where Supplt
conmoditoes rebtets cobweb 4eme
a lag d ne
p e wetn
reats te ().
Pa Pt+ Ut
Suppy - B)t
randam
not
(se t ae
)n s
Consugtdn t-|+r
Logs Ba nome
Consnptes t = Bi +
vavables 0s
AutoTegrusin one ot the explamtog
t the degendent variable
lagged value
3 Manipatton ot data Santer polten

4 Data Tonstomatoon
()
B1 + Ba Xt t ut
Consunpte ncome
Expendite olds in
teme
also
tee
teme t
(2)
So

cqun (il) - équ(12)

Yt Yt-t
Vt ().
(2)
Vt
callad
2quats ((3) iternu operatr
is called as frst Some tcmey we
Over teme
shos the chane than

re
does not bol)
But Vt in e4uam (a)
(ut Wt-) +0
and ths Cov
the

Hon stothatt ey
S. Non stationarty
steteonan teme on varl amt
that s
Covamanu) are
M k ho t the
( mean, Va f that is
d notot chage nstttoay temne Seey
thony do sents.
Case| we hare a
 B + Ba Xt
YE = th and thenetre
are nonstate o hang
non stet Onagi the errer tem
the esml aso
ane aultoare ton
will

Wt-|

wt-|
)
(Neutve Ato corektn)

rvelatn
Posste Ate cw

n the pres ems ett

Conseguences & Atocorrelaton eto cow el atdon
ased.
OLS estematos are wn bj
)
Cosistent
() mene mum Vorae

No longes
BLUE.
nt be
thet
Deteton ot Auto covrelato on.
araphcal Metnod
li
agamst tene
Teme Saguene Plot plo Hog

purelg t
we well no
Yeas get a patte
and hy ho
auto Co relates.

hwsa pattem.
 The Runs Test

sh the grapal meth od w

The patte rns
ntueten ot urelaten
he cheu ed
Sometmes
Puns teyt the RRuns teyt ads0
Cheang test p we note d wn the

over 4ne
resi daa
Sign t the
t+)(-- -)
(-: (+ t++ +t +++t (5)
(>) )
Length t Run
There are hree Yn.
fiast
the reti dua
there se teoo
atocerreett)
bew autrueltn

Test
3 oen- Ntson d
Dutbinanwatson.