Ma ais, D. J.; Smi , E. V.D.M.; Con adie, W. J.
A icle
Mic o-le el es s o a ional expec a ions in Sou h A ica
Sou h A ican Jou nal o Business Managemen
P o ided in Coope a ion wi h:
Uni e si y o S ellenbosch Business School (USB), Bell ille, Sou h A ica
Sugges ed Ci a ion: Ma ais, D. J.; Smi , E. V.D.M.; Con adie, W. J. (1997) : Mic o-le el es s o a ional
expec a ions in Sou h A ica, Sou h A ican Jou nal o Business Managemen , ISSN 2078-5976,
A ican Online Scien i ic In o ma ion Sys ems (AOSIS), Cape Town, Vol. 28, Iss. 1, pp. 15-26,
h ps://doi.o g/10.4102/sajbm. 28i1.785
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/218165
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
S.A .J.Bus.Manage.1997
28( I)
,~
Mic o-le el es s o a ional expec a ions in Sou h A ica
D.J.
Ma ais,
E.
d
M.
Smi *
G adua e School o Business, Uni e si y o S ellenbosch,
P.O.
Box
61
O,
Bell ille,
7535
Republic o Sou h A ica
W.J. Con adie
Depa men o S a is ics, Uni e si y o S ellenbosch, S ellenbosch,
7300
Republic o Sou h A ica
Recei ed Janua y 1997
The a icle in es iga es en ep eneu ial expec a ions
o ma ion
along
he
lines o
he
a ional expec a ions hypo hesis.
I
u ilizes mic o-le el business
su ey
da a
om
he
Bu eau
o
Economic
Resea ch
and
dis inguishes be ween phases o
he
business cycle, consume
and
capi al goods indus ies
and
a ious
deg ees o sec o al economic concen a ion.
Ve y
li le
e idence o
weak
o m
a ionali y
is
p esen
in
he
da a
which
concu s
wi h
simila in e na ional e alua ions.
* Au ho
o
whom
co espondence should be add essed.
In oduc ion
In
he p ocess
o
cons uc ing mic o ounda ions
o
mac o-
economic heo y i
is
equi ed ha en ep eneu ial beha iou
in
a dynamic economic en i onmen be in es iga ed. Busi-
ness su ey da a, such as ha
o
he Bu eau o Economic
Resea ch (BER) a S ellenbosch, o e an unique oppo uni y
o accumula e empi ical e idence on expec a ion o ma ion
and decision-making pa e ns a a mic o-le el. Such empi ical
e idence allows he e i ica ion
o
ela ed mic o- and mac o-
economic heo ies in a manne dis inc ly di e en o ha
no mally ob ained om agg ega ed da a analyses
o
eco-
nomic beha iou .
This s udy in es iga es en ep eneu ial expec a ions o ma-
ion along he lines
o
he a ional expec a ions hypo hesis.
The nex sec ion p o ides a b ie o e iew
o
he hypo hesis
while a hi d sec ion deals wi h measu es
o
o ecas ing pe -
o mance wi hin he amewo k
o
con ingency able analysis.
This
is
ollowed by a ou h sec ion which deals wi h he da a
used and me hod ollowed, while he esul s a e p esen ed in a
i h sec ion, ollowed by a conclusion.
Ra ional expec a ions hypo hesis
Acco ding o Begg
(I
982: xi):
'The
Ra ional Expec a ions
Hypo hesis asse s ha indi iduals
do
no make sys ema ic
mis akes in o ecas ing he u u e'.
I
is
an economic iew
e y simila o ha
o
he classical economics. In essence i
claims ha (a) people make he bes possible use
o
he in-
o ma ion a ailable o hem, and ha (b) p ices and wages a e
su icien ly lexible
so
ha he ma ke always clea s.
The a ional expec a ions hypo hesis (REH) was in oduced
by
Mu h who summa ized he unde lying ideas
in
hese o en
quo ed ph ases:
' ... expec a ions, since hey a e in o med p edic ions
o
u u e e en s, a e essen ially he same as he p edic-
ions
o
he ele an economic heo y' and
' ... ha expec a ions
o
i ms (o , mo e gene ally, he
subjec i e p obabili y dis ibu ion
o
ou comes) end
o be dis ibu ed, o he same in o ma ion se , abou
he p edic ion
o
he heo y (o he "objec i e" p oba-
bili y dis ibu ion
o
ou comes)' (1961: 316).
The essen ials
o
he REH, men ioned abo e, a e p obably
s a ed oo aguely. Shaw d aws he a en ion o he dis inc ion
be ween wha
is
conside ed o be a
's onge ',
and a
'weake '
le el
o
in e p e a ion
o
he heo y:
'Whils i
is
he s ong Mu hian e sion
o
he heo y
which has domina ed academic discussion and gene -
a ed he majo implica ions and policy conclusions,
o he weake s a emen s
o
a ional expec a ions o -
ma ion ha e in luenced popula deba e.
A
one ex-
eme, o example, he s a emen
is
aken o imply no
mo e han ha economic agen s will o m expec a ions
op imally by aking all a ailable in o ma ion in o con-
side a ion whe e a ailabili y is de ined wi h espec o
cos . Such a s a emen amoun s o li le mo e han he
belie ha agen s a e u ili y maximise s.' '...
A
a
somewha s onge le el, ... a ional expec a ions o -
ma ion amoun s o an asse ion ha economic agen s
will lea n o elimina e sys ema ic expec a ional e o ,
and his e sion ca ies a g ea e implica ions o he
conduc
o
mac o-economic policy' ( I 984: 58).
The REH has ce ainly p o en o be con o e sial. The ba-
sic assump ions
o
unbiased o ecas s (no sys ema ic e o ),
and lexible p ices and wages (clea ing ma ke s) ha e
come
unde in ense sc u iny
by
he Neo-Keynesian mo emen , and
ou igh ejec ion
by
some Pos -Keynesian undamen alis s.
Some p oponen s ecognize he limi a ions
o
he REH (see
Pesa an, 1987), while o he s like Lucas ( 1980), Begg ( 1982)
and Sa gen (1983) s a e ha e en Keynes ecognized he im-
po ance
o
unce ain y and expec a ions in mac oeconomics
bu lacked he echnical ools o de elop his insigh s. A new
app oach
is
o a gue o 'explo a ion a he han con on a-
ion' (W en-Lewis, 1985) o encompass he REH in a mo e
gene al Keynesian amewo k (see Ge a d, 1994).
To es he REH a a ie y
o
p ope ies
o
he heo y ha e
been ans o med in o es able o ma . I
is
possible o con-
dense he a ie y
o
es s in o basically ou di e en es s.
She in ( 1983) summa ized hese by s a ing ha he heo y
equi es es s o (i) unbiasedness, (ii) e iciency, (iii) o ecas
e o unp edic abili y, and (i ) consis ency.
.
Unbiasedness
Le he a iable
V,_kl,
indica e he epo ed expec a ion o
a iable
V,
in pe iod made in pe iod ( -k).
16
I is common o hypo hesize ha an expec a ion
o
a pa ic-
ula a iable is an unbiased p edic o
o
he a iable. A e-
g ession
o
o m:
,
= a + b
,.kll
+
e:,
(1)
acco ding o such a hypo hesis, should yield he coe icien
es ima es
a=
0, b = I, and E(e:,)=0.
The s ochas ic elemen ,
e:,,
in he equa ion should
be
unco -
ela ed wi h he expec ed alue V,.u·
I
ha is he case, hen
e:,
mus
be
co ela ed wi h he ac ual ealiza ion,
V,.
Hence, he
a iance
o
V,
is la ge han he a iance
o
V,.1<1,·
I
expec a ions a e o med a ionally, i means ha :
e:,=V,-V,.kJI (2)
Thus, he e o e m is eally he di e ence be ween he
e en ual ealiza ion and expec a ion
o
he a iable. Fo ex-
pec a ions o be a ional in he Mu hian sense, i is a neces-
sa y, bu no su icien condi ion, ha he p ope y
o
unbiasedness is no ejec ed.
E iciency
E iciency implies ha he pas his o y
o
he a iable is
u ilized when o ming he expec a ion, in he same way as he
a iable would e ol e h ough ime. In he ollowing wo
eg essions:
V,
= a,V,.1
+ a2
V,.
2 + ... +a,,
V,.
0 +
0,
,.,k
= b, ,_, +
b2V,.2
+ ... +
b. ,
.•
+
11,
(3)
(4)
he REH equi es ha lli =
b;
o all
i.
I
he REH we e o be a
ue ep esen a ion
o
eali y, one would expec
0,
and
11,
o
be
iden ically dis ibu ed.
To
elimina e he p econdi ion ha he e o e ms in hese
wo equa ions a e o
be
iden ically dis ibu ed, Mullineaux
(1978) p oposed ha (4) be sub ac ed om (3) o yield:
V,-V1-111
=
(a1
-
b1>Y1-1+(a2
-
bi)V -2
+
...
+(l3n-b
0
)V
1
.0
+(0
1
-TJ,)
(5)
Th~s
ela ionship does no equi e homogenei y
o
a iance,
only mdependence.
The
one-pe iod o ecas e o
o
a pa ic-
ula ~ iable is now ela ed o i s ecen his o y. The null hy-
po hesis,
Hi,
: (l i -
b;)
= 0 is hen es ed o all
i.
No e ha
in o ma io~ equi e~ in equa ion (5) is limi ed o he his o y
o~
he pa 1c~la anable only. Rejec ion
o
he null hypo he-
sis does no imply ha he e does no exis an al e na i e se
o
in o ma ion which could be used o educe he o ecas e -
o .
' ! e e icien~y es is also e e ed o as he es o o hogo-
nali y. A
_spe~1al
case
o
he o hogonali y p ope y
o
he
REH, which
1s
ela ed o he abo e Mullineaux p oposal as
exp essed in equa ion (5), is ha he expec a ion e o s a e se-
ially unco ela ed wi h mean ze o. E ans & Gulamani ( J 984)
p oposed a es o se ial co ela ion which is based on he e-
g e~sion
o
he o ecas e o
c;
(as speci ied
in
equa ion (2))
on
i s pas alues. This is es ima ed by:
n
£,+1
=
Ld;&,_;+u,
i
=O
o which he null hypo hesis,
Hi,
:
d,
=
O,
is es ed o all
i.
S.A .J.Bus.Manage.1997
28
(1)
Fo ecas e o unp edic abili y
To ea n he dis inc ion
o
ull a ionali y, he p edic ion
e o
o
he expec a ion mus be unco ela ed wi h he en i e
se
o
in o ma ion ha is a ailable o he esponden a he ime
he
p edic ion is made. This will be a su icien condi ion
o
he
a ionali y concep .
I
will esemble he s a is ical concep
o
a
'su icien es ima o ', which
may
be loosely de ined
as
an
es ima o ha u ilizes all he a ailable in o ma ion
in
he
sample.
This equi emen implies,
amongs
o he s, ha he
p edic-
ion e o
o
a a iable has o be unco ela ed wi h his o ical
in o ma ion
on
p io ealiza ions
o
ha pa icula a iable
(as
equi ed in equa ion [5]).
This
is gene ally e e ed o
as
he
weak e sion
o
he REH.
The
s ong e sion equi es ha
all
o he a iables ha migh be known o ha e an e ec
on
he
p edic ed a iable, a he ime
o
p edic ion, also ha e
o
be
unco ela ed wi h he p edic ion e o .
This
is no mally
no
de e minable when es ing he
REH
agains su ey da a,
and
in es iga o s gene ally e e o es ing he weak e sion.
Consis ency
When o ecas s a e
made
o a pa icula a iable a di e en
poin s in ime, hen he o ecas s should be consis en .
In
he
ollowing eg essions:
V,.,k
= a1
V1•1
+a
2
V1_2+ ... +
a,,
V,_
0+0,
V,_2h
= c1
V1•
211
•1
+c2V1•2+ ... + c0
V,_
0+m1
(6)
(7)
he REH equi es ha
C;
= lli o all
i.
Assuming
ha his
is
he
case, hen sub ac ing (7) om (6) yields:
V,.11
-
V,.2h
=
a,
(V,_,
-
V,.
21
,.
1
) + (0, -
m,)
(8)
which is he well-known e o -lea ning model.
These es s migh appea o di e , ye hey a e me ely al e -
na i e es s
o
he p ope ies
o
condi ional expec a ions.
Fo
example,
i
a,
: :
b,, bu all
o he
coe icien s in equa ion
(5)
a e ze o, hen:
V,
-
V,.
111
= (a, -
b,)V,_,
(9)
and acco ding o equa ion (2) he di e ence be ween
he
expec ed and ealized alues o ms he p edic ion e o . Thus,
i
a,
: :
b., hen he p edic ion e o is co ela ed wi h he p e-
ious ealized alue
o
he a iable, and is he e o e biased.
In
o he wo ds, he o hogonali y p ope y
o
condi ional expec-
a ions is iola ed as long as
V,_,
is
con ained
in he in o ma-
ion se . I can he e o e be
concluded
ha he unbiasedness
and o hogonali y es s a e ac ually equi alen .
I would be desi able o expec a ion mechanisms o su -
i e a leas one
o
he abo e-men ioned ou es s. Howe e ,
condi ional expec a ions, ha is condi ional
on
all in o ma-
ion a ailable a he ime
o
he o ecas ,
mus
sa is y all ou
p ope ies.
Measu es o o ecas ing pe o mance
The ou es s desc ibed abo e
a e
app op ia e when
quan i a i e expec a ions
da a
a e a ailable. Howe e , he
su ey
da a
o
he BER, which will
be
u ilized o es ing he
REH
in
expec a ions o ma ion by he Sou h A ican manu-
ac u ing indus y, is quali a i e in na u e.
The
con en ional
es s o unbiasedness and o hogonali y, desc ibed abo e, can
S.A .J.Bus.Manage.1997 28( I)
he e o e no be used. Ins ead he analy ical measu es
de eloped by Kawasaki & Zimme mann ( 1986) ha e been
adop ed.
Whe eas nea ly all empi ical s udies
o
he REH ha e been
done wi h ime se ies and eg ession analysis on agg ega ed
su ey da a, he Kawasaki-Zimme mann app oach p o ides a
me hod
o
s udying expec a ional phenomena a a mic o-
le el. F equencies
o
indi idual i ms' expec a ions and eali-
za ions
o
speci ic a iables a e no ed in a con ingency able
c oss-classi ied by p edic ion and ealiza ion. The e ms 'p e-
dic ion' and ' o ecas ' will be used in e changeably wi h ex-
pec a ions. Kawasaki & Zimme mann s essed he
impo ance
o
es ing expec a ional beha iou a he mic o-
le el using he ollowing simple example:
'Suppose ha he whole indus y consis s
o
wo ho-
mogenous g oups, each
o
which con ains he same
numbe
o
p oduce s.
The
i s g oup p edic s, say,
I 0% inc eases in hei selling p ices, while he p ices
dec ease
by
I 0%.
The
second g oup does exac ly he
opposi e. The e o e, he p edic ion is o ally w ong a
he mic o-le el. Howe e , he p edic ion a e agg e-
ga ion u ns ou pe ec ly co ec . Al hough his is an
unlikely case, such an agg ega ion p oblem ce ainly
pe sis s in eali y o some ex en '
(I
986: 1336).
Kawasaki & Zimme mann (1986) ollowed he Theil
(1958; 1966) p oposals o a sys em
o
measu es o quali a-
i e expec a ions which a e based upon c oss-classi ied ables
o p edic ion and ealiza ion da a. In hese ables a epo ed
inc ease (p edic ion o ealiza ion)
is
indica ed by a
'+',
no
change in he a iable by
a'='
and a dec ease by
'-'.
The el-
a i e equencies
o
p edic ion and ealiza ion o indi idual
i ms can be summa ized as ollows in he cells
o
he con in-
gency able:
+
P edic ion =
Realiza ion
+ =
{+,+) {+,=)
(=,+) {=,=)
{-,+) (-,=)
(+,-)
{=,-)
{-,-)]
The sum
o
he ela i e equencies
in
he diagonal om
(
+,+) h ough
(
=,=) o
(-,-)
indica es he p opo ion
o
p e-
dic ions ha u ned ou o be co ec . The p opo ion
o
inco -
ec p edic ions (labelled EE)
is
he e o e measu ed
by:
EE=
I - [ (+,+) + (=,=) +
(-,-)]
(10)
Theil also p oposed a measu e
o
o e es ima ion
o
le el
(labelled OEL) which is he p opo ion
o
p edic ions wi h
le els g ea e han he ealized le els:
OEL = {+,=) +
(+,-)
+ (=,-) (11)
Simila ly, a measu e
o
unde es ima ion
o
le el (labelled
UEL)
is
de ined by:
UEL = (=,+) + {-,+) + (-,=) (12)
The bias
o
p edic ion can also be measu ed
by
conside ing
changes as opposed o le el. A measu e
o
o e es ima ion
o
change (labelled OEC) indica e he p opo ion
o
p edic ions
which a e exagge a i e, ha is, changes we e p edic ed bu
none oc1:u ed:
OEC = (+,=) + (-,=) (13)
17
The measu e o unde es ima ion
o
change (labelled UEC)
is
he sum
o
ela i e equencies indica ing conse a i e p e-
dic ion, ha is,
no
changes we e p edic ed bu some ac ually
occu ed:
UEC = (=,+) + (=,-) (14)
Kawasaki & Zimme mann ( 1986) o mula ed bias indices
(labelled BL and BC) o he Theil measu es
o
le el and
change
as
ollows:
BL =
(OEL-
UEL)
(OEL+
UEL)
and
(OEC-
UEC)
BC=
(OEC+
UEC)
(15)
(16)
No e ha Kawasaki & Zimme mann (1986) e e o hese
measu es as
Bl
and B2 espec i ely. Howe e , o enhance
adap abili y
o
he wo measu es, he e e ences o le el and
change a e p ese ed he e. These pa ame e s a e eminiscen
o
he de ini ion
o
he Goodman-K uskal
gamma
coe icien
o wo-way con ingency ables. The indices BL and
BC
p o-
ide simple measu es
o
he di ec ion
o
bias; hey measu e
he deg ee
o
o e es ima ion ela i e o unde es ima ion
ou
o
he o al bias. A alue
o
'+I'
indica es no unde es ima ion
and only o e es ima ion, while a alue
o
'O' indica es bal-
anced p opo ions
o
o e es ima ion and unde es ima ion, and
a alue
o
'-1'
indica es o al domina ion by unde es ima ion.
Since p edic ions and he associa ed e en ual ealiza ion o
indi iduals can a y, i
is
app op ia e o conside s a is ical
p op.! ies
o
he abo e measu es. Assuming mul inomial
sampling o he icho omous p edic ion and icho omous
ealiza ion, which ix he o al numbe
o
obse a ions, hen
he maximum likelihood es ima o s
o
he ( ela i e) equen-
cies a e exac ly he obse ed sample ( ela i e) equencies.
Since he o ecas ing measu es a e all well-condi ioned unc-
ions
o
he ela i e equencies, he calcula ed (sample) o e-
cas ing measu es a e
jus
he maximum likelihood es ima o s
o
hese measu es. The es ima o s a e he e o e consis en and
asymp o ically no mal. The asymp o ic a iance
o
a measu e
(M)
is
hen de e mined
by,
ei he using he d-me hod (see
Bishop e al., 1975; and Ag es i, 1984) o by calcula ing i
om:
Va (M)
= [a~ ] ·~[a~
]T
(17)
whe e
is
he ec o
o
ela i e equencies which appea in
he c oss-classi ied able, and L is he 9 x 9 asymp o ic
co a iance ma ix
o
he es ima o s
o
he ela i e equencies
(Kawasaki & Zimme man, 1986). The co a iance ma ix is
de e mined by:
(18)
whe e
- he indices i, i',
j,
and
j'
can each ake alues
I,
2, o 3 o
'+',
'=',
and
'-'
espec i ely:
-indices i and i' a e used o p edic ion, while j and
j'
a e
used o ealiza ion in he o iginal c oss-classi ied able
o
ela i e equencies;
18
-indices i and j a e also used
o
indica e he ows
o
L
while
i'
and j' indica e he columns;
- N
is
he o al numbe
o
obse a ions;
-
i
1
is
a ela i e equency; and
- he e ms
O;i
ake on alues
o
I when i =
j,
and O when
i :; j.
Besides Kawasaki & Zimme mann (1986) who used he BL
and BC-measu es on he Ge man IFO su ey da a, i has also
been applied wi h success
by
Buckle e al. ( 1990) and Buckle
& Meads (1991) on he New Zealand su ey da a, which a e
e y simila o ha
o
he BER.
Da a and me hod
Since i
is
so di icul o de e mine how much addi ional
in o ma ion is needed o indi idual i ms o be able o o m
hei expec a ions a ionally, only he weak o m es s
o
unbiasedness and o hogonali y on p e ious ealized alues
o
he pa icula a iable a e conside ed he e. The unbiased-
ness es can be applied on quali a i e da a wi hou u he
ado. The c oss-classi ied able
o
p edic ion and ealiza ion,
desc ibed ea lie , al eady ep esen s he s uc u e
o
o ecas
e o s.
I
he o ecas s a e unbiased, he ela i e equency
pa e ns in he o -diagonal cells
in
he able should no be
sys ema ically biased o e ime. Consis en bias would
iola e he unbiasedness p ope y
o
he hypo hesis. The BL
and BC-measu es de ined abo e can be used o measu e
sys ema ic bias om he diagonal cells; he hypo hesis can be
ejec ed
i
ei he BL o BC
is
consis en ly biased o e ime.
Kawasaki & Zimme mann (1986) made use
o
he p opos-
als by Mullineaux ( 1978) o e o mula e he o hogonali y es
(equa ion [5]). Mullineaux did no in end i o be used on
quali a i e su ey da a, bu he e o mula ion p o ides an op-
po uni y o apply i wi h a ai amoun
o
ease. Acco ding o
he o hogonali y p ope y
o
he hypo hesis, he p edic ion
e o (V -
V,.,1 )
is no sys ema ically ela ed o he pas his-
o y
o
he a iable.
To
es his p ope y, i
is
necessa y o
cons uc a new a iable e e ed o as a 'su p ise' (deno ed
he e
by
S,).
The abo e-men ioned p edic ion- ealiza ion able
is
u ilized o de e mine he a ious elemen s
o
he one-pe iod
su p ise, S,. A su p ise
is
conside ed posi i e (indica ed by a
'+'
in
he able) when a pa icula a iable ealizes a a highe
le el han has been expec ed. The opposi e would yield a neg-
a i e su p ise (indica ed
by
a
'-').
Expec a ions which a e
subsequen ly ealized a e equa ed on he diagonal. The su -
p ise able is hen cons uc ed (wi h de i ed changes indi-
ca ed in he cells
o
he able) as ollows:
Realiza ion
+ =
P edic ion
:1.____.__:
I
_:
~'
:1
Thus, he new a iable,
S,,
is also icho omous like he es
o
he su ey da a, and i can be cons uc ed o any a iable in
he BER su ey da a o each indi idual esponse.
Since he BL and BC-measu es p o ide indica ions
o
sys-
ema ic bias, Kawasaki & Zimme mann ( 1986) sugges ha i
be used o es he o hogonali y p ope y as well (simila o
he way i
is
applied o he unbiasedness es ). A con ingency
S.A .J.Bus.Manage.1997
28(1)
able ela ing he su p ise, S,, o he one-pe iod lagged
change
in he a iable,
V,_
1,
is
cons uc ed and used o es ima ing
he
BL and BC-measu es.
Howe e , he e is a p oblem wi h his applica ion
o
he
BL
and BC-measu es.
The
null hypo hesis o he unbiasedness
es e alua es he e o bias when a di ec ela ionship
be-
ween he expec ed alue and subsequen ealiza ion
o
a
a -
iable
is
an icipa ed.
In
he Mullineaux e sion o
he
o hogonali y es , he null hypo hesis an icipa es no co ela-
ion be ween he su p ise and he one-pe iod lagged change
in
he a iable. The s uc u e
o
he BL and BC-measu es
is
such
ha i can only be used o de e mine he pa e n
o
equen-
cies (o p edic ions)
o
cells which do no appea on he
diag-
onal ela ing he posi i e p edic ions o posi i e ealiza ions,
and he nega i e p edic ions o nega i e ealiza ions. The
null
hypo hesis o he o hogonali y es an icipa es ha he
ma-
jo i y
o
he obse ed and calcula ed equencies ( o p obabil-
i ies) will appea
in
he diagonal cells.
The
BL
and
BC-
measu es will he e o e be
o
limi ed use in his case; i
can
eally only be used o e alua e he e o bias appea ing
in
equa ion (5).
A mo e eliable me hod
o
es ing he o hogonali y
p op-
e y would be o es ima e he Goodman-K uskal gamma
coe -
icien o he con ingency able ela ing he su p ise,
S,,
o
he one-pe iod lagged change in he a iable
V,_
1• The
asymp-
o ic s anda d de ia ion
o
gamma
can also be es ima ed
om
he in o ma ion
in
he able, and he associa ed p obabili y
es-
ima ed o es ing signi icance. A signi ican non-ze o
gamma coe icien will p o ide he e idence o ejec he
null
hypo hesis o he o hogonali y p ope y. This app oach
was
p oposed by Buckle e al. ( 1990). Despi e he ac ha he
BL
and BC-measu es a e no used o e alua e o hogonali y ( he
Goodman-K uskal gamma coe icien will ul il ha
pu -
pose), hey a e also es ima ed and epo ed as an indica ion o
e o bias.
Besides measu ing sys ema ic bias, he o ecas ing
pe -
o mance can also be e alua ed by compa ison wi h some
na-
i e o ecas s. A simple me hod
o
gene a ing nai e
p edic ions
is
o assume ha he u u e alue
o
a a iable
will
be equal o he mos ecen ealiza ion, H0:
V,
-V,_
1 = 0. This
is
e e ed o as s a ic expec a ion o ma ion, and needs o
be
dis inguished om he REH o which he unbiasedness es
null hypo hesis is
Hi,:
V,-V,_
111
= 0.
The
p opo ion
o
inco -
ec p edic ions (EE)
o
such s a ic expec a ions can be calcu-
la ed simila ly o ha
o
obse ed p edic ions
and
ealiza ions.
In
his s udy he S a ic Expec a ions Hypo hesis
is included o p o ide a con as backg ound o he REH.
The da a o he a ious BER su eys a e accumula ed
by
means
o
business ques ionnai es.
The
su eys a e based
upon he 'Konjunk u Tes ' app oach ha has been de eloped
by he IFO Ins i u i Wi scha o schung, M inchen,
and
which has been u ilized o moni o he economy
o
he Fed-
e al Republic
o
Ge many since No embe 1949. These su -
eys ha e a longi udinal cha ac e as hey moni o indi idual
esponses o e ime and a e cons uc ed o be endency su -
eys a he han opinion o ma ke esea ch su eys.
The quali a i e esponses in he
BER
su eys e lec ends
a he han measu able quan i ies. The esponden s a e asked
o compa e cu en business ac i i ies (including plans) wi h
ha
o
he co esponding pe iod a yea ago and hey a e

S.A .J.Bus.Manage.1997
28(
I)
eques ed o eply ei he
'up',
'same' o 'down'. No mally he
answe s o he ques ions a e quan i ied
by
balancing he 'ups'
and
'downs' and epo ing he di e ences
in
a pe cen age o -
ma .
When colla ed o e ime, he su eys o m a quali a i e
ime se ies wi h a longi udinal cha ac e and s a is ically ana-
ly ical in o ma ion.
The BER manu ac u ing su ey da a seem o be mo e com-
p ehensi e han mos encoun e ed
in
li e a u e;
i
conside s
no
only expec a ions and ealiza ions
o
selling p ices, cos s
o
aw
ma e ials, labou cos s, employmen , ou pu and s ocks
o inished goods, bu also demand o p oduc s, s ocks
o
aw
ma e ials, un illed o de s and hou s wo ked. The da a is
collec ed wi h enough auxilia y in o ma ion o enable di i-
sion
o
he manu ac u ing indus y in o
21
main sec o s.
Two
dis inc phases in he economic cycle we e iden i ied;
an
ex-
pansion pe iod om he second qua e
o
1986 o he i s
qua e 1989, and a ecession phase om he second qua e
1989 o he ou h qua e 1991. The phases could be subdi-
ided o examine qua e ly da a, o es o consis en bias.
Howe e , he numbe
o
qua e ly eco ds a ailable
is
o en
no
su icien o conduc es s which yield signi ican esul s
in
he a ious main sec o s. The e o e, he da a will be
g ouped in o ou subdi isions
o
he manu ac u ing indus y
o
in es iga ions
o
he expansion and ecession phases, as
well
as o he ull cycle.
The sec o ial subdi isions a e as ollows:
-Consume goods;
-Capi al goods;
-High economic concen a ion (which is associa ed wi h
ma ke s exhibi ing monopolis ic compe i ion); and
-Low economic concen a ion (which
is
associa ed mo e
wi h ma ke s app oaching pe ec compe i ion).
The CRIO (concen a ion a io o he en la ges i ms)
measu e is u ilized as an index o ca ego izing he h ee-
digi indus ies (Du Plessis, 1978). One se ious limi a ion
in
he
cu en da a se is he ac ha expec a ions and ealiza-
ions a e eco ded o indi idual es ablishmen s o one-pe-
iod in e als only.
I
he his o y
o
pa ame e s ex ending wo
and
mo e pe iods in he pas need o be examined
in
he o -
hogonali y es , he numbe
o
consis en esponden s dimin-
ishes ma kedly. Too many eco ds a e los ha
way
o
wa an good es esul s om which p ope conclusions can
be
made. The o hogonali y es is he e o e conduc ed such
ha he null hypo hesis associa ed wi h equa ion (5)
is
e-
duced o H0: (a1 -b1) = 0.
In
he ollowing sec ions he a ionali y
o
p edic ions
on
sales olume, p oduc ion olume, o de s ecei ed, un illed
o de s, s ocks
o
inished goods, gene al business condi ions
in
hei indi idual sec o s, numbe
o
ac o y wo ke s em-
ployed, a e age hou s wo ked pe ac o y wo ke , he a e
o
inc ease
in
a e age o al cos and selling p ice pe uni
o
p o-
duc ion a e es ed o he abo e-men ioned sec o ial subdi i-
sions
in
he BER manu ac u ing indus y su ey.
Resul s
To al manu ac u ing indus y
All
he esponden s
in
he BER manu ac u ing indus y
su ey we e pooled and analysed o e he expansion phase,
ecession phase, and he ull economic cycle. The esul s a e
p esen ed in Tables
1,
2 and
3.
In
he expansion phase he
19
es ima ed measu es a e all signi ican a a 10% le el, ha is,
he null hypo heses
Hi,
: µ = 0 (µ ep esen ing any one
o
he
measu es epo ed
in
Table
1)
a e ejec ed a le els ma kedly
lowe han a
10%
le el
o
signi icance (associa ed p oba-
bili ies p <
0.1
).
Excep o six cases
in
he ecession phase
which ha e associa ed p obabili ies la ge han 10% (e.g. H.,:
BL=
0
in
he unbiasedness es o s ocks
o
inished goods is
ejec ed a a 50.5% le el
o
signi icance [p = 0.505)), he
a ious measu es a e gene ally non-ze o a a 10% le el
o
signi icance.
In
he ull economic cycle ou ins ances
o
associa ed p >
0.1
occu , again indica ing ha he null hypo-
heses
Hi,
: µ = 0 a e gene ally ejec ed a signi icance le els
below 10%.
S a ic expec a ions
Con a y o wha would be an icipa ed, he s a ic expec a ions
a e (consis en ly) sligh ly mo e accu a e han he co -
esponding expec a ions o med
by
he en ep eneu s (smalle
p edic ion e o ). This
is
a su p ising esul and exac ly he
opposi e
o
he esul s ob ained
by
Kawasaki & Zimme mann
(1986).
I
can p obably be asc ibed o he poli ical ins abili y
which plagued he Sou h A ican economy igh h ough he
pa icula expansion and ecession phases conside ed he e.
This appea s o con i m he no ion ha a s able poli ical and
economic en i onmen
is
conduci e o eliable and consis en
expec a ion o ma ion
by
en ep eneu s a he mic o le el.
The Sou h A ican e idence seems o indica e ha exogenous
a iables can ha e a de imen al impac on expec a ion o m-
a ion
in
endogenous a iabl.es. Odd as i may seem,
in
imes
o
economic ins abili y, i
is
p obably be e o a he o ecas
acco ding o he s a ic me hod, which assumes ha cu en
changes a e going o pe sis ; he p obabili y
o
encoun e ing
an
e o
is
lowe han o he case whe e expec a ions a e
o med acco ding o some complica ed ( a ional) me hod.
Unbiasedness es
Conside ing he BL-measu e o he unbiasedness es ,
i
is
concluded ha en ep eneu s end o unde es ima e le els
o
change
in
sales, p oduc ion, o de s ecei ed, un illed o de s
and gene al business condi ions du ing he expansion phase,
and o e es ima e le els
o
change du ing he ecession phase.
The ne e ec o e he ull economic cycle indica es a
signi ican unde es ima ion
o
hese endogenous a iables.
On he o he hand, en ep eneu s seem o consis en ly expec
ac o y wo ke s o spend mo e hou s on he
job
han hey
e en ually do, al hough his conclusion is no so well
suppo ed
in
he ecession phase. The e also seems o be a
gene al endency o unde es ima e he a e
o
inc ease
in
a e age o al cos
o
p oduc ion, and o o e es ima e s ock
le els
o
inished goods, albei no so p onounced
in
he
ecession phase.
The BC-measu es o he unbiasedness es indica e a en-
dency
o
o m expec a ions conse a i ely o all a iables ex-
cep selling p ice in la ion a e and a e age hou s wo ked. The
ac ha en ep eneu s a he expec a iables o emain un-
changed (while changes e en ually occu ) may also be a con-
ibu ing ac o o he abo e-men ioned su p ise
o
s a ic
expec a ions being mo e accu a e han mo e complica ed
me hods. Changes
in
he ex e nal en i onmen we e p obably
so dis u bing, and happened a such a apid pace, ha
20
Table
1 To al indus y es s: expansion phase: 2/86-1/89
S a ic
Main sec o Numbe
EE
EE
Sales
6265
0.3657 0.3984
P obabili y 0.000 0.000
P oduc ion
6160
0.3769 0.4037
P obabili y 0.000 0.000
O de s ecei ed 6061 0.3829 0.4242
P obabili y 0.000 0.000
Un illed o de s 4811 0.3498 0.3901
P obabili y 0.000 0.000
S ocks
o
inished goods I
183
0.3466 0.3156
P obabili y 0.000 0.000
Gene al business condi ions 6 275 0.4376 0.4696
P obabili y 0.000 0.000
Fac o y wo ke s employed
4677
0.3517 0.3761
P obabili y 0.000 0.000
A e age hou s wo ked 4 707 0.3612 0.3867
P obabili y 0.000 0.000
P oduc ion cos in la ion a e
4044
0.2507 0.2713
P obabili y 0.000 0.000
Selling p ice in la ion a e
4679
0.3411 0.3458
P obabili y 0.000 0.000
en ep eneu s ound i di icul o espond o he ques ionnai e
which equi es ha a iables be • ... compa ed wi h he same
qua e
o
a yea ago
..
.'. Al hough con enien o he s a is i-
cian, he equi emen ha seasonali y be elimina ed om he
en ep eneu s' answe s is conjec u ed o cause di icul y
in
gene a ing hese answe s
in
imes
o
poli ical and economic in-
s abili y.
In
an uns able economic clima e, en ep eneu s
would, he e o e, end o a he espond wi h a conse a i e
'no change expec ed', e en hough changes ha e ealized
in
he pe iod
in
which he expec a ions a e o med.
I
i
can be assumed ha he en ep eneu s did unde s and
he ques ionnai e co ec ly and p o ided ue opinions, a sig-
ni ican endency o conse a i ely unde es ima e le els du -
ing he expansion phase would indica e ha changes, which
e en ually occu , a e no mally upwa ds. Simila ly, conse a-
i e o e es ima ion
o
le els du ing he ecession phase,
would indica e ha changes a e gene ally downwa ds.
Agains his backg ound, i
is
no ewo hy ha he en ep e-
neu s we e gene ally op imis ic abou he p oduc ion cos in-
la ion a e wi h a consis en , conse a i e unde es ima ion
o
le els; al hough he p oduc ion cos in la ion, compa ed o
ha
o
a yea ago, was expec ed o emain unchanged, i al-
ways inc eased, ega dless
o
he economic phase. This esul
does no appea o be consis en wi h he e olu ion
o
p oduc-
ion p ice in la ion which, acco ding o he igu es supplied
by
he Sou h A ican Rese e Bank, showed a gene al down-
wa d endency du ing he ull economic cycle
o
1986 o
1991. Howe e , i could
be
a gued ha en ep eneu s did no
eally conside he change
in
cos in la ion a e, bu a he e-
sponded wi h hei es ima es
o
cos in la ion a e le els
which always ended o
be
lowe han he le els inally expe-
ienced
in
he las qua e
o
1991. This a gumen would e-
S.A .J.Bus.Manage.1997
28(1)
Unbiasedness Onhogonali y
BL BC
EE
GAMMA BL
BC
-0.1571 -0.2350 0.7464 -0.1505 0.4795 0.5728
0.000 0.000 0.000 0.000 0.000 0.000
-0.1411 --0.1964 0.7334
-0.1968
0.4524 0.4988
0.000 0.000 0.000 0.000 0.000 0.000
-0.1637 -0.2256 0.7497
-0.1920
0.4212 0.5256
0.000 0.000 0.000 0.000 0.000 0.000
--0.1486 -0.2813 0.7842 -0.1795 0.4365 0.6588
0.000 0.000 0.000 0.000 0.000 0.000
0.1377 -0.2281 0.4718
-0.3319
-0.1866
0.0927
0.006 0.000 0.000 0.000 0.000 0.059
--0.2128 -0.1903 0.7232 -0.3038 0.2331 0.2920
0.000 0.000 0.000 0.000 0.000 0.000
-0.1620 -0.1185 0.5279
-0.2900
0.1657 0.0755
0.000 0.000 0.000 0.000 0.000 0.001
0.0780 0.1114 0.5022
-0.2556
0.2733 --0.0742
0.001 0.000 0.000 0.000 0.000 0.002
-0.1139 -0.1462 0.8677 -0.6161 0.7526 0.7687
0.000 0.000 0.000 0.000 0.000 0.000
-0.3115 -0.087771 0.8288
-0.6626
0.5395 0.6541
0.000 0.002 0.000 0.000 0.000 0.000
jec he assump ion ha hey in e p e ed he ques ionnai e
co ec ly.
O hogonali y es
The gamma coe icien s, as well as he
BL
and BC-measu es,
o he o hogonali y es s, indica e ha he e we e consis en
biases
in
he ela ions
o
all he a iables wi h hei espec i e
one-pe iod lagged changes. F om hese esul s i can sa ely
be
concluded ha su p ises we e sys ema ically ela ed o
he
in o ma ion inco po a ed
in
he ( ecen ) his o y
o
he
espec i e a iables. Al hough his conclusion is ob ious
om he gamma coe icien s, which gene ally indica e a
signi ican in e se ela ionship be ween su p ise and
p e-
ceding ealiza ion o each
o
he a iables, he conclusion
migh no
be
so ob ious when conside ing he BL-measu es
only. The BL-measu es o he ull economic cycle analysis
indica e signi ican bias
in
o de s ecei ed, un illed o de s,
gene al business condi ions, numbe
o
ac o y wo ke s
employed, and hou s wo ked pe ac o y wo ke . Howe e ,
he bias appea s o change di ec ion om being all sig-
ni ican ly posi i e
in
he expansion phase, o signi ican ly
nega i e
in
he ecession phase. No e ha he exac opposi e
is obse ed o s ocks
o
inished goods. E en hough
he
biases do no seem consis en igh h ough he ull economic
cycle,
i
is signi ican
in
bo h he expansion and ecession
phases. The e o e,
i
is
concluded ha o hese a iables
oo,
he e appea s o be a sys ema il: ela ionship be ween
he
su p ises and he in o ma ion con ained
in
he pas his o y.
The gamma coe icien s indica e ha ecen upwa d mo e-
men s
in
he a iables a e p edominan ly associa ed wi h
neg-
a i e u u e su p ises, ha is, when upwa d mo emen
is
S.A .J.Bus.Manage.1997 28(1)
21
Table
2 To al indus y es s: ecession phase: 2/89-4/91
S a ic Unbiasedness O hogonali y
Main sec o Numbe
EE
EE
BL BC
EE
GAMMA BL BC
Sales 5
167
0.4438 0.4720 0.0381 --0.1186 0.7434
-0.1483
0.0200 0.3891
P obabili y 0.000 0.000 0.059 0.000 0.000 0.000 0.214 0.000
P oduc ion 5 108 0.4383 0.4571 0.0433 --0.0921 0.7222 --0.1920 0.0323 0.3391
P obabili y 0.000 0.000 0.036 0.000 0.000 0.000 0.050 0.000
O de s ecei ed 5 027 0.44IO 0.4663 0.0802 --0.1187 0.7571
-0.2152 -0.0515
0.4161
P obabili y 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000
Un illed o de s 3 970 0.4033 0.4307 0.0573 --0.1990 0.7922 --0.2075
-0.1320
0.5829
P obabili y 0.000 0.000 O.ol8 0.000 0.000 0.000 0.000 0.000
S ocks
o
inished goods I 124 0.3254 0.3390 0.0341 --0.2000 0.4607 -0.2047 0.0697 0.0289
P obabili y 0.000 0.000
o.sos
0.000 0.000 0.000 0.094 0.573
Gene al business condi ions 5 130 0.4158 0.4355 0.0734 --0.0835 0.7409
-0.3147
--0.3160 0.3694
P obabili y 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000
Fac o y wo ke s employed 5
103
0.3296 0.3471 0.0198 --0.0658 0.5414 --0.3354 --0.21
IO
0.1694
P obabili y 0.000 0.000 0.405 0.008 0.000 0.000 0.000 0.000
A e age hou s wo ked
5066
0.3360 0.3421 0.0179 --0.1090 0.4955
-0.3573
--0.2032 0.0607
P obabili y 0.000 0.000 0.456 0.000 0.000 0.000 0.000 0.008
P oduc ion cos in la ion a e 5 108 0.2678 0.2872 --0.0484 --0.0588 0.8375 --0.5199 0.7265 0.7002
P obabili y 0.000 0.000 0.063 0.034 0.000 0.000 0.000 0.000
Selling p ice in la ion a e 5 095 0.3209 0.3386 0.0713 0.0409 0.7973
-0.4072
0.6790 0.6029
P obabili y 0.000 0.000 0.003 0.108 0.000 0.000 0.000 0.000
* Figu es p in ed in bold indica e ejec ion
o
null hypo hesis a signi icance le el highe han I 0%
obse ed in he mos ecen pe iod, he a iables a e gene ally
p edic ed o be highe han he ac ual subsequen ealiza ion.
The in e se
is
obse ed when he a iables ha e adjus ed
downwa d in ecen his o y, ha is, downwa d mo emen s
o
a iables a e no mally associa ed wi h posi i e subsequen
su p ises.
Accu acy
o
o ecas s
The es ima ed p edic ion e o s (EE in he unbiasedness es )
o cos , s ocks
o
inished goods and p ice in la ion a e con-
sis en ly smalle han ha
o
any o he a iable. This pheno-
menon was also obse ed by Buckle e
al.
(1990) and can
p obably be explained by he no ion ha du ing pe iods
o
in la ion (and in la ion a e changes) i would be easie
o
p edic he di ec ion
o
mo emen in p ices and cos s. This
iew
is
suppo ed by a conclusion in Ne lo e & P ess (1986),
as
well as Konig e al. (1981 ), ha F ench en ep eneu s end
o
co ec ly es ima e p ice changes, o a leas ha e mo e
consis ency
in
he bias be ween expec ed and ealized p ices,
du ing pe iods
o
high in la ion a es. Fo he o he a iables,
such as sales, p oduc ion olume, and ac o y wo ke s
employed, he e may be compa a i ely li le change om one
pe iod (qua e ) o he nex , making i mo e di icul o p edic
he di ec ion
o
change. This con i ms he sugges ion
o
Buckle e al. ha
' ... he p opo ion
o
co ec expec a ions would be e-
la ed o he dis ibu ion
o
epo ed ealiza ions ac oss
he h ee ca ego ies (up, same, down)' (1990).
The ela ionship should be
o
an in e se na u e, in o he
wo ds, he highe he dis ibu ion, he smalle he p opo ion
o
co ec expec a ions, o he highe he es ima ed e o , EE.
The o ecas e o s es ima ed o he Sou h A ican manu-
ac u ing indus y a e consis en wi h hose ound by Theil
(1966), Kawasaki & Zimme mann (1986), S alhamma
(1988), and Buckle e al. (1990). Howe e , he New Zealand
obse a ions and esul s epo ed
by
Buckle e al. ( 1990) ap-
pea o bea he closes esemblance o ha ound in analysis
o
he BER manu ac u ing su ey da a.
P ice and cos expec a ions
One addi ional in e p e a ion
o
he en ep eneu ial beha iou
on p ice expec a ions needs o be men ioned. The EE-
measu es in he unbiasedness es indica e a ela i ely s ong
associa ion be ween p ice expec a ions and ealiza ions. As
poin ed ou
by
Ne lo e & P ess (1986), and Buckle e al.
(1990), such an associa ion may occu ei he because p ice
expec a ions a e e y good es ima es (con i ming he REH).
o because he en ep eneu s a e se ing p ices a he han
accep ing ( aking) p ices se by exogenous supply-demand
o ces. Howe e , he consis en bias obse ed
in
he
unbiasedness and o hogonali y es s, indica es ha he weak
o m REH can be ejec ed a he I 0% le el
o
signi icance.
The e o e, assuming he abo e-men ioned impac
o
in la ion
can be igno ed, i seems easonable o assume ha he
manu ac u ing indus y en ep eneu s a e gene ally p ice
se e s, and no p ice ake s. Whe he he ela i ely small
es ima e e o
is
a s onge unc ion
o
he in la ion a e, o
o
he cha ac e is ics
o
en ep eneu ial beha iou , will be
22
S.A .J.Bus.Manage.1997
28(1)
Table
3 To al indus y es s: ull economic cycle: 2/86-4/91
S a ic Unbiasedness Onhogonali y
Main sec o Numbe
EE EE
BL BC
EE
GAMMA BL
BC
Sales 11432 0.4010 0.4317
~.0606
~.1785
0.7450
~.1102
0.2723
0.4931
P obabili y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
P oduc ion 11268 0.4048 0.4279
~.0518
--0.1466 0.7283
~.1564
0.2636
0.4294
P obabili y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
O de s ecei ed
11
088 0.4093 0.4433 --0.0474 --0.1755 0.7531
~.1467
0.2057 0.4774
P obabili y 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000
Un illed o de s 8
781
0.3740 0.4085
~.0505
~.2432
0.7878
~.1440
0.1781 0.6253
P obabili y 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.000
S ocks
o
inished goods 2 344 0.3360 0.3268 0.0863
~.2142
0.4664
~.2702
~.0641
0.0619
P obabili y 0.000 0.000 0.017 0.000 0.000 0.000 0.026
0.081
Gene al business condi ions 11405 0.4278 0.4543
~.0894
--0.1450 0.7312
~.2129
-0.0171 0.3274
P obabili y 0.000 0.000 0.000 0.000 0.000 0.000 0.117 0.000
Fac o y wo ke s employed 9 780 0.3402 0.3609
~.o708
~.0919
0.5350
~.2767 ~.0333
0.1252
P obabili y 0.000 0.000 0.000 0.000 0.000 0.000 0.016 0.000
A e age hou s wo ked 9 773 0.3481 0.3636 0.0487 0.0031 0.4987
~.2993
0.0279
-0.0046
P obabili y 0.000 0.000 0.004 0.860 0.000 0.000 0.051
0.780
P oduc ion cos in la ion a e 9
152
0.2603 0.2802
~.0764 ~.0957
0.8509
~.5569
0.7383 0.7312
P obabili y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Selling p ice in la ion a e
9774
0.3306 0.3420
~.1140
-0.0173 0.8124
~.5363
0.6108 0.6278
P obabili y 0.000 0.000 0.000 0.357 0.000 0.000 0.000 0.000
• Figu es p in ed
in
bold indica e ejec ion
o
null hypo hesis a signi icance le el highe han I 0%
elucida ed
in
he ollowing discussions
o
he a ious g oup-
ings
o
he economic sec o s
in
he manu ac u ing indus y.
A simila a gumen can be cons uc ed o cos expec a-
ions.
Ra ional expec a ions hypo hesis
The signi ican biases de ec ed o he indi idual a iables
subjec ed o he unbiasedness and o hogonali y es s, indica e
ha he weak o m
o
he REH
is
no suppo ed
by
he BER
su ey da a when all he esponden s a e pooled oge he
in
he manu ac u ing indus y.
Consume goods indus ies
The da a o he expansion and ecession phases, as well as
o he ull economic cycle we e analysed
in
he same manne
as o he o al manu ac u ing indus y.
To
conse e space,
ables a e no p esen ed he e, bu a e a ailable
in
Ma ais
(1995).
Ve y simila esul s a e ob ained (and cha ac e is ic beha -
iou concluded) o he consume goods indus ies, al hough a
g ea e numbe
o
null hypo heses we e ejec ed a signi i-
cance le els highe han I 0% o he BL and BC-measu es
in
he ecession and o al cycle cases.
I
can be concluded ha
he weak o m
o
he REH is no suppo ed
by
he BER su -
ey da a when conside ing he consume goods manu ac u -
e s.
Fo he consume goods indus ies, he unbiasedness es
es ima ion e o s (EE) o cos and p ice in la ion expec a-
ions a e consis en ly smalle han hose
o
he o he a ia-
bles. Whe eas p ice in la ion expec a ions we e gene ally
Jess
accu a e han ha
o
s ocks
o
inished goods o he
o al
manu ac u ing indus y, he opposi e appea s o be ue o
he
consume goods indus y.
The conclusions de i ed o he o al manu ac u ing indus-
y, wi h ega d o p ice and cos expec a ions, can also
be
de-
i ed o he consume goods indus ies, ha is, p ice
expec a ions appea o beha e mo e like plans han o ecas s.
The en ep eneu s
in
he consume goods indus ies,
he e-
o e, seem o be p ice se e s a he han p ice ake s.
Cos s
appea o ha e a simila plan-like cha ac e .
No
u he ligh
is
h own on he ques ion (men ioned
in
he discussion
o
he
o-
al manu ac u ing esul s) abou he s eng h
o
he ela ion
be ween p ice/cos expec a ions and he gene al CPI and
PPI
in la ion a es, on he one hand, p ice/cos expec a ions
and
en ep eneu ial beha iou wi h ega d o p ice/cos se ing,
on
he o he hand.
Capi al goods indus ies
The esul s ollow a simila pa e n (al hough wi h a g ea e
numbe
o
ejec ions
o
he null hypo heses a signi icance
le els highe han I 0%) o hose
o
he o al manu ac u ing
indus y and consume goods manu ac u e s.
Fo he capi al goods manu ac u e s, i can also be con-
cluded ha he weak o m
o
he REH is no suppo ed
by
he
BER su ey da a on a a ie y
o
business ac i i ies.
The abo e conclusions ega ding cos expec a ions can also
be de i ed o he capi al goods indus ies -cos expec a ions
appea o beha e mo e like plans han o ecas s. Al hough he
REH is ejec ed when conside ing all he a ious measu es
es ima ed o cos in la ion expec a ions, i is no ewo hy ha