Journal 4,' '..di~uirwa and I ppJfiJ4 d Econlomc,fl. Aix:rfr1 Jro


place. The basis *.' the determination of "like
product" centers on the economic priisiple
that if two or more gtiol are perfectlyv ho-
iln(IgciicuiLi' then all ,* lheir prices must
move together h ritcll) If a market consists
of' "like god,." the prices of these rooJd.
would also move ir~gihei. although not per-
Il c [ If the prices ol gu .od, in a market move
independent I. then the ioodsI. cannot he "like
goods."
 Jordan and VanSicklc (I''15..i used this ap-
proach to determine whether FlhnriLi and Mex-
ican fresh tomatoes were integrated in the
same market. Thc, concluded Ilr'i their study
that Florida and Mexico were in fact intelgt-l-
ed in the same market. Mexico was primarily
a shipper of vine-ripened tomatoes and Florida
was priin.irily a shipper of mature ngrLenI to-
matoes. Within the context of trade Llipue..
the conclusion reached by the authors ueg-
ec'-ted that the two producIt could be consid-
ered as "like prtodilt."
 The ,.pcL lil u.1,n used by Jordan and
VanSickle ( 995.1 to determine whether two
commodities, a and b, could be al.litineul as
"like products" in the same market was


(1) P. = + p, + .



 ;
 + 2" V + e,

 61. + p,


where P .in- I (: represent the price and quan-
tity of L.IIII.mIm1dJi\ a in time period and like-
wise P/ and Q represent the price and quan-
tity I' ~m.illiiin .lu4 b in time period t. I'h
Iclnglh of the lags used in the .inial\.i (as-
sumed here to be / prii r0. can be determined
Lhrou Lh a series of statistical tests. The distur-
bances ,' the equations arc assumed to have
the rlllh int. priq lnic,.

* 1 I I l .-[ .0
 r[... 1 = 0 for t ~ ., and
 for i.J 1. 2


 Lir-.. 1 = a

 It is re.iitil seen that the parameters of
each equation are irlkittd, I"The above sicin
can be estimated by any simultaneous eqila-
tions estimator. However. i.elln the cross-
equation h.poihii' tests to be examined,
thret-l.isc le.i-ia-quare is a convenient esti-
mation procedure with Lksrirhlc properties.
The ft\ll,_,-inig tests are conducted to deter-
mine whether ilhe markets for commodities a
and h are inrteLraicJ (the iL.i'iiinioiJiiic are
"like pridlut" I or, if not, lich extent to which
the markets are separated (market segmenta-
ion The null hI.t pthe-i, is that the market is
not inti-Ir.ie.i (i.e., it is Lrien teid). iher est
the alternative hb p ithic.,' is that it is integrat-


4 1 I ,


1 0 t i I ,


Hence, rejecting the nill hilp.the'.i. in the
above test implies that the products are inte-
t.ruted in the market and can be considered to
be "like products."
 A more gencrl approach can be specified
fI.r an arbitrary number tOf coLnmmndiCtle. In the
presence 'I more than two ..uimni tiiie., how-
ever, the system is most v.i..ily seen in matrix
form. (.i.'i.'r.tnil/iin to In Lomltii illiik. the
.elliialiuln, can be written as:

(4) YV + -B E.

with the IOiull..i iiL te Jdtrllliiiii l '


 P11
Y-
 I 'i


ii.,

 PJ
P,1r,


 X=lI P-1 -P Q -QI Q ]

 Pill P .0 PP2I
 pitl PIE

 I P~, .I


I iis a Tcricimcni ivlum.n of oneCs


I