Maritime containerized export flows

MISSING IMAGE

Material Information

Title:
Maritime containerized export flows South Atlantic and Gulf ports, 1974-1976
Physical Description:
xi, 150 leaves : maps ; 28 cm.
Language:
English
Creator:
Marti, Bruce Edward, 1943-
Publication Date:

Subjects

Subjects / Keywords:
Containerization -- Southern States   ( lcsh )
Unitized cargo systems   ( lcsh )
Shipping -- Southern States   ( lcsh )
Geography thesis Ph. D   ( lcsh )
Dissertations, Academic -- Geography -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1982.
Bibliography:
Bibliography: leaves 143-149.
Statement of Responsibility:
by Bruce Edward Marti.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000334432
oclc - 09384383
notis - ABW4072
System ID:
AA00002184:00001

Full Text














MARITIME


SOUTH


CONTAINERIZED


ATLANTIC AND


GULF


EXPORT


PORTS,


FLOWS:


1974


- 1976


BRUCE


EDWARD


MARTI


A DISSERTATION


PRESENTED


TO THE


UNIVERSITY


GRADUATE
FLORIDA


COUNCIL


PARTIAL


FULFILLMENT


DEGREE


DOCTOR


REQUIREMENTS
PHILOSOPHY




























Copyr i gh t 1982
by

in 2011 withllunding from


University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation















AC KNOWLEDGEZE NTS


This work began with an


i dea


the mi d4-1970s


and has


ong educational voyage.


udy could


been


Undortak en


without the continuous support


and vauab


s; ghts


of all


comi ttee


members.


I should part cul ar ly


to thank


rginia Hetrick,


who served as comm


i tt ee


cha i rpersoni


a result


of her cons tant


e nthu


and con f


d"i ce,


Uis ~ ,aiia


has been


t~:


comply' eted.


Spec i a


thanks


I


"4


Schul tz


Or, arnI


Sour ock


who both


Kf ,- c


S?


L 12S


t;vde


:cri


cisms


d(IV ice


Las ii i


r)f(~ pv)Vj


p ~- +8


(n I wi


F ~
tj~y


gave


me invaluable


of hr2


iJ C "


Sinvo


S:r R I.


;i`"' i :0


in;S


i;c: u


,i


h a LJ r"















TBiS' 1


OF ONTUNTS


PAGE


ACKNO WLEDGEMENTS.
LIST OF TALES ,
LIST OF FIGURES .
ABSTRACT. .


* 4 9 4 4 4 4 4 4 9 4 4 4 4 9 4 4 4 4 4 4 ~i1
4 4 4 4 4 4 4 4 4 4 4 9 4 4 4 4 4 4 4 4 4 V11
* 4 4 4 4v~11
* 9 9 4 4 4 4 4 9 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 ~iX


CHAPTER ONE


THE CONTA


SYSTEM PROBLEM.


4 4 4 4 41


roducti on


4 4 4 4 4 4 4 4 4 4 4 4 41


storical 0
Mechanics
Status of
arch Goals
aier Syst
Distance H
Econoretri


verview .
of Containerize
Contai nerizati

em Suppositions
ypothesis ,
c Hypotheses.


Movement t s.
Data ,


V 4 4 4
* V 4 4 4
* 4 4 4 4


* 9 4 4 4 4 4 4 4 4
* 4 4 4 4 4 t 9 4
4 4 4 4 4 4 4 4 4 4
* 4 4 4 4 4 4 4 4


Sunr aryj


CHAPTER T WO


METHODOLOGY


Introdu
The Gra
Model A
Hypothe
Data Co
Sum ary


0on, .
ty Mod
umpti o
s Test
ct ion


V 4 4 4 4 4 4 4 4 4 4 4 4 4 9
* 4 4 4 4 4 4 4 4 4 4 4 4 4 *
* 4 4 4 4 4 4 4 4 4 4 9 4 9 9


CHAPTER THREE


DESCR I PTIVE


ANALYSES.


hion


The Conbin ed
The individual
The Exporting
The Exportinq
The Importing
The IrmDorting
Summiary .


xport
Expor
States
Ports
Corn tr
Ports


4 4 4 4 4 4 4 4 4 4
4 4 4 4 4 4 4 4 9 4 4


4


roduc









PAGE


CHAPTER FOUR


STAT I STI CAL


ANALYSES.


Introduction. 4
The Gravity/Regres


C, f ft


sian


Mode


Procedure .
Gravity/Regression Analyses
Regression Analyses South
1974. .


Regress on Analyses
1 u78


A

Re
Re
A


Comparison of
Regions, 1974
gression Analy
gression Analy
C nparison of


1
and 1976.
Regression A
Regression A
A i ,s m arison


Gravity/R
Country
I ntrodu
Number
1974
Number
Number
I976~
Number
Other R
Other P
Summnary


nai3


es;s~i


auiJth5


t i an t


1 ant


-Gu

:Gulft


the South Atl ant ic-
and 1976. ,
ses South Atldntic,
ses South At antic
the South At antic


lyses G ulf
1 vses GulL,
f the Gulf R
on Analyses


Contain ner


o'
U.


1974. .
i976. ,
egion, 19 i
AMoreqated


tSo. th Atl


ant i


f Tons South Atb lati c-Gulf


of Cont


Tons


ne-rs


ft~u *


egressl I:ns


Sour:~h


anti


Al anti c-Gu4f


Siont ainers
Containe.s


4oa


oast


1974


and 1976.
on Forei gn


c-Gu


~oast


Coasts1974?.


c-C;u


stlS.


ft ft S ft t
Coast4 1976.
d Tonnaoe4 1974
d Tonnage, 1976


CHAPTER FIVE


SUMMA~RYi


CO(N LT S i


ONi;S,


AND RE COMMON DATT


Sugnary .. ... ..... 8'
4 5 S f
Conclusions 100
Recommendat-ions 104
e^o' i noln Ol -r 0 ,* o T.^^ ^ 4 i












PAGE


APPENDICES


IMPORTING COUNiTR ES OF


SOUTH ATLANtA


CON


AND GULF


A INER fl


t~



REGYONK


E X P OR TE


AND 1976


4 107


UAPr(O~TtiN


SOUTH


IIUPIU'


Ai I 3>NT


, NTA IER ZEO C
ND GCUL REGION,,


FROM


4 AND


GE OCODES


JlT1L IE


0 FOR SOUTH


ATLANT


AND iiL F


POPT 4.


ft 15S


GEOCODES U ILiZED


FOPF{GN


POPf Ti>,


ft ~ a, ft 11


UNSTRAINED GRAVITY MODELS


1974


a 4 ft


REGRESSION ANALYSES OF THE SOUTH


AT LANT C GU.LF


COAST S


MAJOR P0RTS, 1974


ft t 44 f a 4 t f ft ft t S f 1 2


CONSTRAINED GRAVITY MODELS


REGRESSION ANALYSES OF THE


1976


4 13?


OUTH ATLANT


(


LF COAST


Mi


BIBLIOGRAPHY.

BIOGRAPHICAL Si


~JORORTS1976. t...3


iL?


x~08


tli;C! r-:














LIST


TABlES


SAB8LE


PAia.


SOUTH


C AND G ULF


CO AST


ONTrA


IZED)


C A: UT ^


BY PORT RANGE


SOUTH AT VANT


AND GULS




lIED


SP4T E.


SnUTH ATLA
BY PORT


!\ND GuLF


erTAI? NE


i XPOR


SOUTH


ATJN1i i?


A ~D


Xp(& j-c: "


BY POri


TEN COUNFRI


clIV


;:i I T


lE


NPB OR


SOUTH ATLANTIC


tab


1974


Aki 197b


CON TA INE R


GULF


TS BY P


P#x

RANGES,


1974


FROM
1976


(-(V
32JTH


SIGNS OF


FF1C


ATLANTIC-GULF


TS FOR ESI !


(X)AS


GLOW,


fl1AT ING


1974


Q9JAMI 9I765
AND vi976


SIGNS OF


CO E i-FFiI E


EcSTf A tUG


EQUATI ONS


aUTHi


ATLANTIC RAN


1 )V


AND 7


SIGNS


(*. CO :e-FFr


FO R ETIMAT


EQUAT


Sii O


KCr< uL,


L


~7~Li~N7 i


frJ r; rl j














L IST


FIGURE

1


OF FIGURES


PAGE


SOUTH ATLANTIC AND GULF POR T HANDLING CONTAINERS,
1974 AND 1976 .


SOUTH ATLANTI

SOUTH ATLANTI


PORTS NUMBER


CONTAINERS EXPORThD, 1974


PORTS LONG TONS EXPORTED, 1974


SOUTH ATLANTI


POR TS NUMBER


CONTAINERS


1976


S 34


SOUTH ATLANT


LONG T


EXPORTED,


GULF PORHS NUMBER


OF CONTAINERS EXPOThO


GULF PRC TS


lONUI


TON S


EXPORTED,


1974.


PORTS


NUMXP


CO NTA i NERS


EXPORTED,


t Cs


GULF PORTS


LONG


TONS


EXPORTED,


1976.


ora T`E O














Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfiilment of the


Requirements for the Degree of Doctor



MARITIME CONTAINERIZED EXPORT


SOuTH


ATLANTIC AND


GULF


1.74


Phi 1osopooh


OWS:


- 1976


Brucecc


II~:.r C


riULS


Cha nnr a:


"q ni a


apartmentt


Hot nc


Dc: ti.r ;neii


G-cjran


This study


anal vzes


and erxp


C"it


the re jul rit Iu .


th' word-wide comiierc


maritime


C~fl a ncr~


:;V5 LA8
4> <"* "' .


~ i 4f


the pri ncipl es


of spati al i nteracti on.


It includes oth iii i


crT gi i s


des t i nat ions of


container zi


0Xi T


S i oJ


from


thIe S


Euth, Atlant ic


and Gulf coasts of the Unirted


S tates.


paucity


research


concerning spatial p


ait ternrts


rri ti i me


trade


was the i itial


ratio onal


for the st


A time-spec iic agnostic


c os s s i ctional


modei


i n t he or "-v i ty


f r amework w


ias uti ized to


teSL


t~ *i


>** -. l X\ ;F i -i 4


I t r >,i


znat

"*1


C~ 3; ~i


It V% arr n


1i


t K'*:


i r ;: rtl


r


1 ,*I.~ f r t r~ Ir~


~ aj rz r












by a modicum of variables.


was further hypothesized that


Increased use of containerization is


nversel


related to


the distance separate ng trading communities;


Increased use of containerization


directly


related to the


state of a nation


economic


maturi ty;


Incre


as ed


use of contain nerization is directly related to the


amount

general


a countryt s international maritime trade in


c argo


SIncreased c


use of


cont aineri zati on is


reet I


rel ated to the


stage of develo


pment r"


Expl nation of contain ner


a nation's land transport network.


vari ance provided


the port-to-


port regression an


alyses


was genera


However,


they


show


that


, first, total


nau ti cal d i st


anrce


com"rl er" ce


maritime general car


separating trading


orts were the be


anrd second,

st variable es


adding to exp anati


976, the


si tuati


on became maore compi ex


with four of


a poss


ible


vari abl es cont ri buti ng to


ex i nation.


most gratifying


resu


lts occurred with the


portl-to-count ry


models.


Moderate to hi gh amounts


Ilanati


wre ~"


c hi eved


four


dom es t i c


reni


major


ports.


The hypot


is that certain deri


better understanding of i


rternati onal


va I ~D al e


cod ."


contain er fl


Slead


Oro v e


to be


true


a ii t ed


extent i


Res u


i indicate


tfn Furt tePr`


mode


~inq


",~~~~~ ~Arit 4l <~V i4. AA 1 i


C I ~ Ai


pnnC


:1 ni.i


Ij ; zrz R i' E


G ,rrir rE


(rr n r*


,1~s, ~











If data could be acquired and


could


incor praterd


into fut ure


models


descr


ibing proximity


load centers, then the


spi lover


effect of distribution hubs could be accounted for.


endogenous f actors hos C


F"inafly,


ich a port can directly influence--should


be consi dered in future models.















CHAPTER ONE

CONTAINER SYSTEM PROBLEM




introduction
-- ^


This study analyzes and explains the regularities of


a portion of


the world-wide comm rcial maria tne


conBftal


sy~s te a


on the basis of the


principles of


spatial


nteracti on.


Spati al


n t eract onT


be defined


the flow of information, people or


comm and


cities from one


ocale to


anot her


C arcol


enent rarity,


Snt vering opportunity and


tran sferab


Ih-


are three


rnetess ary


pre-conditions for the seat a interaction


process


(Ui lman,


362i:- 880!)


placl


es are


c o~ I r i t


arv if demand


one pl ace can


e isatisf


m-ipg


lnfotllr. e


However


, interchange


hetpween the


areas


only


taki3ice


place if no


Suciire


upply or


demand is


nt pros ed


transferabi i l it


intervening


-> anlce~


betweer~




mf 'r1: ~


If the f


SU~lPT


factor


j I i, j :I |


i interaction wi


o3C r"


even


w th


enflt~f a':


-nd


i n ter veni n


opportun i ty


A paucity


i n i ti 3i


cnfC~l Pg~ i


rationale


f B" ie


4?


pauter rrnsF


st udv


of att i-e


posi on~ -


4-


c~ne marl-


system


-~~j


in his stud


ude;S


bot


or i gins


~nd d e stI;irn a:


ti ors


it w~~ nta jnnQR~


maria


t ~m


containerized


exvor ts


from0"7


*InbC~~k *1n~o


<1c~c nEt"X"


FLt


C t ,


r" .t K.


A


%.. V ~ I U X ;~t~ V fl LI


r e se aruch


time


s-iE'1


dr


perf eC:;


1


i ac ii


*7!


F












Carolina border while Gulf


ports


are those


adi acent
^. "'. -


to the Gulf


Mexi co


A History ical Overvi w
*. -- ...- i -..:.:-- "-.^ '* 'i~m--ma.^ < -^ 111 ^ **- l -*** .iiiiin


Containerizat ion and other forms


un i i zat i on


have


beco me


standard methods for transporting general cargoes


across


the oceans


today.


The object of a!l forms


un i ti zat on


c arv


i age of


arg e


quan i ties of het


erogeneous good


friom origin


to estinaL


Without


ays in t'rsferi^ n g between mo


Without


i ns pectio


or Ot


dmi nistati


Incumbr ances


cross ig


Snter nat i onal


K" u fluaritT 2


W i tot break


li< ~ir:tii


arv w~i


OrCce


final


s t i nat i ri


() iQer


I -,
'At/i,


Al t uah


coiru~ie r 1:


f~la y


rrn~


cont i nci


their


iM -i rc? 7 L


appear ance for te


d1in i n 9


at 1950


'x>i (ICC


pack agin goods iu thi


p nir~


U. G~


rr


fl ~C 1 :id


irodei n


comI er


mri time '


cc nt a' tMCI


movements


3inO~ ar ~d


vertis ient i n


Natpona! M~qaz~ne
- -~- -~ ,- -. -


n April


Lift-


Vans wVere


provided by


ing G een Storage


mnpany


New York


household goods shipped between the Unite.


Sttes and Eu rope.


From


World War I onwar d,


the unit ed States Armed Forces used a small


version of the container (CONE


) for overseas shipments to insure


intact delivery of va uabl and pi ferable mi i tary cargoes.


In 194 9


John Woolam, a shipper, began packing sporting goods into a container


;


1,1


"i r"5 i':


ar i i


(






3




However, it was not until Malcolm McLean, a trucking magnate,

acquired and converted two U.S. flag tankers and offered a coastal

service from New York to Houston that the modern era of containeriza-


tion began.


In 1966, once again under the direct influence of McLean,


Sea-Land commenced with a transatlantic service from New York to the


European ports


otterdan, Bremen, and Gran gemouth.


Other early


attempts to prove


de containerized servii


include Grace Lines'


ervi ce


to South America and Matson Navigation Cocpany' s


servi ce


to Hawaii.


Thus, the


new innovative technology in mar


ime transport was not


adopted unt i nearly 50 years after its concept on.


Some of the


principal reasons for the industry's earlier reluctance to readily

accept containerization include the following:


it is a


api t al 1 ntensi ye system, thus it


beyond the


financial ab


ty of many sh


i power s


as wel


as entire


co untii ;


The veh


c e of the i ndrus t ry- -the


s O ---i"i


a ifes


pan of


OVer


twernty


years


t:hecr ezF-


ore~


t~r~jdes


are uti -ed
C"E '- 4 L d^'' .* -


si owly


reqtU i res


a sufficient


eni i er


C gi


i ni hlot b


rec.~


tions because empty contain ers


timen


tsll


boar


space


as frul


cZrieS


Neverthe 1


~he


y ars


Gcomp


et e


new concepts


i nvoivin


contain neri z ati on an


ot h er


un i itized


met houds


ha ve


taken ihol d


world-w de,


The Mechan ics of


Conta nr zed


41o tne


i~ ;i;TI i"


~S Ii












suffi ci ent


to insure the


&uiipply


both


iehic le6"


capacity.


in i' international


mari t i me


tr ade


icecrinrial
i: a


cGener a I


C~r (Jo


Lransport medium


conta


inei z ati on.


The contain er


revolt ut on


became


a reality


in the


ate 16Os


resulted frmxn


three


f acto


--rapidly


i ncreasi n


Slabor costs, enonaou


expansion of international


trade


in 47nufactu e
'niu .cture


co0ds


:ncreas ing


B*pet ition and


l power


profi


n the


sh pi
hr v'


-n u str
U t-


am~-~oe9


1971, 54).


Dur ing


the I960s


?e


(XiSc


Crd2


ing doub led


percent.

necess ary


while


hcit4.-(ost


b ecau:js


the cost


t~ic eflt 3


i ar i ne


eqi r"' l


ncr 'iase


t. e rt


2 ntir f


'6flt ,3


I ~: ci


i t&t1E .


A >


(i nno va i


h ..b.a .


5 icr 9 "3Y;?d


i3or~ urn "


*r


Cont a i


nerizat on


i n i e


rcu~aC e. a


K7 i


/i 7i it


Av 2


of a


size


rent L


i~nq thcii


rr


ed~n~nn


Wfl&) "


handled most effici


cntl


f at ons


4197%C


S t-os t


C3i rtL di1


are standard unit


feet


i enjgt


ten f et


wi th


maximum


t ieet (Wjh it takicer-


'1975,


major exceptions to


the International


standards Or ga nizati


recommended T


went y- foot


Equi va ent Uni


(T.E.U.


are Sea-L


and'


foot size as well


as the 17-foot and


24 foot


sizes used by


gr ace


Matson L lines.


Recently, MAerican President Lines called for bids for


construction of an


initial order of


45-foot contain ners


Ports and


);


[p~ ;ifl


41


i


/ '1


162I


~ers


Srs.o


~:: i : i -al


r~iiic


~ tilj3:?:~i





r5




I.S.O. degrading the present interchange of containers among different

carr iers.


Specialized contain ners


have


been


designed to


carry


a wide variety


of cnw oddities.


Some of the major container types include


The covered dry container, the most conmon type, which is


suitable for manufactured commodities shipped


bagged


n cartons


goods.


ulated and refrigerated container which


utilized


for perishable


carQoes


such as meat, dairy


products,


fruit,


etc.


op loader container which permits over-sized cargo


loaded from


to be


top.


The bulk liqui contain i ner which al s for the carriage of


wine,


spirits, chemicals oil,


etc.


bulk powder container wh


i dea


for shipments of


fertili


zeIrS,


client,


etc.


The bin container which hand


buIlk


gra in,


coal, and


simil


cargoes


et m


conta i ner,


simi ar


to the


covered


type


without si dings,


used


pping drums,,


wire


cert


ai n conlnero


ve h i c l es


Branch


1977


188-


189)


hough not all


cargo es


can be


count


ai neri zedr newr


cont ainer


types


sure


ernerr


qe to faci


it Iat e


carr rce


of goods


not present











point of destination (consumption) is t


pci fbi a


obj ec ti Ve


of con


tainerization (United Nations, 1970, 3)


The funct i on


a container


is to ease the storage and transport of


them during sh ipment


den Burg, 1975


, to protect and preserve


and to insure their effTicient distribution (van


, 3).


3y tihe mid 1970s, containerships were operate ng on


mos t


of the


world's major trade routes.


trade routes was caused by


Early


carr i PIer overcapacity on certain


an art


Iy hi


level of anticipated


profits, and high capi t a intensive investment require ents c au ed


operational losses


i the early


ye ars


The sh


pping i


industry


general and


espec


the cc vita ariezed


trade


suffer


2> w~


excess,


capacity alternating with per iods


high demand.


Ve~s sel


capacity


responds quite


to. cha niges


-n demand and reducti painfuby


slow.


In the


s ho rt run r th e supply of


vessels


is ine


astic; thus,


shortages of ships are re elected


i, upward price swings for the move-


ment of cargo.


On the other hand, voyage c02t


are fixed and are


Largely independent of the amount of car o carried.


T` rieref~ ore,


economic


ac cept


pre)sures often exist


ca rgoi


-4 ie! iUU5J


power than usual freight rat


excess capaci c ty


thus causiln! rate wars,


Future growth


conta i rner


novem ents wil


certa inly be c osel


tied to and adjusted by both


ical mark t forces


and increase es


trade


i itt O~l(OY


1 1975,


-30 ).j


The Si


atuls


u -C L~r~td~tF2Urieiz :I diif


Data


worn


510"1~











i nteracti on mode


are approp rate conceptualizations for meaningfu


research into the geography of


international trade flows.


The spatial


interaction theme has been applied to total


between nations


international trade flows


Johnston, 1976), but the difficulties of obtaining


port-to-port United States Foreign trade data have


gated the


majority of geographic maritime research to studies anphasizing site


rather than situation.


Site refers to local area conditions, while


situation refers to the effects of events in one area on another


Transportation flows


research about


are one process pertinent to situation.


import ance of


Notable


uaticn i n maritime geography


includes those by Patton (1958), Carter (1962), and


yon (1970).


Since the early 1970s, the United States Maritime Administration


(MARAD) has


been


e C i


annual container statistic


Unfortunate-


Sy, MARAD'


pub i cati ns"
i^^ \^ O \r sw SI


include only aggregated container data on the


essential trade route level and only report cargo volumes which pass


through principal domestic and


orei gn port


However, intense compe-


ti tion result tin


fron the container rev


ution


is also experienced by


medium and small ports and their status i


nor e


precar ious


than that


of the l arqer


will be hi


ports. T

y dependent


he survi va


upon


successful


to media um


l e~r no~r~t


regi


compete


i onal ports

ition both


among


t hern s


as wel


3s with


ports


The contain er


revolution has rermi


tted a


r at on


of both exi


sting


and por ts.


The number


of qenera,


cair go


Sarr


S1 5


en gaged











servi ces


on fewer ports (load centers),


these |inehaui p


ti rts


connected to extensive tributary


-*t ~
a~


h i nter an tds


Therefore


efforts


attract a reasonab


level of


cargoes


are successful, the


future of the smrnal


to medium regional


po rt


be precluded


res ult-


ing either


thet


limited use


as feeder ports or even in their


Closure.


Research Goals
. ^: C-u \ &~--TI -W ~>-


This study will differ from prev ious i nvestigalo i


s ice


it wili


consider only the movement


cont aineriZ cl


ca;rgoes.


Two distinct
0 ?T S'J Ji* "w*i ?wtr


research goals


are as f o w 1 s:


To monitor changes in the system between two years (1974 and


1976);


To provide a detailed iuant itative breakdown


the major


f actors


ectin g


a container system.


These rese arch goals


have not been addressed, eva uat.ed, r tested


either in the g-eorap i c


or in the contain nerizat ion


Container i zt i o n is presently the dominant


means


i erat

by whi


ure.

ch general


cargoes are transported


over the wor d's


sea routes.


This domi nance


rests on the premi se that the majority of international trade is among


nations with


de Vel


hoped


ec onomi es.


P ese


',tinn


are the


P "imarv


traders.


of containeri mov ements wil


o nt ri


bLte


further geographic understanding of


trans o : >f t
^ ,- '|/' ^ ^ ,: ^ I*E I **


networks


descri b i ng and cl assifying emp r ic


requla 1i


.


t" ,






9




sufficient level of explanation can be achieved, the first goa


monitoring change within the system is straignht-forward.,


Once


vari abI i ty


isol ated, questions may


addressed as to why changes


have occurred.


Answers to such questions could provide additional


input thus aiding regional port planning decisions,
Container System Su pos itions


Generally, the thesis


of thi


resea r ch is that, al though flows of


international trade in maritime containers are dis


persed


world d-wi de,


containerization will continue to be concentrated between and among


i ndustri al i zed economic es


The working hypothesis


rather than between developing economies.
that spatial interaction and spatial


variation of a container system' s flows


can be


1 ained by a modicum


of variable es.


More explicitly; however, it is hypothesized that


Increased use


of containerization i1 i


fi v e~s e'ly


ated to the


distance separating trading communities;


Increased


state


a nation


cont a er i zati on


ec onumic


directly related to the


m713 Alri5ty,


Increase eI


amount


V~i;" C


contr n t ri z at i on


a country's


i ntern.at, iona


dir ect Iy


lug' -itrme


re ated

trade in


to t11


qe orr I 'I


cargo;


use of container zati on


direct ]y


related to the


stage


deivel


opment of


a nati n s


transriporte


network


The Distance Hypothesis


Ta











seems to be avoided


because


contain ner


r eut0'1 e


jr cons~i der~ed


i "-)


stable when c:npared to


conveitl ond i


time


rout es


required for


ii ty

cargo


occurs

hand Mi


because


equip nent


'e high


cont


capi t 3


craIes


investment


ports.


Sh ipowners


have


effect
I-^


Sve y


transF erred the


reso


ty of


providing


cran es


to the


bse iiUSe


ioutll


such


d~Il


pment


the ports


cannot attract


ca rri


and,


wiltlu:Ut.


carr ers


revenues of the ports


would surely diminish.


Mention has been


mnade


of the optimum


one-way


Sdis n C _lC


4


that


a contain


scale


e (Brother


N nations


port


S on


, 1970


comp ared


a qi


y~en


wiz


1966


SEven


nmust


t ( ve


L 7" ~i, ~ni9,


t hou gh


tojCC cnvent~ ; oncfl9


spe I


e~or;onmi es


I CV>


y) nt a n ~ssshios


~t2V C>


econo1in w)


Uni tei


tIe


~r e ~COC


profound on hau


i I T1 DOC It


amce


Ft 3t i


'ji c ri c o;


in n O*


r 2)S


* >: a


1 i i i ;h In


1 nc:>

( atham- Koeniq i


surrogate


1971,


for t ranspor


656-6 7>.

t costs,


t I1t


ii


lter iCt Vt i


Sin iual


co ntainer t


ewsi~


has not been


S P2


cd V,


address l ed.


Mayer, (1973


S 150)


not ed


th at


mo~r e


&itx


perce nt


the tot"a


ori gi n-to- desti n ati on


mov-nents of


Cacao"


con ve nation al


br~kakfltt k


methods were


attri


buted to


port


costs, when a


ship


is e


tiner


loading


or discharging cargo.


However, because


earl'


y anophasis


n conta nership


was on speed rather than fue


economy, recently


ri s ng


costs


have


tad


i:;ir 90


ns i 31 1


~+ti!,


21;


rif:


"cr


:e? at i va





11




economical to run vessels up to 33 knots because in the prior six


years bankers (fuel


had risen from $18.00 per ton to $190.00 per


ton (Ha!l, 1980, 32).

The Econometric Hypotheses
The last three of the four hypotheses i manate from readings of


containerization literature


where a notion that certain characteris-


tics can contr ibte to an


expl nation of


a maritime container system


is implied hlut never tested


Coup er


S1972


London


Chamber of Co1 merce,


1968; Meyers, 196


United Nations, 1970).


151-154; Rath, 1973; Strom, 1972,


The major vari


46-57; and


les allude to include


we ll-developed economy ies


high degrees of industrialization and/or manufacturing;

large throughputs of general cargo; and

mature I and transport infrastructures.


A well-devel oped economy influences


a country's buying power,


therefore, enouragi ng


demand


for di


ifi ed


products


i increased


trade,


The degree


industry i ~zat on


af f"cts


the supply


pri ncipal cargo carri e


in cantainers--manufactured joods.


largee


volumes


qener a


Cargo


ar necessary


Ueau


the hi gh


capital costs


vesse


port.


Sj Li b s


qen:erzai


cargo flows do


d co tainer


opera at


in icur


because of unused vessel capacity and


space.


bal anced bec3us e transporting empt


he flows must al so be


; enerates no


revenue.


fixed


i~u *e s











The statistical hypotheses was that distance is not


import ant


as it ha


been demons treated


in previous maritime stud i es


Niievert he-


less, it was expected that distance wil


contribute to an overall


explanation of the geography of conta


i ner izatL i on.


ant icipated


that the


independent explanatory variab


will enter the model in the


fol ow ng sequence:


General cargo volume


B a lance of maritime


general


Distance; and


The proper on


intern action al


U'r ade?


i n manu featured goods .


It is further hypot hesized


r erai r i p


Stwoi


vaniabl


es, tihe


eve i


econorMi c


de ye'


I opment


and trans r tat t i


nfrasL'ucture, will


not contri bute sign i gri i f c nt


i0 t he


no lel:


latter


were i introduced


pub:


bec a ujser


I tr1ir


qua i ati


(Nagors


ret acnc as


55:


mr ;


and Sun


S;c'7


:iLl amos


156-162).


So ImarI


APP ro ?x Frat 31 y


fifty


years


assedS


unt i


the maria ime tr. nsport


industry


dop ted




as a working


i flflOV3


a~ti 've


~fl o C


Since the l ate 19


container zat i on


ot her


urit ized methods


hadVe!


been accepted


as the maj


forms


i nternation i


Sma it i i e


transport a-


tion for the carri


of general


cargoes.


f uncti ns


mechanic sms:


necessary for the perf o ance
!G.~~~er o- j i t

a maritime cont


1 fnelr


tem have been


u theil


Li; cr t


ii itri?~l:5


C











It also sought to provide


a detailed breakdown of the major factors


affecting the


abo:Ver


conta i ne


hem


lid i Z f l9


sei-Verall"~


qualitat ve


variables all


uded


to with n the l i terature


co nt ai eri z ati on,


study tested


the validity of statements stati


tic-alt


Chapter Two


defines the stud ys methodology, while

descriptive and statistical analyses


Chapters


Three and Four report


final chapter


, Chapter


Five, suiiarizes this


work


and pr videos conclusions


reconmendat ions.


r k ers















(if57r


MFT!"(10_ HODO t:


Accordi nq


to John


S ta n


1976,i )


thc usi of


sySt


C onIc

geography is re at ie y


recent.


System


be defined


as a set of


i nterrel ated el Sneents i r whi ch


el FI" e~ne


i. erac. tis


Smith


one or more


of the others,


h 7PSes


n~ cCL 4:tcr~s


are ne- cessar y


system' s funCLion.


5zir: g


9 l ~i nf cit c to fl P


Siir


annI 2)


ach S


general fi ndings, in


I rJ


r196%3,


thre


h:a s 1 I


tomprcI inet S


ma') tinme corwa ly>


stem


ic ude
-i. u' ~e"y


ports


whi "


r*


cnrac~ 17teir


c> ii l U?:3


P at I; 1;


ccc ;


~ t4d2 :1


opo


V


se 0|
-> ^^ of\


- a 4 w- 'i


:~p Aj~42~ri


4


)ktl th~r


h fi i~ *? KJ


d~fl0 f xi: I"


"i-si rst,


tV/K) C (X nOOC ~ it


< ~fl


54


<~


system


a pa


A) t'rr


1'!i


Smoonrnt


pat tern


iter3cti on


river


Since


proposed


research


to nxpflun


t he s ati
I~ Sc I If ^\II I*


~:~:Lita i on


Sconta; i nFr


fl ows, emphasis
'W S I IS r*M ^ ^**E^l^^'"


was i


ally
a:'v


ire cted


towarJ1 I


the str uct w- e


container system


a exc usi ve time


per id.
opC" 0 i o


Therefore, a time-


-i


:-V


j~


I,


rCIClif rt3ilien ts


II tit? i ""


r~i


1`? 1)5


:: j? cr


, 1






15




The Gravity Model


The gravity model has been


used


successful 1


in geographic


investigations


urban movement and international trade phen mnena,


among many oth


er"S


SGravity and potential concepts


are determi


stick


in nature and have been used both to discover fundamental


rel ationships


as well as


prov


a basis


t ow ard


the development


theory


of traffic


flows (Carrothers, 1956, 94).


C .p


stated


gravity concept of human


nteracti on post i


cities


that i interaction


between two centers varies directly with some function of mass and

inversely with some function of the distance between them \Abler,


Adams, anrd Gould


Successful examp


1971


221-


of derivations of gravity models applied to


geto-~


graphic


rese


arch on conmodi


have


been


dCCO~fl


i hc


Yeates


(1968, 134-141


nd Bt ack


97 2, 1 0S7 -413. .9


-i e ,-


qrav vy


f r am ewor k


is suggested C


as aprop


r I


ve ari d


j?4flthe


r~egtfl1ari~tu e


a mar t i Pi


ai 11


ot v t"i*a.


Al c1acert


on pagc


in Chap er


are qual i t 3ive


because nowhere


o Re. >


the i i e"


report t hem


"4~


ai~toi ie


set.


The complex ities


st at. istica


testing


ueli


tattve4/f


siqet


that a conversion of such variab


to a quant


it ti ve


appropri ate.


Where data for the specific variables are unavai labe,


surrogate variables were sought

A cross-sectional model, in the gravity model framework, was


C~ir*1 5


3 i iii r* ~


~et~


m;?de












aniial7ys is .


The variable


in the modep are


as fco


louws:


-- f(lNP


EGYj, TCM, RiT *, CVH DST4)
VJ J


where


IVC. was one of the following
3


die n de nt va a5' d1


sTNXj as
S _i


Clc44IUrbi


of cont a i nr s


exported


-


Atllt ic


C) ~?ii


IW


K wid h


pe "


the :nitel


tate


TNSX


n UmJ~rt


lI )


$


GCo Un-


C"


; 1


- I 4'


:oas<


Str 4t,


GN~j


~ias the


r:at iilc


A
. S


icj ,?Jdt rv


tirve


year;


was the energy


20 cl ~;73 1; ?34


c. P ta


coutr ,


n the


res pect i ye ar


:wnicn is


S surrogate
t S' *-


tie de


gree


rnanru


facturi in country


T was th tota maritime oerce in general cargo of country


j in the


res pect


ye ar


TEj was t h r atio "


mar time ge eral


cargo


imported


over


that


exported from country j in the respective year;


CVHj was the number of commercial vehicles in


country '/


j in the


respect v


ye ar


, which is a surrogate for the degree of deve


opment of the land transport infrastructure in country j; and


DSTj was the distance of the foreign port in country
3.! iJ ^ I^ l C '3 '1 T 'h^ Tf^1a f^ 1^^li lS ^'^l^ 1 !


from the


-c i


: or


E-'rY^


:xii


s


I


ij ro~S












manufacture ng; and


number of commercial vehicles


representative


of a country'


land transport infrastructure.


The function of the model


is to test


the hypotheses


create


iaw-


like staternent(s


he two


riloda 3e


srde CF i t edl


twi ce~,


once


1974 data


and again for


In this


cetnp~ir I SOr1~


of both the


coeffici f cents and r

Significance of es


cetA

mat in c


Sati on


equlat


on;r~


;:Ft~


each


ac i ICC


? model


wS co'


rg'5 rss i 1 n


and the co eff


i e' t


ridE>~


pendent vari


tables


wh i ch ef entered


rf ere nti j !


statit I


were uti


ized


because


the ohen rnena of co ct i nerized


were


exami ned


for two discrete


t im r


Pe, cinrs


In f


f 1ows


of containers


are continuous;


and two discrete populations are not


necessarily


indicative of a continuous process


Model Assump tons


The use of


assumptions is


a co ron practice in scientific research.


Assumptions simplify reality


so that the rese arch can move from the


specify i c to


he general.


An analysis of the container system therefore


requires the fo1 ]wing assumptions


That no artificia


boundaries or barriers restricted the


movement of containers; no specific trade policies existed;


That countries which were


andlocked utilized the nearest


adjacent maritime facility; no restrictions because of


political situations


Pvident:


i j6 90 i


r? aC~t


WP"SP


C h re r.l r3rl


~II ~""'"1












That


a uniform


type


carr; jI K


or s h


ser vi ce


areas;


transport


i nno v at


acting


ODrit


Ca rrI


introduced


Th at


duOr i n


the cost


the~iA


the i. udiv:


F


tr a; Do<


V
~ >2 K


ti CA


spec


That


rate


hu:tan


Ar LB. turn


(~j5


(fli?


aout r


reFlected


<~ii nd<


y 1ng


I


V V? i,.


effects


s i 3ccl;


p n nomi K


-~
t iv> rwe';cnt 4


above


a?5s u'r tr3 l


a- gqenrcc


to most


j ]Kl 1:: :

~~> I- e"is


In the


case


erna


t -r -I


assl ir: 1o r-


trade poli


ci es


4 t
crei


cml n ated


such


pr acti


aS <0;


nation status and


second assumption


lso


quot as,


, no restriction


tar iffs


bi;ecaus


emzbarGoes.


of p


si LUt OI


precluded the


uat i on that


high overland transport


costs


a nation wil

, Assumptio


have


"n three,


to incur


an


uIrneceSs


qual


maritime distance


n aNl


direct


ions


that ,


i. th


the exI-ept


navigating around land masses, the


cost.


of maritime


transpo rtt


equal to t


distance


trave led


fourth


ass umpt


, no transport


innovation affecting container


car-ri ers


during


he t


peri


of the


study


, did not permit the use


ves se


wi th varying


sizecs


, speeds


capabilities; thus


no cost


advJa nt ages:


were


eve


u~t i


iziny


type of ship over


anot


her,


Assumption five,


no special


rate


s uructure


310u'h


3K


1?i i:


j. ilt1


*- 1 < '


flc1 Z


S


tt


was


1 t~rF~


" i


J i ; t Lf r;i:


i _, I


i


i .i:


d


tr ade





19




stated that importers obtained goods from that place having the


combination of the lowest production costs and


east transport cost.


2ot hesis Testing tj


In the


initial step


the nu hypothesis that spat i a regular i ties


do not function within a container systen was


tested,


Paramet r c


statistics were used because the data


were


in either


ratio


or interval


form ,


In these


cases,


par asn et


can be app li>


af er


data have been standardized.


This


permit ts


use of


inu~re


powerf i


methods which


increase


until ity


research


mon i tored


cha~n q


w i t i i


the system


a thr>~- y


time frame (1974


1976


~iqn ir; i cane


r a r amet er


est imate


were


utilized to


develop


Si ho~rt-


c~nn


tlhr eCho 1:.


valui"es


S 2Sten


onCIiPo,1Cnent-


System


anqes


bc th


pu s'


Ut i:,


can be


entialy y


explained


advances


or dec i


n t he


eS C, 0~1?x 1


it VcI1


-being


individual import ng nati


and the


leave


1 nvesrnierit


appor ti roned


to specify


dcmesti


maritime fac il


ties.


Thus,


the real


effects


interport corn pet


can be considered.


Data Collection


Dependent


variable


data for the conta


model


obtained


the United St


ates


Mari t


Admi ni str


at ion


(rutARA~j


MARAD


the only


agency currently canp


c nt ai


Stics


vihi oh5 f


i Vto


port-to-


port data.


3Je~ncy~)


i~ ?ires


Conrt 3 niCr~


t "-f^ I I


an annual


report


which


aqy1 -ire &J t E s


:iurnmb er


V.-, I?


SIi ner>*


and tnt:


4f-"


,in I?)Si


~t at i sl:i cs


s 1,3 i, i













disaggregated data


were


ge rco


i dent i Fy


jfll 1`: u :? i'


;li OaS


or gi


eating


at South


~ tl ?i; It. 1


coa t


ports


d. t iV
^ c r 1f^,ct^


Cc~r


ports.


Data


the i depe deni


var, iabei


acqu ired


F ro~;r


\3r iJ OU


sources.


Foreign country


perta ini: i


co total


and the balance of qenera


Cargo


ow C


Swel i


er capital


eflc~rjvg


consumption, and number


of co mn erc i


ver hicle


w ere


_ .itr ac ted


frcxn


United Nations S at'i stica Year book for


appro pr


ye ar


Gruos s


national product statistic


we re


obtained




the Europa Yearbook,
++ ^ +@w w <'-- 4 ^ ".- *fi **.- .__ .-^ --m><,,__ .-


Finally, distance


between


Uni ted States


*i *- :<


~ia


c3 i 1:idesif


i n forei gn


cOUntri es


wn"et~e


calculi ated


f" or"


1 i cct i sn


Di stances B:et|ween Po rts.


o Ufly5i ,E!


rati onale




44
(2> s>


f ""~ii


toi 2.i',


doil


concerning n


~pati +ifl


patternls


hI

' ,


t:ai neri zed


tnV{ mQ~i


he sn ~:d


f- i`


Ii I


C


P &t U *


COmm elir c


i ft.: etr niat 0%


:i:< fl:t


t (K Fr ~i


";citv"<


states


So uth
J %


Altat vc


and ";u i'


-rflc


o ur ts


OHI C Q


K ?


0 -


les of


spat i al


~4c~ut~K 1 h


analysis s


deter i r i St


in na t ur e ,


??mp icsd


q3 I


framework to


test the


extent


spatial


a rit i es


in thi


~i it er -


national trade subsystem.


was achieved, change within the


signr if i


systen


cant


leave


was moni to red


expl nation


over


3 thrPee-year


ft:?i


1


3 i:t 3


i at e


1 I? s


1 i? i", --


: ;3:












regionally based container rlow


pat, t~rnrS


hr?"Uiii~r


Ymle


can p ro:v de


pl manning too


specific


contain faci it '


1mpr;l- ?V:<.iicen i;













p~ 4 ?) rr1


S:SCR-+ PT


:T i E <1


i: i: K


i rtrrL ti -


Asa


ri rst


step


t ov ar s


A bet~ I


*unt rst adin


n.Pt Vj br 4tutfjl


tic-Gulf


Coas


exnV:.3rti


cont


t11i r


vi es c' i


i c~~t v-


analyses


wO?"


performed pri


to stati


sti


VS e< (


s tr atie


provided insi ghts


re a rd


th i dt a's tepora


r~evu


ies and


rr el-


u] ari ti es


Ana yses


were


pe"irfOnrTT


on t#40


eve1 s:


From


perspect


iye of the


doniiS t


pa~r ts


ci~r acr


as export


forel hands;


From


sition


of the fore ign


Qrto s


of desti nat i on


import


hands.


F ore] hands


accord n


to geori aphers,


are the


En1d


areas


wihi Ch


lie on


seaward


port is


si de


connected


port


by oce an


beyond


marni


carr iers


4 r


T h.us


Sparces,


and wi th


the concept


which


"forel


and,


as opposed t


"hh nterl and


" eight


ap~pl i ed


to all


tuat


prov i dedl maria


13S5-2003


time


thi


traffic


viewed


study Un


from


at es


g~end i


ports


are export-


, 1958


fore] an


foreign ports and forei gn


ports s


are import


oreIands


for United


ates


ports


. The fol owin aanal


ys es


ncl ude


n tmbers


coi" tai ners


tonnage shipped for the years 1974 and


!176.


r; u 17 S /;5 ~:1ITI S


port











n ~e Lombi ne


R~r~


r anqeli


COnS


Fst;,


riia A


: C j hti


t. fl 1 it


s pe c i 5F


cd i I'


def ined


q~o qr aoh
4


2?gre


4,C j?


dual par t


into a port range,


st atem ents


can be


made rega di


an eni


or sub-system


m'Fan


time fac


t. es.


ranqes


aue st


d here,


the South Atlant ic and


the Gulf


coast


Over the three


year


study period, regional cona I


ner i zed


exports


increased in both n umbers and tonnage


in the initial


study


year


(1974), South Atlantic and Gulf Coast ports handled 94,


3 containers


boxes) weighing 1,415,824 long tons (2,240 1bs.)


byl


the volume


had grown to 138,762 boxes and 2,176,236 tons.


The locations of South


Atlantic and Gulf ports handli ng export container shipment in ei their


year of the study may be seen


in Figure


Between 1974 and 1976


regi onal exp


orts


of contain ners increased


.60 percent


tonnage


increased 53.71 percent (See Tabl


These i


increases


demonstrate


the overall


vital


of the c bined port ranges.


However


as would


be expected, each range grew at different rates.

The Individual Ex ort Ran _
i v~i~u a E~xL..... jes


When the performance of each range


was considered individually, the


South Atl arnt


range was clearly dominant in both years.


In 1974,


South Atlantic ports moved 54,526 containers or 57.61 percent of the


ent ire region'


total, and 785,078 tons


or 55.45 percent of the


regi ona


total (See


able I).


By 1976, the South Atlantic range


F: :"Fi:; i-"


s _v sit. em





































cm


ci



r.
~
-1


C,
cv

4








LtJ


Ks


(7)
H-
etc



U-




S

A

0
I-
F-

CC
-j
H
CC


-n
C)
tfl





































































S -
-. ~ *


* 4


to- -S *- -.- *-. S- -


*J


jcUO

I


* *-c-~~.C


;L(1)(S


- .


I.






26


SOUTH ATLA NT IC


GULF


<1<


R ANGiTE


CON TA i NERS
RANGE NUMBER PERCENT


M- P--R -..
i N 1UMBER i rL PP \^-L_1
----^^.~--- ----" -- ||||||C E NT ^-F--~. ^-


South Atl antic 1974


Gulf 1974


630,


Total 1974


I00. 0


1, 415,


South Atlantic 1976


Gulf 1976


50 *)37


Total 1976


Source


SeDar tent


of C cm e


MAR/AD Contanerized Fare ii


C,


T ra d1e %i3t>:i


iA Ju; i *


I 5
MAR/AD


Department of Comecc


:iiar ; t20~;


vals:e


[ JL *


"."; >tr


COAST


BY POIT


>tr I


peni7


~t


t. "ui


GO~IA I ilE ii i if I!


It













per cen t and tonn


grew


72,3a2


perant,


whle:-l


t he mo vements


for the


f range


arew~


onl~


25244)


cnnl it


9233


n~r- centi


inc tt>

The Export g *n te


Only mnor c


:iisPari ties


were


:iheni r


COY r t+ ~i oerI" .


U'


ter


evaluated


oil t 1i ?


"~ C:


Ic


State of


South


r i 1 ini


a led


ii


?Cnlci : flU


while the


State


Cd >213


cflnta pr i zen


t CnniidqCji


By 1976, South


Caro


was f


both


cat eg


[)>"i irC flb


on''S


Cnn *nT a ?l 4c


ani 21. 93


percent


the reqic s:


tornnaue


See Tabl


With regard to containers


hand l


ed in 1976


Texas


rep r aced


iourisr an~a


which held the


second


position in 1974,


l1rid


i ncr as ed


pl aces in the


standing to number


three


Stat es


with the


exce


pt ion


of Alabama and North Carolina


, shipped


more


than i3


containers in


1974; and only Alabama, with


boxes


in 1976, moved


l ess


than 12,000


export containers in that year.


All states, save Alabama, handled


more than 200,000 tons of containerized cargo


n 1976.,


Alabama'


rel a-


tively poor position was further


1iustrated with negative growth


rates for both containers handled and tonnage moved


dur in


g the study


period


whil


North


lina and Florida


experienced


the greatest


pos-i


tive growth rates
"S /ft A-lW : :JT ~ T ^


The Exporting Ports


Fl ni rrrra c t


Q hw th incatinn nf


,' II 4i r #trVIIII I 1il ~ )ElllFIIi


fln nt*


tin


1
A Wil .-^ W^4^ A;^


i" ri rl 1


~v~l


ily


C i


:di


C ar o


F. r\r) E*


hRurl


1;


S


























C> C) 'K
Cj
C) C)
-
K>
tin

'V. C;
0


17; 0 2


C"
-! V2

in & 2
ZN -
t K I
(7) Ut 0 -
Ct 2> CC
4 4
a> 0 Kr >0 0
o
cm 4,>
Ct
K)
tU 4

.> (2 t
C) *
C) ~fl
21 4~
4) -~ fl~
-~ (N> C>
V1 a>
LU C?
r ~
t
(Nj (At
ii
-4 I
4 4 4 4
C)






if <2 Z



--3
*


cm
.
C
-~ -~
4- 4-
(4
~C .
K> V 4
*
A
0< 4
<2
Ct, 4.~ t.
a
H- C> C>.
~ A-t >
'C
K> -t C
C
C) -~ C C> 4 C>
< v --
,*~
:11 4' 4 t K'
>0 Ci -at C
-4 <.
-. C>
a
C -,
C>
4 a a
~n c:

0 C iz
ci
Al C) C? *
S *r it r U
4>
tu1~ 25 N
C> r > a





































FIGURE 2


4fLH l TI rA:1T TO


NUMBER OF


;1Cz>4c2<;;


~X r"r`1 p


; op1


1~1*;1'1"'"~













SOUTH


ATLANTIC


PORTS


NUMBER


OF


CONTAINERS


EXPORTED,


974


- .- S *


r S- S -- -


- *~ S .


S ~ -.


Wilmington
Wilmington


S S1111


---S"


Charleston


Savannah


Brunswick


5~*~ *
S S ~* S -


Jacksonville


CONTAINER


< 250


,000


0,000


10.00


15.000


T Evergladesf


1w.t
































F IGUP


SOUilH AiTLANTIC


C


TO<' 2


F XrT D


1Q i


-"O i 1T S













SOUTH

LONG


TONS


-ra-


- S


PORTS


EXPORTED,


1974 ..


^-. .-e Wlmi






"h^" -, Wilmi


Charleston


Savannah



Brunswick


e* *
a. .~
a -. a ~ ai


Jacksonville


TONS


,000


20,001

200.001


- 20,000

- 200,000


Port
Everglades


300.000


ATLANTIC


-r


ngton


-7


i






































F: ~


SOUTH ATLANT i C PORTS NUB kiR CF


K. :


i ,"~













SOUTH


ATLANTIC


PORTS


NUMBER


OF


CONTAINERS


EXPORTED,


--VL


Morehead
City


'4-,
a ~ -
-. a


Wilmington


-


Charleston


Savannah


Brunswick


-.a -


._._._..-.-.----'


Jacksonville


CONTAINERS


< 250
1,000
0,000


n -nf


IR (0nnn


.vergladei


1916


~*


r

































F iGUREi ai.


SOUTH A' LAMt 3C tC


f


1>3




36




SOUTH ATLANTIC PORTS
LONG TONS EXPORTED,
1976



1-.-^- '" -s e C
ff





\
a" *Ch



B Savannah

Brunswick
..-."------"-'-"'~ Jacksonville







0 0,000
2O0,O00 0 Port
n n_ Everglades










































p.,




C
t ii

a:
C
CL

U]


a:
LU
KO -
C-
LU
a;

C) C)
- ~**~ C)

LU
C)

0
w
00
a-
Zn
2r

La
F-
cc
C
a.
LU
-I

C)



















































* -


* .,


a-.


* w
S
I.~S S -. S S S S -
I


* S -S


--~'


I


ow0


-LI


I-T


























































LU
F-
C








ri^


0
FK--


S-J


F.-








0
CL

U-
?"*,D
-" 0w




















O 0
O0

-o
VI C>


*ra


.--.r
.
I


a)


r \ f i


S- I


aL1S


I
a
I
a
/
a
/
S
I
a
S
I


*--'


S











































kO










X
LA
cc



O'
COT










0

CL

L-.
.UJ
l-
cc
F-^







G
U-
LU




C-t



D 0





















































,* .C *


4-.-


r
Cr


* (U

(U,,
cat,
.4 (CC


*0
'U
* (U
'Co
Sc
I,


j-e.4- -- a ,-* J- 4 -*
i o
\ ~ '"*"----- '"0


0
C
(U
S
-5
S
3
z


N..-.


4-.


--4s .


I


-p7

















































Ct


0
LU


C)

-


Lu (A

- 0
CS H-
0
-r
C)
-J
(A
1-
cc
C


U-
-J
Zn
0






















00
0 0
- o
VI o


* -..-


. -*


S ~
I.-


* Cu,',
I
S-i (U
'go.


(UU
IC

'a


S ci
S S S -. -
See S S -. qua
I 0
aa I


S0

a,
2


'1,


* -


i,5%~
a
-S


I"
'C,
*r~ ~)
-3,
tU0.g
-J


*


I


1* ^














SOUTH


ATLANT IC


GULF


en V;; i.


KUP /9 P


STATE


PORT


TAI >4EV


Pflhs CE NT


f ynSi


i Ft:2 rE ~: r


abama


Mr trfib


. 007


Flori da


Jacksonv i e


t .970


Miami


t) 73fl


Port Everg


adles


0.01


Tampa


0.00


Georgia


Brunswick


0.00


0.00


Savannah


15,744


235.038


Loull


North Carol ina

South Carolina


New Orleans

Wilmington

Charl eston


20, 551

4,061

21,009


21.71


4.29

22.20


310,


77,361

281.149


21.94

5.46

19.86


Texas


Beaumont


0.00


0.00,


Freeport


0.00


Galveston


0, 00


~648


1.81


Houston


16. 740


7,6i9


19.93


Total


94,653


99.98


1,415,824


99,99


Source:


U.S. Department of Commerce, Maritime Administration, 1974
MARAD Containerized Foreign Trade Data, (unpublished report),


""1 ::


:~ b" Pil~ T












Fourteen


dome t i:


ports


C U (it 3 t: U


1974,


seven


from


~ac h


port


ranq il


See


t "


i i it>


Ii t


Co rinerc


varn e


frcE:;


a i qh


0 3


I


flr ICY


one unit


wh.Iil C


tonnr age


varied


"over


H r~l F


eif i


tons,


The Port of Charl


eston


At her


rprts


cont3


neLtS id


movement i


with


22s.201


rYercEnc


thuc r~eg ion


total;


now ever ~


third with 19. 06


per cent


of the


tonna: ifeI~


By 1976


there


ports


export rt


co nt ai ers


froui


the South


anti c-Gu 1


req on, thirteen fran the


and eight


frm O4


South


Atlantic (See Table 4).


-our


ports Char


ton,


ouston, New Orleans,


and Savannah exported


more


t han


20,000


cont iai ner's whfle~ iF


at the 1


r~1w


end of the spectrum the Port


contain ners


Or angel


Five ports, Charleston,


Texas ,


o~F ly


Orl means, Houston


annah1


and Wilmington


moved


more than 200,


tons,


whi I


the Port


Pan ama


City handled only nine tons.


Within


the study period,


the ports


' "1. i ngtPon


Charleston,


Houston, and


savannah expert


enced


an abso ute


oroivth?


in containers


greater than 5,000 un


ith


d to abso


growth in tonnage


handled


Charl


eston,


Wilmington, Jack


son vi~


HoustonF,


Sav annah, and


Ne Orleans all increased their hand


wu~re


h:I-an


75, 00
^ y(J


tons.


When the dat a were viewed by cacnbinin

with respect to containers in particular


the five


year s


ncipal ports,


ports


handled


45 percent of the containers


1974


and 79.14


percent




.


. .


e x js 31


:"i


: 'i ii:


L,,


i; ~3 :: s


d


61e~


regar





4;7


SOUTH


ATLAS NTI C


AND GiULF


COAST~


CONTAIlN E


S37 476, BYi


Fx0: 51


STATE


CONTA! NERS


RECENT


TONS


P" ER E:NT


Al ab a a


, 948


Jack sonv i 1 e


10,642


M iam


Panama City
Pensacola


7.67
9.73
0.00
0.31


Port Everglades


165, 969
178,743


4,573
7,755


Tamnpa


Georgi a


Loui si ana


North Carol ina


South Carol ina


Tex as


Brunswick
Savannah
Baton Rouge
Lake Charles
New Orleans
Morehead Cit
Wil i ngton
Char es ton


21,902


22, 533


12, 830


BEa uont 7


Corpus Christi


Fr report
Galveston


2, 64k3


Houston
Orange


0.02
0.06
15.78


0.00
16.24


9,25


0.06
0.00

0.03
1.90


312, 649
14,579


0.01
0.04
14.37
0.67
0o00


38 7,403


213,413
477,232


1) 071


38,588
368,362


0o00


0,01
9.,1


21.93
0.05
0.00
0.02
1.77
16.93
0.00


Total


138,762


100,,01


2,176,236


100. 00


Source:


U.S. Department of Conmerce, Maritime Adninistration, 1976
MARAD Containerized Foreian Trade Data,. unoublished report).


i" i~h" T


F~I)RT


bile












Although the comb ined


or. port


percent aye


of cot a i n s


tonnage decli i ned f


1974 to


1976


the cont i nued


comianance


tnese


ports


i rdi cated an


Si. ted


ni tial


development


cenerS


throughout


I 91Ofl~.


[U]l ;13CTI


ii: ;riSi


cor Lai


cargo


o in the


i`!


we ire


e 1rCv ced


-.s


4I

t` >ii er11


t~i I


count


ame nr


vesse


V >fl


C~C3 piej;


mfs 2112 I


j` -V


U 1W


decks


of s uch


I<~~Q~ rA <1t~ ~YUflt~' ~. -.


nfne cHi: h x


ists


I ev ; nti


S:n try


- 1 .^


tonnage


during


The South


t'
K~


(2*11 o


exr'orted


containers


Lo SS


aI i ff er en


countr


Lu tJi


st


Per" a d,


64 in


1974


in 1


country


count 1


ners 1"


ranied


Y (XII


to 3


in 1974


to 48,129 in


975.


onnage


cownit


r an e d


fromx


to 455


in the fonner year and


f ram


to 79


matter year


The Netherlan ds


,n both years


co untries i


1 ;p rcr t i l g


containers and containerized tonnage with


per cent


or more


of the


total cargo imported from the two Unitei


States


port


r-angqEs


(See


Table 5),


The top ten


sporting nat


icni: luded


Australia and


countries in Europe, Asia, and


south America.


In each year


five of


the top ten count ri


for importing from the


were

South


in Europe indicating that region


h Atlantic-Gulf region


affinity


In total, the top


ten importing countries handled


percent


the domestic ranges'


aerj


i! ani: I


~


Forts





69





TAPL[


TOP TEN


COU NTR i ES
FROM tOUTH


rE CE: i iG

1974 AND


COA P NTA I 1iE i'

1 t76


MPORTS


RANK


C UNT RY


?372


RLNK


r, tT


-~ ~


Netheri arods
West Germany
Japan
United Kingdom
Australi a
France
Honduras
Brazil
Bermuda
Italy
Total


i ct ,ert and
Wesi. C rmariv
a an
Uni cej K ingdotn
aO" A
Franr,
Australia
Hordur as
Brazi 1
Ttaly
Costa Rica
Total


RANK


COUNTRY


1976
PERCENT
OF CONTAINERS


RANK


COUNTRY


PERCENT
OF TONS


Nether 4 ands
West Gennany.
Japan
France
Italy
United Kingdqom
Australia
Hong Kong
Venezuel a
Spain
Total I


34.7
14.5
8.9
4.2
4.0

3.0
2.6
1.9
8$1.0


Net her ands
West Geniany
Japan
Italy
France
United Kindom.
Hong Kong
Australia
Spain
Venezuel a
Total


Source:


U S. Department of Ccrmerce, Maritime Administration, 1974
MARAD Containerized ForeAn Trade Data, unpublishedd report,.
... ... .............. ..... ..N'N"-"l' _W~. ..a ,u p l h d r p r


1t_. flnartment nf Clmrsp Mar itimn Admi nitratinn 1976












clear that


co( nt ai lei" :ed


c>lot


4C r


c" to^F~


sl i t


tend c i l iC e


K


I C


V


AppeJdi x


! t 4


perti t


cii


p U! I


V V: iY ;i i,


& II t n


durirna 1974


1976.


Vhe So;t


At I 3rV i c"1;


"r`t (i cr


>s K


contain ners


to 184?


Different


t1 o rciS


in ports


1,,2 .


t, [j' stu


per od(


in 1974 and


32, 821 in


i.n i97S6~


1974


orei qn


fOrt 1


and frn cr


SCon tasi ~


Cuiitlo. uCr:~if C


ton nae


by port, ranged from 3


to 44


in the


if o^


year


an roi


791,131 in the


l atter


year.


The Port


Rott erd an


the Nethe rlands,


for each year


tudf!i ed


, led a1! foreign port i


uvuport~ i V


o, n lin ers


cent ai ner ized


tonnage


with 31.5


percent


or more


the total


C ar go


exported from


the two Un


States port ranges


'S,3


Tabte 6)


. After


Rotterdani, the only foreign port wnich hand e


more


than


10~0


percent


of the South Atlan tic-Gulf


regi on"


containerized exports


was the


of Br ieerhaven,


WyesL


Senriany.


Similar to the result t


wnen


aggre-
g ,^^^.


gated by country,


exports


to the top ten


f ore i


port S


al so


included


ports in Austral


co untries


of Euro,


Sout


h America.


Once aga


the predom


r i ance


of European


markets,


especial


ly those


the North


Sea,


coirpr is


the maj


ority


imports


South


At anti c-Gulf region although growth


had occurred


in both the Asia and


Latin Anerican markets.


In total, the top ten


impor ting


ports handle ed


76.6 percent of the domestic ranges'


contt


ainers


1974 and 69.6


ii


fr- CZ:i


Ci: n C iZ:l t i ;; 8. 1


raj;~ed


f


i: c~u~ I:


F r ~ii
















FROM


SOUTH


CONTAINER iZED IMPORTS
ATLANTIC AND GULF PORT


pvi orw~


1976


RAN K


POR I T/C(UNTR'


1974
PERCENT
OF (UN VA INR
-- ** --- :


f~ANK;


PU: r>U-~R


PER CE NT
OF TON S


Rotterdam, et h.
Brernerhaven, FRPG,


a C,~?


Greenock,


Sc't.


eal ~rj~
~
*1
~ 2<
T9 kvs.
H
~ 91
Paeotg CL:]


Kobe, Jap.
Tokyo, Jap.
Le Havre, Fr.
Puerto Cortez, Hon.
Melbourne, Austri
Santos, Braz,
Sydney, Austrh
Total


Santos, r
Yokohama,
Totburne,
Total


Sctt.


irt 1 f


M(Rn.


Utz


U


ap.
Austri,


RANK PORT/COUNTRY


1976
PERCENT
- T -a- R-E


RANK


PORT/COUNTRY


PERCENT
OF TONS


Rotterdam, Neth.
Bremerhaven, F.R.G.
Tokyo, Jap.
Le Havre, Fr.
Kobe, Jap.
Hong Kong, H.K.
Livorno, It.
Greenock, Scot.
La Guaira, Ven.
Puerto Cortez, Hon,
Total


347 3
12.6
4.2
3,9
3.1
3.0
2.6
2.0
1.8
1.7
69.6


Rotterdamr, Neth,
Brenerhaven, F.R.G.
Tokyo, Jap.
Le Havre, Fr.
Livorno, It.
Hong Kong, H.K,
Kobe, Jap.
Greenock, Scot.
Puerto Cortez, Hon.
Hemburg, FR.G.
Total


Source:


U.S. )Department of Com erce, Maritime Administration, 1974
MARAD Contai ne_ rzed Foreh n Trade Oat a (unpublished report).

U.S. Department of Conmerce, Maritime Administration" 1976


i,











tonnage s showed


some


di verificat ion


)n connerce;


nowever L ~*


i1


center concept


was still


r-eachi


y apparent.


ary
- ---


Contai ner


export


data


fromir


Un ited


States


South


Atlantic


and Gulf


ports experienced


growth


peri od


from 1974 through


S9 7 &.


The data


included many large movements


handled


on conta nershi


as sell


many


sinai


ler numbers moved on convent


-ionzf


veQssels


The South Atlantic


range handled the maj ority


the regional


ot vetm en t s


sustained


a greater abso


owth


ra 5t


compared


to the


Gulf range.


With regard


to individual


doci-es:t


regi Ofl"3


ports,


Mobile,


abat~nas.


diec meIind


in botht


numbers


adl Ci


or;n a~e


of conta ner


asnipa


beginnings


oridba, s


centers


s howercci


uu Ljl


1 fltt


doit:est2


l oss


r and es


in tonnage.


denti f i ed.


As for


the im porti ng areas,


there


was


at van atcn


both


number of boxes and


tonrnage.


The Netherlands, wit


the Port


R otterdam5, was


clearly the dominant destination for imports


originating


at the two United


States


ranges.


On the


co untry


Sevel,


the top


nati ons


revea


moderate growth


indi cati


a concentration of O


des tin3-


tions for U.S.


exports, wh


on the Fo~t
port


leve


ether


than the two


ri ncipal importing centers diversify


cation


occurred.


ii however


import markets conti nued


grow


as evidence
--^ ** '-f ue r^ ced~.* *


increas i n


acce ptance


of South Atlantic


Gulf


exports,


V i oe













CHAPTER FOUR

STATISTICAL ANALYSES



I ntroducti on


Prior to the actual


statistical analy


model


ing, it became


necessary to geocode all fow data to


l i tate the


investigation.


Worldwide Geographic Location Codes (1972), published by the Genera


Services AdRinistration, was utilized for this purpose.


If a domest i; c


or foreign port


ity


was not included with


the vol ume,


then an


nal four-di g


code


was assi


gned


to that


geographic


locale.


Appendi


C and D contain


the codes used


dornest


states


port


c i ti es


foreign countries and port cities.


Thus


each


origin or destination


of a container movement


was a


allotted a uni


numeri cal


desi


nation
J


Next, the mi eage between ports handle ing containers was


computed


using the


work


U/.S.


ncl udes


Navy s edition of


nauti


D instance Between Ports


cal distances between major worldJ


965)


trading po


This


and also distances


ween


mi nor ports and


marl time


unc ti


ci nts--


ari ds


, capes and


major


ports


The determi


nat i on


distance


between


ilities


ha~nd hinq


contai ners


a~ss 4r: ~


the sh


oirtest


all-water

Also.


route

each


T oVC go


ports

trading


lioarJ


30(1 DUCt


art neri


N ~:~o


d schar


Prod uct


obtai ned


for each


~ear Oi


the study.


HFilC


ti


cv
a


straids, is7










commercial vehicles for eacn


Forsi


i~icouri ctry


Shndi i r


conta i .


obtained from The United Nations


Statistical


Iearbook


(1975


and 197


The data analyzed included all


cont&a neri zed


experts foon


South Atl anti


region with the exception


those


to Taiwan


those to cent


ly planned economies.


Dat a


3 1 W anr


were


elXc i


uded


because socioeconomai c statistics


that nation are not include


the U.N. Statistical Yearbook.

to the Council of Mutual Economri


except for Poland, were excluded


trade involves other non-freemarket


Those for communst natl


Assi stance ,


cause hi gh


economic


former l y


ropor(tions


thus


belonging


o tecon"

of their


contributing trade


bias into the


models.


in fact,


diff


frc~


other


"'C cxneccn"'


rnati ons


ince:


recei v es


most f ,avoredu nation (MFN) treatment fr om the


United


states


and, therefore,


does


pay very


h igh


tai- ff


rates.


The Gravi t/Re" session Model


Var i at i ons


the basic


gravity model can be


used


as a framework


for stu


pati al


interaction in general.


An excel


stori cal


review of the use of


gray


ty mode


can be


found


an artie


e by


Carrothers (1956, 94-102


However most


graphic


texts


, with the


exception of King (1969)


(1977), and Y


eates


1968), have


failed to address


or test


the assumptions


I near


r eqres


mode


when they are util ized


to estimate gravity


par meters.


classic


essay regard n


the assumptions


the 1


r e ressi o)R


rna t~jt


written ~in 1971


Poole


and 0'E


arrell


will


bE~ dfraw'i


U~Ofl












The classic
Y-a+-


linear regression model may be represented as


bX. + u
1


where


s the dependent variable;


X1 x2,...,Xi. k. ,


are independent variables


a and b4 are the regression coefficients, which represent


the parameters for


u is


a specific population


stocastic disturbance-termr, which may be interpreted


as resulting from the


fects of unspecified independent


variables and/or a totally random elanent


the relationship


specify


Two dist i nct


model


exi (


because of the


natr ;ire


ndeoen dent


vari abl es.


Onre model,


the fix .ed


mode 1


in depenientl


Yd~iablt


held constant experiment


at certa -i ,


iJh i I ?


random-X


rinod&


a'r~e S'2i


eci-ted?


al adm


t N9


modelJr_ i


11 be uti


ized


n t[h


LuG 7.


The research obj


Lives F


which


the rer; ession
^ : *


nde'


employed


are gt general


cli ds i f' e


into


threr*


r ioup s


compu t at


of point


estimates


v'ation of


n t;r~v~ji


431 idi l~


test


ing of hypotheses.


Although testing of hypotheses


tion, the computat ion


the major


of point est inmates


concern


i s fundament al


this

when


di s serr a-

compar i ng


in "h


deni


fix~cr


110 "; ,












The cri tical assumptions


the line ar


regression


model include


(Poole and O'Farrel 1971


145-158


Each value of X. and of Y is observed without measurement
1


error;


The relationships between Y and


each


Sindependent


variables X


4 are li near


pdar afllet ers


specific


f:i~ rt oi 11C rtnn rkroscnt;


Each


cond i t i ona


dis tri but


u h as


a mean


zer ~xi


The val


u are seri 1


i ndpeo de


var i ance


the co:ndit i


ur:n -*.


b t i o n


J S


const an


S iu r1


di st


ri but o


i ndep endont-


vIa r ['C5i r


each


xed-


model


, '~


tw: condi o o.


m u s t


icrnalin ri


These assumptions


as they


apply


to the


count a Iner


equa-L anyl?:


are ad


dressed


at er


sect] (,is


Regression


ana i vses


wer e


oerf f *r Aed


0on four


evel i


for both
(* /^^ h.''t"l


of the study


and i included both number


Contal; d


rlevs


anid tonn age.


First, estimating equations were


movements fromn the South

po i nt-to-poi nt movements


oDtai "l C


A tlntc u1


Seco ind


equ at ions


4
norm
whi


vr~e ai


ci hi


aine" i ed


ncl utded


the two


subregions


individually util


izing the


sae cri ter


hrr~


the data


Cuite


y : ij*^pe :t.


ri-a


l wrin


f(lat~


di=l, ri


"t~P"i~


c0 as s











data were parti toned i nto individual


expor ting


ports


and regres


SHY'1


anal


yses


wer e


alccoinp l


shed


nort


faci


Procedure
ununsan -Anx m adil -a w er


In the case


this


udy, all data were


derived


Frcy


secondary


sources,


Thus~


the first assumption of no measurement errors cannot


be satisfied without


reser vation.


However


rnmeasurMent


error


indeed


did exist in the dependent variables, one can


stlii


that


were


r an don


optimistic


incorporated


vi 4."


4~ ~


in the residual


error


asst i n


4
cnn.




i:rr knowni I


measurement


error


The normal


not prej


Cur? ivc~:t


iidi.


Sr~~e~r


trjtv I"


ng both


s des


eaCh


atinq


St: Inslc


lC" co:n~li


ished for


each


' ~ ~ ~ z '] "


t ormul


tI~c ~j


0 < .4


* K? I


vari ab


w ertf


tr ran sF a cm e


i nto


S t 4~lcik~


S- sc s .


) ~~ndar


'"ize


values for each


obse rvat i on


in each


,tri buti on


&L ht neui


subtraction


me anl


dividing


the result


by the standard


attn


each di str i buti on


this


regression equation becoies


manner


zero.


Y- inter :ept of


As sump tio ns


regard I


each

the patterns


disturbances were discussed on1


for the final


two series


equations since they provided the most explanation.


The last assumption, that the


independent


vanr


were linearly


independent of each other, assures that multicollinearity did not


exi st.


To assure that such a condition did not go unnoticed, Pearson


i : ~ i &


I:1Vj"?


EB S I1Tn r~


41 sir,


ht


e ; i.


1!!1~3











statements imply ied


within


i:1 :t t.i


>~YP IL1-,ts


t~c~ij nfl


highly correl


afted


independent


a!) I '3


C"


equations at


statist, ical


rrlii I ca nt?~


~~'


S:,ucK1




'>: eS


identified


, thereby


1 Si s ( r 1 i nQ


that


POSS5 IAb


s~r ad i "i s


r"4 '"s


recognized.


A stepwise regression pr ocedure was utilized


ins ad


calibrating a single regression


equation using all


independent


variables suggested.


This approach begins with a biv


ar iat e


equa I on


and proceeds by adding


one variable at a time


until


he ccinpiete


equation is finally


aa i rat ed.


Ther" F Y`efoe


order


in hn


vari abl


entered


regress~t


sequence


was not arbitrary.


Indeed,


it was dependent upon


eaicn


vtan ab


contr ibution to


tnhe expl


nation


of the unexpl ained v ri


each equat on.r


ance


remnai


one of the


in the dependent


import ant


Va ri


property


stepwi


procedure


is that vari


which are


ti col i near wi th


previously


entered var


Ik


to enter the regression


equation at a stat


cally sion


fi cant


Estimate ng equations


wer e


considered at only the 0.05 alpha


Seven


for both the


regress


on and the


coefficie n ts


the i


dependent


variable


Neverthe1


es s,


the si


o: nonr- s i Qf? if i can t


independent


vari abl es were


ev~al


uated .


was deemed


necessary


to appraise the


direction of the relationships of


the literature.


Thus, additional 'in Uor, at on,


sugg9ested


al though


S ,I ,


wouii'


Therefore,


e el -


specify ied variables


~~r,'i p i? 4 fi











.rai .i j/P c- .... A "


i ndepen dent


vari abl e--first


var ables


the number


~g rc


rgcres1


of: Co3n~t a inersc


e a ch


export ted


dependent


(NCTNX


Sees ml


number of tons


xported (NTNSX).


The i dependent


v ,ri ab


udel',?


= naut


distance


between


ports


= total coerce


in general maritime cargo


import ng


country;


RITE


= the rat


of general maritime


cargo


imported by


to that


ported by the foreign country;

= per capital energy consumption of the importing count


- gross


national


product of the importing


country;


= number of commerce a


vehicles


n importing country.


he abo v


independent variab


were


utI;ilizeFd


1 subsequent


regress ons.


The bas


mode


was expressed


Ivcj


o 1j


i RITE


. GNP


*EG;Y.


. CVH.


The B represent the relative i mpor stance of each independent variable


in explaining t


volume of containerized


export


s through the South


Atlanti c-Gul f


coast


reg-i


express


the equation


n;reiar


form~,


natural logarithms of al


var liaMl


were taken wh ich


following estimating equation:


IVCt


ogTCMI +


1 ~loR iTE


4 B logGNP


o3gESGy


r loqCVH


ogDST


I,


+ Uq


DST 1


og So










Regression Analfses South Atlantic-Gulf Coas t 1974
~-~-~-~ ~ S ~ -


The results for the


ogNCTN X


observations in 1974 were:.- ,


= -0.281ogDST + 0. 30!og
(24.91) (27.60


R= 0.32


- 0.10


19, 00


ogNTNSX


= -0,.281ogDST + 0.351icgTCM
(25.39) 38.60)


0.35


=0.12


-- 23.59


where R


the multiple


correlation


coeff i c ent,


the coeffi-


i ent of


determi nation


ad just ed


degrees


Sree dom,


aidE


ratio of expl aCned to unexpl b:.i ned


parentheses below the coeffici


var ance.


ent s


F r at


are writt


The above equations as well as


those to follow


are si


gnificant


for the


reregession


the coefficie ents


at the 0


leave


error


(tWO t


) or" be tter


The 1974 regression equations for the


entire


export


region


plained only


bl es--ten


a modest


per cent


amount of the


and twelve percent


v<3r i dance


r esecti rely,.
r :sp* t


the d en


Only


dent v


ar i a


two of


designated si x independent vari ab es,


toti~d


cammilflf1 er


Iin m:ar i-i


general


cargo


arid iauticai


dista


~ae re


Aqflli


FiCant-~


The vol ume


general


car go


hand 1


.t fore


countries


~qas c


ear ly


s t a I: i ~ tr i:;?l











that nations which received large quantities of genera


cargo usually


have adopted cont


cargo.


ainerization as


innovative method for moving such


Nautical distance, as expected, was inversely


related to both


the number of containers and tonnage.


This inverse


rel atio


nship


confirms the negative eff


ects


of overcoming the friction of distance


endemic to transport problems.


An analyst is of


residuals could lead to the development of new


explanatory vari


not mentioned within


the existing container


l i terature.


Exami nat on


the residue


revealed that the majority


of those residuals greater than 1.5 were those of the primary world


ports, i.e.


, Rotterdam, Bremerhaven, Hong Kong, and


ngapore


example.


n the other hand, over-predict


for secondary world ports


as well as those


as general


n developing


the rul


econo ies,


we., Stockholm,


fast


and Genoa, and Hamilton, Mar"icaibo, and


Fortaleza.


For the purpose of this study


deve


loping


economies


those which possessed 1


than two times the worL


aver a


per capit a


GNP in each respective year


Thus


Sit may


deduced


the mode


performs poorly with


regard to wel l-established contain ner ports which


i n addition perform the


serve


ice of


centers


and


also,


some


foreign national


orts in which


the predominant


was nor-


container


iztablie


tr affic,


lReresslon Analyse; :South Atlantic-Gulf Coast, 1976


The es timrnat


equations


488 observations


in 1976


were


t v" a il~i?













ogNTNSX


- 0.33


= -0. 17ogDST + 0.321ogGNP + O.i8togEGY
(14.31) (12.99) (12.31)


= 0.10


- i.201oaCVH
(5.315


=- 14.45


Four variables


a Dos sib i


entered


each


of the


orecedi ng


equat i ons


a stat


sti ca tly


icant


e eve


They


luded nautical


Distance


gross


n atrin a ~ i:


product, energy


cons


umption per capital, and


number of c ol7 cial e i n


cies.


cxle v er


exp~l


an ati


(W2) w


Co~rk


sidered


to be poor


N auti


di stance


contri


buted


the most t


toward


xpl anti


because


05'19


min :'


cont ai ner


routes between the


U.S. and Europe and


the U.S.


apanr


were


n place.


Dist


anices


between these


n 3ti ons


are gene,


Sithi


the mid-r


an rge


of containerized movements (5,0


to 1


es.


SO by


corv


tainerized


trade


was estab,


*1he


pri~rnar


betwcer~


ncaist


rial


ized


well


-deve


hoped


economic es;


therefor,-


the positi


the coeff i-


clients


LII.h


and EGY


were


exnpec ted~


as


- i! e


to enter


fi cant


lvi v


nu r fT "


cc~rineru


v~hicle


? 7VE rtver .c


rel at


onship.


:ccurred


one r more


'ri


V^ :0 "I


The smaln


ati ve


rjee qriCap ic


siz


Ct: U


hand Ii


amoun LS


,r nwt


SIt


nl* ?>


201


smaj aI ~1


Fleets


7I and jhase


conMTrc


vehiccs,


Sr i~de


wit h


foreP


oad center, s


Tn both


Ir I "3


and e


contributed


to under


on of /i


motor


trucking


favor


rnri~tn t


coastal


in land


WatetF:"rw:3Ys.3


C


p ,ir~n~r-:r


~:.i~


E as t


't iz~?Ck











hne vy


use of


unit trains


especial iv


in co ntinental


Europe captured


potent


count ai rer i zed


motor


truck


movenents.


Al though international


ma r itime


count a i ner i zed


mo ~v emtlent s


1arge, many contain


were


still being


stv ipped


(enpt~i


and stuffed (fi lled


within the con fines of


port


f-ac f l


U es


wor d-~wi le thus el i mi


A cursory examination of the


nati ng


or-to-door


residua


showed


1i p;Wuenr


f-n i


residuals


greater


appeared


devel op


It :os


and secondary


ports of deve oped ec)nomin


1 es


frj jcP


c ainr


i: F}i~ tnooec i


)er orm CedC


poori


estabi ishled


co ta ns;,,1r


P> cI';


:4lr


handling fei


cont a i ers.


A Comparison of the South


Ai' nt ic-uun


~r<'~ 3m>


> L;
I flit;


ree xami


n iati or


>tecire.ts (0?


t5325


ye "


over


ten o e : cen


e17 pi


anat i on


was 3 iii


4S;i:


nd i vidua


trro 44% 1


COs S


devcn2


to be


rob~~st


S1 rCe


a m inor


amount of


anti on


was acconp


Ij n;s


i important changes and


genera


Si t i es


(C irc r1v:ed


the systeln


dui

the t threee


'I'>"-


nerlo


In 1974, nautica


distance and maritime


commq er ce


we re


stati


stically


si gnif cant, but by 1


in addi tion


to distance,


gross


national


product, energy consumption per capital, and number of commercial


vehicles contributed to explanation.


The equations


the earlier


i ;


fli~


,' 1


> I


Ni~ r


Qr" ? I


~


Clz r G


Li Cj(r;C::?.jr


r;h3 it











system ref 1 ect i ng


i I ncreas


cOi flt, V


Serjizd


t r Dd(


hetw


ndustri alized nat


i~fthe wnr id


When


final


the es l


step


Cf1U<~t


hr Pstepwil s


DC L ct


I~~~ ti~03 3


Yt


(5<30**


r,671~


q~~!: j ~I


I 1


dependent


240

werT


sta st >


14 i: CanL


6 6e


alpha 1ev


H~ 0v'J : sLve


most


Cl flfl


Table 7)


excep


ton


Was h


th~ii rt


ports of


giener


alca


go (R


whicGh


erfilt eredc


the equation


a nega-


t i ve coefficie nt


1974w


reve.rsa


the expec.


gn further


supports the prl demise that


tem in


1974


ws expert enci


an i


nif a!t


growth stage because


lanced flow was


requl hired


to maintain


continued growth.

Regression Anal ses South Atlantic, 1974
R ^^ t r J *S!\ i c!1I i\"*3 nJ^-14<'Tt ^ < r d d'e


Follow ng the strategy


pr es cr" i bed


ear ier in this


pter


flint.
I


sion analy


were


next perf formed on the data which


were


d sa gregated


into two


ubreg ions


The function of this procedure


was to ascert ain


whether major differences, endenmi


to each subregion,


were


pr eval ent


If indeed regional differences in


radTing partners


flows existed ,


then khopeful


higher
h S p1


evel


Sanat;on


achi eved


by dis-


aggregation.


The results


the 145 observations in the South Atlantic region


for 1974 were


1 oaN CTNX


S-.37o1gDST + 0.251aoTCM


(15.51)


(6.77)


115nc;


)fl~~r


I,


i in >i





6"5





L: L "JI


COEFTF I" XI&\'S


SOUTH


ArlNFCU


ESTMATNC


UiA FIGrlfA


1974


: 7.


NJ CT N X"


NTN $X


N CT N X


NTNSX


TQM

RITE

GNP

EGY

CVH


Source:

Note:


Author's individual

Asterisk indicates c


regression


c effI client


an a i ys es.


statistical y significant


the 0.05 alpha level (two-tailed)














ogNTNSX


S- 0.4 ogiDST
(19.61)


r
4(II


oc" C


- 0.36


= 0.12


The 1974


regression


equations


the South,


uAt ? anz~ici


ranqe


pl gained nearly


the same amount of variance as did the


xis for


the entire domestic region


of 0.09 and 0.12 respect vely).


addition, the same two


V anl


nautical distance and


total


C onuneree


in general cargo, entered at statistically significant levels.


Nautical distance, which had a higher F-rat


was C


early the most


important explanatory vari ab


Extr ne under-predict ion, indicated by residuals


greater


than or


equal to 1.5, occurred


major trading centers inc lding Rotterdam,


Tokyo, Melbourne, arid $vemerhaven, whi


fiSO


fun cti


oned


as entre-


pots along with the ports of Hong


K~ong


and Kobe.


The Port of Greenock


was an exception to the previously mentioned rule; it expert enced


under-predi action due


to the return of conta3 ners


the eventual out-
get ; ,?ri, *Y i- | ?*%i I


bound shipment of spi


ti ts


the most


part


ports


in developing


economies experienced over-prediction with


duals


less


than or


equal to --1.5.


Typical ports in this category were Manta, Valencia,


Paranagua, and


anto Domi ngo.


In adi tion, relatively new ports, e.g.i
In add ,^K JE on,*r!w1^ '^stL'fY-^j 1'T I^J ^i -


Fos sur Mer, in France, and some of the older established ports, such


as Hamburg and Adelaide, also showed high nega tive residuals.


Once


ex-


20


ecyu :-; tj


-






67




Regression Analyses South Atlantic, 1976


The estimating equations for the 226


observat ons


1976 were


IogNCTNX


= -0.351ogDST + 0.381ogTCM
(25.98) (30.72)


R = 0.32



logNTNSX


= C, l


RL = 15



= -0.35ogDST


=- 0.16


- 19.75


+ 0.411ogTCM
(36.22)


= 22.12


same


two vari abi es, nautical


distance


and total volume


general cargo, entered the equation at statistical


signf i cant


levels.


However


, in both equations total commerce replaced distance


as the primary


expi anat


vari ab


with F-rat ios of


30.72


36.22


respect vely.


Although


he model performed better for this later year,


exp gained var iance


was still r elat


y low


both number of contain-


erns (I2


- 0.15


and number of tons (


= 0.16).


An exami nation


the residual


revealed


that


under-


nrcdj


ction,


indicated by


resi


uals greater than or equal


to 1


fell t


category es.


The first was major world ports which al


fun cti owned as


transit facilities, i.e. Rotterdan, Hong Kon


Br emerhaven.


second catego


waS an aqnoma


cons


ted of developing


nat ion


ports that


rece


ived


larger


than


expected


amounts


of traf


due to the


construct i on


a conta


mencr


such


as Manta,


i Ecuador


D aman,


S auii


Arabhi


ver-peed


rt ion


th~ extre m


~~pe w












facilities was capturing genera


cargo which


uS ua


lly flowed through


convent onal ports,

A Comparison of the South Atlantic Re ion,, 1974 and 1976


A comparison of


regress i on


equations


over


the three


year


time


period showed that in


the


atter year the model


beccne


more robust


since exp


increased sl i ght y


Neverthe


esgs


the 'evel


expl nation


left


much


desired


It was


interesting to note that the


Sa~e


two var iables


(OST and


TCM) added tn explanation at states


tica


anificant


evel s


in both


years of the


study.


In addition, these


arn~e


van"


wer e


only independent vari.ab


which


st a iisti a -


i cane in


the equations


the entire


South


antic-Gulf


coast_


1974


The above


subst


anti ated that


won 1 `


c Int


system


was in


a dev el


opmlent a


stage


si nce


;ame


yarn ab]


CO i 1 stent


in both


recgi


ona1


in fact,


model 4


coniine r e


Srigit at i


i3r ts


of the South Ati


range


t (


ribut




towI: d


an a-
-ft If


tion,


HiOwever ir,


976 bec


ause


chan ~i


"in e Atory


for the larger


doim es t


r"eg i in,


can be


dednced


St /te?. ~>ri


o h an< es


were


occur ng.


Esti


mating


equat i ons,


wh ich


prC; i 'ti


Inn ??iI


u.X V


vari abl


were


created


Wi V7


I r* Cjt (>


were not


stat st


ica 1ly


si nifica-


at the 0.


pl 1


ire


Table 8


e .,-,


PTndeSnt


vari ab


1 es

and NTNISX


a tw~


were


1 Coef i i ens


1


.


I







69






TAESLE


S GNS


GO E FFr T I (

P AN G E6


L5IIMA
I 1974


i HI


t


MnL 1976


1974


1976


NCTNY


NiNSX


N'>NX
~ -


RI TE


Source:


Author's individual regression analyses.


Note:


Asterisks indicate


coeffi c ient


statistical s gno ificant at


the 0.05 alpha level (two-tailed)


1[TN


PTL4NTi












containers in 1.974.


The bal


&nced


a i a 31)1?


p !ii


rmat ye


associated in all


regress ons


support ing


of the to w ay


traffic requirement.


most


Cd~s 9


cienci


either


inbo~jun


outbound container cargoes


could d


otenti al


ance


domes t


or foreign co


astwise moanents.


of Gr


onria I


odu~clct


was inverse


the earl


year


nt\~


in the lat


ter.


This


shift stated that in the fi


year of the


study economic affluence


was not important in


determ


cnnt


a "n i zed


owsJ


howi~ie ve r


in the


Scatter year it was..


sign


per capital being positive refl


coeffic ient


eced e i


iport an ce


co nsarmVp on cr


ndu~trs


i zed economic es


eer* n


or receiving


C 2
Jid. ai nerl 230


negative si g.


i" the f


flmZEry j J


tans


41 19~


oiiiy


nyu i.c; i


iC YEii~


bIte


.A! A>Y4in ,


! I fl&I 1


numnoer of


the st


trucking


co0 i:* rC"


per i ocd


L~~re 0et


veh i ::wc;


: cat flu 1


$i ~~r C""i 2


V i


>~i ~


cy xre


V' i .


Ta> !


s erv i ce


at foEI "an


oesti n at


ressi"on Anal ses Gulf, 9
.. .. .. ....


estimate ng equate ons


the 190


observat on ;


1974


w ere


'lo gNOYNT


S.341 ogGNP
(23.87)


= 0.34


= 0.16


- S.1


C' ara1


>1iv,


C.


K) 5i:*i


"K


: 1


17


cner Qy


I


I- ,,


:i:


i "


r






71




Gross National Product was the only variable to enter the estimating

equation for number of containers exported (NCTNX) at a statistically


significant level and it explained


even percent of the


var i arlce.


Three dependent variables, total commerce, Gross National Product, and


commercial vehi


es were


statistical ly


gnificant and expl ained


sixteen percent of the variance in containerized tonnage (NTNSX)


the second equation, GNP cl


early contributed


most


toward


expl anat i on,


followed by total conmner


performance of the model


commercial veh


the l atter


si tuat


ci es.


better


the


first


the study,


oth withI regard to


exp I ar ati on


supporLt:i :


variable es


ind i cated that


toan age


was pro


ba~y


a better


ndicCat or


2>c


was absolute number, s of conta


ners.


Number


C: nnt~l n nr< r


does


account f or varations in the phys ca


size


bot7es --?()


eet to 60


feet in length, while tonnage inc


orpo ates


su ich


v ~r~yn' ic


dimension ns


In order to obtain additional informant


with


re*


d to model


performance, resi dual


were examined noting their geogr


aph c


ori enta-


tion


Under-prediction, in dictated


residuals great


than


or equal


to i,5% appeared for three groups of countries..


first;


group was


internal ional load centers


, i.e., Rotterdam, Brererhaven, and H


Konng.


The second consisted of


the two maj


Aust;ralj an


ports, Syd ney


and Melbourne, which both


OS nadonal


centers


Ports in


developing economies which had


coc tl ij ;rucet


co]nt, ai ner


UtoHi


,ties


Santos Tuxpan.


di the port


area


on: the


i skm ndi


A ru bha


F nrq~


, i.e


.


9


* *| I












Regression Analyses Gulf, 1976


Two hundred sixty-tw


observat ions


contr i buted


to the


FoE owcing


estimating equations for 1976:


oqNCTN


= 0. 16


= 0.l16ogEG


= 01.03


I ogNTNSX


2S8 oqGN4iP


+ O.22Zq:u


K> 32


- 0.27


fl.tt


Only one


v ar~ al)ii


en -1


C


Sti on


:3 c zt I tC< y


gnificant


1 fev 4


wh ?ile


thr ee


var iab


5 [67


0/H)


WereF


nifica nt in the


second


equation.


In t h


>U0tt flP:!IUr~Llt21{


contained irs


equation, energy


cons umpt


ion per


ai ned


on I


3 nmeacer


amountit


of vari ance.


However


it was


fi rst


ancTe~


that e


:~. e


cons Ui p-


tionf alone contributed to explanation.


tinr al l


close


economic


ties between


Pers


an Gulf


nations


and the U.S. Gu


region


result ng from


inr;du str"Y


ers between the


intern s


areas.


t i i


encour aged


Koc;eiqn


mrnov e ien t s


moies r r


C i


Wr,1ve


to si zabi


i nboun o


~novenients


man uf act ,r d


>20s flS uWiirri


Stains .


Though exp anti on


W ~ I ~'vihdt. I.) CC t Q:~


sti i left much


to be


les r i~e:


C:


C (ins


(21 t t


deav


was the greatest contri butor


t ei l na


- 7


)


1K


"It'in


; i




r.


in~t


li


i







73





By examining the geographic distribution of very high or very low


residuals, a better understanding


weaknesses


in the model became


possible.


Severe


under-prediction


was evident


i nt ernat i


anal


l oad


centers including


R ot t er d a


Brer ~errhave nr


14"

over-predi c-


tion res ulted


for the modern


deve o ring


nation s


prYf ~t K


of Pi ueto


Cortez,


Santos


?> arie- ro(


as well


as the


p~rI2S


of Me lbourne


and Sydney


in Austral


th twin


port


ofL no


Nigeria.


oped economies


ver-p edi cti on


such


once


as ttockhol L


gain typ


0


in older


and Cato" Cow


deve V

Thus,


the model


does


Uio pe rfTo m


nebli


for f


Ih i


a "1 C~


i:i the


ver"`Y


or very sma]Ll

A Cariparison of the Gulf eon, 1974 and 1976


Unlike the regression equations for the


the three year period, equations


south


theG


Atlanti c


regi on


region


te;ss


over


robuLS t


s i nce expl


anat


on decreased.


fact,


an~iat i


< Jeach


l owes


point cmp


ared


to all previous models.


The period for the Gulf range was


one of


tr ansition.


earlier equations,


G ross


National


Product


was the maor


0 ~Jtar~a L10flf


variable,


whfl


in the


latter


year


Ih .t ^


tOOik


p~s~ct


nn~c in2r` g,~


consumpti


per capi~a


sec


7Th it;


: ii


4p Pt C(~,~


was cause


it' \ ,


a~i n>?r


iizeci


xoor ts


produce rng


n Ofi:lS~


Sn t h


ci mY mflfrs K n t ZY'Ht1h


K, ;


Pe 4>


lif


Dtlii7


; ~rqi:


: ii r' t~ ,~s


C~$I" I~; j (,


'i iE?~


;y;in i ;;sl Z 1"1~


rS1 ;r ir5 I r PI"


. :


r,






74






TABLE 9


SI GNSf


COEFFICIENTS


EST I MAT ING EQ1UAT I ONS


GL LF RANGE, 1974 AND 1976


1974
Jt if l"


1976


NCTNX


NTNSX


NCTNX
*-:-* ---llll:l--m--


NITNISX


TGM

RITE

GNP

EGY

CVH


urc Ce:


Author's


ndi-: id a da


regression


OrI:: I >r; es;


Note:


A.t er I sks
the ,0.05


ndi ca ri


eve


Co erfr


ic(3i en!i:


statIst :c~ lv


C~~~ ts-t


simnihicant












number and


of contain ners


sh pped


ws not si ini icant


change in the sign of total


commerce


ge!iler a


car go


Fromset


pot t


negative reinfor


idea that


to ri j)


p ro ci cnir


4 S


was increasing


al though


nternat:i ona


trade


was Qroving r


a decl ning


rate.,


However


, the posi tive si


ratio


i npo r


ts to exports


on the other hand supports


bal ancedr trade


air qurent


6r~e


N national


Product was statistically sign


ficant


in all equ


a tjnnci


i ndti cabny


that


containerized trade occurred primary ily amon


wel -


devel aped


economy es


and also oil


producing


eco;norni


Energy


consumpt


per capi


bet ing


positive supp orts ~


pran i se


of container


I.


d~orn fn~lh:t


U rd :jt~t I


alized and


rodu ;


nq not i oilS *r


4ii~


the oe a:


~~cjn oF


coierci a vehi


v i (:n


C: 4i


once


agal fl


cat


coulp


movement s


world


-wD &


Gray i ti Re res ;" e"


Anal s1


A
Muir e~a cet.


r r~


i ntroducti on


Fur t her


yses


performed


O:n thr?:


statci t


( Th -?


greatly improve expl


Sa nat i o n


I heref oreC


resull V


>ii~


Sha


n~ot bacr,


reported


Because


4f the


overall


poor


resul


var once a c


reevaluation of the


numerous dyads


previous


or small


abso


methods


was undertaken.


dependent


aren't iy,


var iabl


hindering the model's ability to achieve higher orders


ex iana on,


o remedy this situation, data


were


aggreg


ated


on the basi


s of domestic


c: [


flj


-~


<~ I


14 :


exFi:, r "i, s


V;ir" i


t


I


i I iii;


Il?:ir~


c











involving shipments frc rn


drn, e~; ti


p irts


vCri u U


3:


so recoriput


tLe distance


var i


() 1; t ~t) ~


C 3


I~ a t ed


observing t


a esu
> A < < ,^ i
a : : ..


fl ow


a r domestic


port


ges t


count erpar


in a


speci


- "


country


then


ass gsning that


distance to the toa flow between


tin~es1 ti c


or gin


90 country.


The rat ionale for the prior ment loned


meti odo


cuseFd


on an


attertpt to be more port


spec


i ic than


former


st udi


Marti (1980

an analysis


anrd Yeates (1969

n keeping with t


However, the i ntegr i

original objective wa


prov hiding


s parti al


conserved

by aggregating the data only with respect to foreign national destina-


tion.


The following regressions involve a


mu "~chl


small


data


set,


addition, because their coefficients of deter mination


rmos t


part


i ndi cated better


its, the usual


assumpt t n oS


of regression analysis


were addressed for subsequent selections on ly.

Number of Containers South Atlantic-Gulf Coast, 174


The domestic port/forei gn country


gr av i ty


model's results for 1974


were found as


i ogN CTN X


- 4.27


- 0881ogDST + 0.


Io gTCMv


(0. 1990r)


(0,09l


= 0.2155


=27.


- 2 1358


~ 1Q6


writere ~


the coefficient of deter nation


freedorn, F


us~t ed


the ratio of mean square explained t


for degrees


mea~ n


square unex-


nli iinprg


ic thp standard prrnr nf pctimatp and n


the' niirnh0r nE


n r


(3 i3v:


be t:lrj eci;


1










The equation f ts the dat a


bet ter


that


equation


which


included port-t


o-port


mov nentsii


as i ndic ated


n 15/


which almost


doubled.


same two independent


variables,


and TCM,


cont


i buiteFd


to explanation at statistically significant


di saggregated equat on.


lev&l


Therefore, one can safely


in the


state


prior


that with the


exception of increased explanation, very little change resulted fr iom
aggregating the data.


The Pearson correlation coefficient for the two


vari abi


independent


entering the equation was 0.38043 with a significance of


0.0001.


Therefore, the presence


mul t i col i near i ty was not severe.


Testing for spatial autocorrelation is a camp


task,


Autocorrelati of


generally is of interest to the geographer when attenpting to define


new explanatory vai abl es.


In fact, it is believed that sizeab1


amounts of autocorrel nation exist among the data given their nature


The contal ner data account for actual port to country


f ows


, but do


not guarantee that the ultimate destination of container f1 ows is


within the same country of entry


i nst.ance, arge amounts of the


container movements to the Port of Rotterdam eventual y move to
nations other than the Netherlands either by overland transport or


marine transport.


Thus, the concept of


a load center does not


necessarily only apply to dilestic final destinations.


Ports, such as


Rotterdam, which handle


i large


amounts of cargo to be transshipped to


other foreign ports are sometimes kncwn


as "pivot


ports.











tests were applied to


results of


te gsavi ty/'egress oi


mode


this study.


The Isual


procedure


'i, nvjcl


group ng


" he J a3

independent vari


able


5 ie n


perf


orm~i


separate


reg re s on


ys es;


on both the lower and upper groups of


observations.


squares from the two regressions


are estimates


disturb


ance


variances, and the test is based upon the ratio


Sthe two mean


squares.


The nu!


hypothesis is that the two d i sturbance variances


are the same.


the c~nmuted F exceeds the tabulated F, heterosce-


dasticity is


to exist, and


must


corrected


a transforma-


ti on.


Following the above procedure, t


equ t on


for 1974


as well


as all re ai ning equations was


tested


heteros


ce dat ity.


After grouping the d


on tnhe


maSS


var'i able


T CM)~


ten ce: ntral


observati ons


were


omitted


rqrgess ion


was run on the two


equally


sized


sanp


The result ts


were


Sower sampl


ogNCTN X


- 7.65


- 0.871 ogOST
(0.1990)


+ O.29IoqgTCM
0O0982)


--. 0.126i8


2-7,


, ,t


- 93,


upper sanp


ogN CTNX


- 8.12


- 0. 371


ogDST +


(0. 5158)


1.4051


ogT CM
4)


= 0.1341


=9


=93,


The residual sum


areas


from the


l power


sample~


n-3)S


372.9173 and the residual sum of


squares


2
(n-3S ,was 448.6017 so that the computed


Frcxn


waS i


3 2


sannle,


= 112030


whlrich-


was less than the


table


F of


1.4000.


Thereft ore,


)7 i>, CS:>'> r Uia -ici


8,











assess


whet her


the distance


vai able


toot bIJ Li i


:I) fl> erf:Is ce -


dasticity, a


si l lar


test


was run after


s ort r


the data


on the loga-


rithm of the


va;r iabl e


*11 tin


ten central


ohs er vati


The results


we re


ower sampi


ogNCTNX


= 5.71


0. 981 ogDST
(0.4900)


+ O.62>ogTCM
(0.2029)


-- 0.0765


= 4.81


- 2.3212


upper samnpi


logNTNSX


= 4.02


- 0.411ogDST + 0.760 ogTCM
(0.6547) (0.1247)


= 0, 2806


- 18.94


- 1.8443


= 934


The residual


sum of


squares from


the lowe sample, (n-3


484.9066 and the residual


sum of squaress from the upper


n-3~2, wa 3
(n-3)52, was 306


. 360


so that the comnputed


2


= L.5840.


Since


the computed


ex needed


the tabulated


a transfoinnati on


was needed


to correct for


heteroscedasti c


Thus, the


regr session


equ at ion, with


adj ustment


he t eros cedas ti -


city becomes


ogDSTi ogNCTNX


=- 14.291ogDST


og2DST + 0. 75IogDSTiogTCM
1264) (o00948)


- 0.2382


- 31.


- 17.


6472


=- 196.


With the logarithmic transformation t


adjusted


coefficient of


deter-


mination i

of estimate


increased


decreased


lightly over two percent and


ndi caring


standard errors


a better fit,


In a


simil ar


manner


, the data were


sorted


on the


ensit


- 93;


e











owner sample:


!ogDST looN


CTNX


= 23.231ogDST


- I. 651og2DST


? J475i r, i D U?-I~cUQTC M:


-- 0.1576


*~ # ij


= 17.8713


~934


upper


sample:


logDSTlogNCTNX


- 17.451goDST


- 0.238ogDST
(O.3302)


+ 0.771ogDSTl
(0.1244)


ogTCM


=- CL 285;6


-- 19.39


~ 93,


IC,


The residual sum


squares fro n


the l


power


sanw11


I


'W s


2874


4. 8034 and t


2
(nl-3)S2,


was 25


residual


042.1360


sum of


so that


squares


the canput


from the


was S


upper
SS2


saznple,


=-1 1487.


canputed F was 1

attributed to di


ess


than the


stance


has been


tabul ated


co r rec ed


t therefore,


for.


,eteros edasticity


Thus, correct ions


heteroscedasticity added an additional


two percent


t093r i


an ati on.


Number_ of Tons South Atlantic-Gu f Coast, 1974
- W I- -' .^ -i U ( >**Iw 1- I L j ^' in i~ y n r fi it T i T i -l T -- T I-^


The results of


the domestic port/foreign count tI ry


gravity


rmode i


1974 were found as fo1


1 ov s :


1 ogNTNSX


- 6.25


- 0 .95 ogDST
(0.2142)


+ O. 801gTCM
(o. 1057)


- 0.2340


- 30. 79


' 2. 298 7


= 196.


Once again


even


after


aggregating


the data,


s ~fne


two dependent


var i abl es


cant levels.

Multi colIi ne


DST)


Adjusted


arity was m


entered


R-square al so

inimal since


equat ion


increased


J

C) e


calrrPel


a even

ation C


perc


teiff


I~:nt


far~~~~r th0 tnntritn4 KC1 nI, P I!'i


x\ni~ ", nh l c~ i


t2 h1r~


signi f -


n-3)


.^*a ^


i FW


F nr


antkr"r fin


I











equation as well as in all further equations


was apparent


y present


for reasons previously ment ioned,


When


assess i


ng for heteroscedasticity caused by


mass


var i able


(TCM), the results for the two groups sorted by the log of the mass


variable, leaving out the ten central observations were


as fo lows:


lower snmpl


1ogNTNSX


-- 10.66


- i.84logDST + 0.17logTCM
(0.2490) (0.2409)


= 0.0982


= 6.01


= 2.2734


=93;


upper sarnpl


loQNTNSX


- 4.67


- 0.551SogDST + 0.17logTCM


(0,5158)


(0.3474)


= 0.1389


= 8.42


.2919


=93,


The residual sum of squares frcm the lower sample, (n-3)


465.1711 and the residual


sum of


squares from the upper s rmpl1e,


n-3)S, was 472.7457 resulting


n c omputed F


2
, (S/Sl2


~of


0163.


The computed F was less than the F-tabe of 1.400; therefore, heteros-

cedasticity was not caused by the total commerce variable.


Testing in the same manner af


sorting on logDST,


the following


results occurred:


lower sample


logNTNSX


- 4.02


- 0.04 1iogDST + 761ogTCM
(0.5252)


= 0.0758


Ann,,
-
-


-2 92'


upper saimpl


ogNTNSX


- Eli1


-0,


2410ogDST
7057)


+ 0.891og
0.1344)


- 0.3148


- 22,13


- 1


9879


=93,


The residual


sum of squares from


t he l power


,2


; ample












the F-table of 1.4000.


correct for heterosrcedas


e eF ore,


city
i C i*i


a trar 3;sv QV!'hfjI t O~n


Caused. by


, < nsace


t, e s. ,tan:


After mul ti n Iyn the quantities on


hr:".nt


s;i :i


?C jddt i r


ogDST, the result


reoress Ci O


4"ua~i;


K> ~


oqIST 1


0o0NTNSX


= 2t 31ogDST


f n ? ;r
1tu, 1355>I^


4 n/


I '
^ JL >^!00 ,.
]i.U-- ^


- 0, 264 5


As a result of the


= 49279


tran sf nn ati on


i8 169122


I ncreas


appropri ate stani dar d


errors


o: es t


iriate


< *decreased


better


overall fit.


A check for the resece


,of heteroscedasticity resulted in the


following regression equations


lower


sampi Ce:


logDSTl


oqNTNX. X


-
t. 9, )


711i


52:1


C~~iI I


0.2<39?s


U IPpar


S lp


iooQS7~1ojN9


17


t" T t<


fl ~V~ ft


U,


rj~j > ~
i


0. 2 iS


resi? du i sum


S, iU are Is


fromr


ni t I O e


Sdv1hi) t


3i65. 5 0324 and


the residual


sum of


squ1iiar es


fr om


upper


s _anpl,


(3S was 29016.1032.
't~~~~~~~ liMMJ./l i W i3 .<- LW wX W


tabu l at ed


The computed F of 1253 w as


of 1.,4000, therefore,


than71


heteroscedastcity attributable


fran thr~ inf1wig~ncp nf th~ di~fnrr~ ~n4riAb1P


waS cofrzct.o1i fnrt


2:9G


U. /i
"7 .j-


"~?


2.


r~Q


i. 9


51 .3 ~











Number of Conta ners: South tl ant ic- uf -Coa ,


The reyr

apsed to in


esss; art


%results

oreigtn


Sor cort a l


import





Ierl


K. :


we~; i e*


Sgi-
oI!m


ogNCTNX


0^f *] '*. T


S0.1538


3~1:<) 1


<0 r. 15<3 i O


- 14,


-300.


A moderate in


G' *\ *A **-ie


explanation


was acc pl i shed


dwrreya3.t'irc


data to


the fore


ign country


l eve


Four


independent


var i abl es


ent er ed


the equation at


stat


ically sign


f icant


levels


Scss


than


or equa


0.05 alpha.


The two highest F-v


the coefficients wer


for the distance variable and 9.56 for the


gross


national product


var i able.


Pearson correlation coefficients or the variab1


en tearing the


equation are recorded below:


1aoDST


ThgTC


1.00000
0.0000

0.16389
0.0044

0.16014
0.0054

0.13595
0.0185


lo-GNP


0.16389
0.0044

1.00000
O.00O00

0.84041
0.0001

0.80175
0.0001


O0 16014
0.0054

0.o4041
0.0001

.00000


0.86217
0o.0001


1oa[VH

0.13595
0.0185

0.80175
0.0001

0.862i7
0.0001

1.C0000
0.0000


The upper number


the correlation coefficient, while


the si gnif icance.


lt should be noted that


the three


mIass


var i abl es


t he i 1


ii""r-


`i iide


i sSNT"


; 2.2(j


1 o~


-












correlated


or n


since the


F


obj ec


tiA ye


l literature


could


eJxfA`


v, al:" aicer ;


I
it: (~OVU .1


2 9


n ear ier par a


graphs,


s pat i ai


correi at ion


the nature


dat a.


were


run to dscCCL3In


whet er


i {t1. UrO Ju>iL^3' "- < ..


wi; caused


by the two


Sid e penenRT Et


practice, all


vari abi es


havi


the heteros ce das t


y Iuti zing sta tistical


trans fo ati


the I
rrni I-


by i i


9 in -:st


be remove ed


1nwev er


cuarit


r oop e




ttfl


paper allowed for


miax i


mum of t-esting for


het p~r o


cedar, st ic


in ont


the two dependent variab1

Sorting on the logarithm


cortr ibut

distance


the most

eavrnq


tow &~Ir d


4-4


expl anati


centr&I


observati ons,


the results


for the darmesti


port!


f oreign


country


mode7


for 1976 were


power


samp


ogNCTNX


-- 0.12


881iogDST +
4630)


.49kgy


50 ogGNP
.241))


- 0.371 gCVH
(0.2635)


0932


3538


- 140;


upper sampl


ogNCTNX


zt 3, 33


761o


gDST


6876)


.081 lgTrCM
0.2185)


+ 0.5
(0.


61 gGNP
2431)


161 lgCVH
0.2085)


= 0.0801


.050


- 140.


The residual sum


squares fr2n the lower sample, (n5)S, was


747.4390 and the


rYsidual


sum of s quares


F ~orn


upper


samphl t


ne~i


4 A


r s i. t-


*w'1 (. o


:I, i)


W17hu


V,


W~<*4


i


1*











The computed


was s ight q y


than


the i


tzt;: <


L.JJUij


th it


heteros ced i ci ty


^cfSult i ngi


fi


LI1 rir <: L ni


var iab ~I C


probl em.


The test


same


" regressi on


sorted


on t)I e


1 c;a thin


gave the following results

l1wer sample:


l ogNCTN X


=-8.92


- O0.881ogDST + O.06o1gTCM
(0.2081) (0.1833)


+ O. 5i I gGNP
(0.2217)


- 0. 171 ogCVH
(0,1918)


-. E.li0t


=-- 5,30


- 1.8891


upper sample:


'I 0gM CTN X


- I 1.17


- 0.271ogDST
(0.4425)


- 0.44logTCM


S1.itOlogGNP


- 0.5610CVH
(0.2506)


= 0,.1224


- 5.85


='2.4J735


=-140.


The residual sum o squares fro the lower simple, i(n-5)S1, was

481.7743 and the residual sum of squares from the upper sample,


2
(n-5)S2, was 825


-table of 1.


oscedasti city ca
The appropr


9978.,


3305


Since the computed F of 1.7144 (


transformation was needed to


2


correct


exceeded


for heter-


:used by the vari able logGNP.


transfom action was to multiply both sides of the


regression equation by the reciprocal of


the 'log of


GNP.


resulting


equation yielded


1 ogNCTNX/1 ogGNP


- 0.36/logGNP


- 1.101 ogDST/i 1 ogGNP + 0. 241 oT CM/ logGNP
(0.1708) (0.1401)


+ 6.3 Io gNP gGNP
(2.0714)


- O m Li iI~r GIf
- 0O34logCVi/IogGNP
(0.1374)


no t:


-4


r










With the reciprocal RT2


increased


sightly
q a


er' r ors


estimate decreased.


Grouping on the lo of


and orriting twenty


central


observations yielded the following two regressio r


c ii at i onrS


necessary


to assess whether heteroscedas ticity had been corrected for

lower sample:


1 ogN CNTX/ 1 ogGNP


- O.O1/1ogGNP


- O.99logDST/,0gGNP + O.06ogTCM/!1ogGNP
(0.2045) (0.1768)


+ 10.(10ogGNP/iogGNP
(2.5917)


- OO08]ogCVM/iogGNP


110.30i57


= 16.30


0 1.2301


-- 140;


upper sanpie:


SogNCNTX/ I ogGNP


- 1.07/1og-GNP


- 0.32o g T/l Io GNP
(0.4296)


+ 0.47ogTCM/1 ogGNP
(0.2548)


+ 10. 51ioGNP/1ogGNP


- O 5610gCVH/o1gGNP


= 0.0611


- 3.22


*- 0,.2156


- 140C.


The residual sum


squares


from the lower smtple, (n- 5)S, was


7.1473 and the residual


Sum of squares


4-;n


upper


sarmple,


2(n-5)S, was 5.4680.


The computed F of 1.1051 was


l ess


than the tabu-


Sated F of 1. 3305,


so apparently heterosce dasticit


attributed to the


gross national product variable had been corrected fort

Number of Tons South Atlantic-Gulf Coast, 1976


Data collapsed to the foreign country


1iv


ye ar


1976


yielded the following regression equation:


s, and:i, r~ ci






87




and increased the adjusted R-square witr respect to the prior model of


port-to-port movements.


Among the three


independent variables the


highest F-value


was for the coefficient


DST (1


H1owed by the


coeffi ci ent


of GNP (14.37


The Pearson


co rre


nation coefficient ts for the significant variables


were


h9Q~i


loqGNP


IoaEGY


1 ogEGY
IH.'"*-**10a'*"'EGY ''


1.00000
0.0000

0.16014
0.0054

-0.09249
0.1099


The mass variab


gross


0.16014
O.0054

1.00000
0.0000
0.51739
0.0001

national


-0.09249
0.1099

0.51739
0.0001

1.00000
0.0000

product and energy consumption


were moderately highly correl


ated


and, t


therefore,


precs


ented


bias


problems associated with niulticalo i


i near


Ne vert hel


es s,


they


~eoec


not dropped


from


the anal


to prio"
- r


i" >: cdSr; l) l)


Tests for h


etercs cedas


fj ci


the fol' win4-g


two: equai t? on rrs


when sorted


on the


of distance


l power


TNSX


S6. 18


ogDST
4


"2 o10aGUNP


0.0899:~r


-",J tu


- ,.


it


upper sanp : i


ogNTNSX


= 4,74


- O. 541ogDSI +
*In u T 1 'J ^1I *


[


43i~og

.121


ogEGY


;ar: pi>


.I4Gli;ciE';'r












882.0209 and the


n-)S was


res~ 0 rU3]


6809.


sum of


s qu rs i


comp t: t


Sc r


iS r 1


-,


s than


tabu


t ed F of 1.


3300


therefore,


cedtc


cdv


~t 3:; Vt
>>


C z~~c


with


regard


to the distance


v arl~ a i, C .?


Sam1e


Pl~lc? il"2


was accompl ished


tort i


on tne 1


GNP.


resu1 ts


regressio


equations


o:er


sampr


logNTNSX


- 12.34


- 0.89
10.230


o0DST
7"


- 0.ia


oGNP!,


. 201 oaEGY
1n 0
*?- I- *


= 0.1172


7,1 i-


1. 063


= 140;


upper


oaNTNSX


-1s


0.181 ogST
(0.5059)


4G
-


7 o rgEGY:


6,0(;3


resil dua


s jm or


i:C~i t n:` \


sec


1ne


4I 2< > )


603.6325

e e
In-4)S ,

exceeded


the r


a
< <<0


Sis!rt


Ki-s


~LKJw~t


L:aL; "2


ol


15C3


nul tip]


Tooth


si des


s ides


; :4 -r V


ioGNP to


adj Just


e e eros cedas i city.


in~~ r:i. tg


:q~t~tA O


V Yin


became


l ooNTNS X/


oqOMI~P


-0.20/1


oqGNP


-,?~?CT I 'oi;N


9.211 ogGNP/1i ogGNP
(1.4959)


+ 0. 61 oEGY/ (0.0899)


.,A t>

I


r- '


vhs3


(t'1 ;


~:flt i


a


1<


i-4


(9:


ice, P *


ir ~f


)1Ci t ~r` 3"


tl: r


0. r35 i


0. 051~(3


71


~ r"


;v" :: i


i


n


*


rtnrr,