こ[:己]甘 田本オペレーションズ0リサーチ学会 2◎04年啓季節究発表会
つPh怠P忘じme且こ‥一之七・三こ翫ユエこ、:n・三七ご五lニ王こ伽deユミnぎlつゴFご・野a皿二(己Se
恐ⅧS鼠m邸$混細『y五感⑳闊⑳貯也ぬÅm暖野鼠旺a
Model,Ⅱstimation,andEntryAnalysis
OllO6726 西南学院大学商学部 小島平夫 KOJIMA斑正ao
arethoseindividuaトand也me−SpeCifice蝕ctstha七are
れi 笠崇競:ご霊監恕h血,血dan。血eeト
fects are触edor random,the modelwi11be corre−
SpOndingレfiⅩed−Orranddm−e鮎ctsmodel・
Mod¢胞w弛払①m皿y瓦md五Ⅴ五du劇¢飴c七郎 mOde旭 W弛払0畑y血豆m¢¢飴c七s(om¢珊ay㊤訂貯①町COmpか 皿e皿缶mo舶A)Fbr七heindividual−0nlymodeJs,丘Ⅹed−
e蝕c七s:
且 二晩触闇触感摘m
Thepapera紬mp七s,inSec也on2,tOb山1dstaticpanel dataeconome也cmodelsofthe3apanegebⅥ虚nessen−
tryintotheNorthAme血canmazketand,inS眈也on
也汀Sand丘Ⅶ鮎calye甜Sl算7thougb2001willcom−
p虚sethepaneldata.
ewe椚CO鮎erned
鵬.1
Emp五町温c舶 五鎚Ⅷ¢S 昆m 五感◎釘m乱缶昆om如 bⅦS五m℡SS
欝漂鬱£慧㌫誌慧慧謂よ冨芸濃
i七woormozena也onB)exist?Th把emajoThypothese8
,thAmedcaisasubsti_
to㌘叢焉芸霊誓欝ミ謂ム
(4)
即=∬β十皿Ⅳ脚十〝.
鮎ndom−e鮎cts:
財=α且Ⅳr十gβ十匹⑳gT+ぴ. (5)
Fbrthe也me−0山ymodek,丘Ⅹed−e蝕cts‥
財=∬β十皿TÅ十〝・ (6)
如ね払rexpor btbe陀glOn.
OrpOrate grOWthin如e willlikely
出 ム:Indl血
resultintheF NorもhAme丘ca. Random−e鮎cts:
財=α底ⅣT十∬β十(度Ⅳ⑳∬T)Å+〝・ (7)
Mod¢旭 w弛払 bo缶払 五md五Ⅴ邑dⅧa且 amd 七五me ¢匹 細かW野㊤訂町0訂正O叩①m¢血mo掴)F奴ed ̄
即=αgⅣr十∬β十皿Ⅳ♯仇+皿丁寧Åヰ十〝.(8)
R烏ndom−e庶bcts:
野=α恥汀+∬β十脚⑳鯨+(慮Ⅳ⑳∬T)Å+肌(9)
geJ∝抽lタggOぬ勒巾βWegJ鮎わ00勒ノ動αJ九脚0仇e−
Sio
thel。Cauyfinalhyp。七山癒c。inci。eswith
b忠蕊諾謂Pleman画d陀幻ⅧeSCO血b血
2 −P忍皿e且]D乱世a M①de胞
Thestaticpanelda七aeconometdcmodeisofforeign
byぬ匹n既鮎皿血臆d五ntbepaper
‥
R瓜の通℡且w弛ぬm℡紬払¢釘五md五Ⅴ五d弧a且皿0町立五皿¢㊤飴c也β
Ol∬fundamentalmodel,tObecoIutaStedwithother
C
,
constrainedmodelinthesenBethatneitherindividual
nor也mevaria也onsocclば:
的t=α+荘;tβ+叫わ 慮=1,…,Ⅳ;f=1,‥.,r(1)
Wheze yit=Japanese FDIin North Ame血ca and
Xit=Obsen,abLe deteminants of the F:)Ⅰ.In vector ぬm,財=α脛Ⅳr十∬β十仏 (2)
t
。。s
叫t=拘+ん+巧t (3)
お忠恕霊t禁芸。諾盟 孟
由
La^isacceptedbyHall斑n鱒nSpeCifica也on the preBenCe
nalbypot血e 1ngpre寧entth
(8
dud
地車,(81年て鱒d与鱒C享畠甲 brrandom e鮎cts.
0抽雌由coi皿Cides w止血 locdly丘皿al
thepresenceofonbrindivie鮎ctsoronly也meeト
bcts,thenretl旺ntOeaSlierspecifica也ontesttable(s)
品
i
for蝕ede蝕ctsor(7)forrandome蝕c七s・
ー288−
© 日本オペレーションズ・リサーチ学会. 無断複写・複製・転載を禁ず.
1払blel:DeterminantsofJapane茫FDIinNorthAmerica
DependentVidable
ratio.Balc6.L8ubsale6PJ8ut6ale6inU Lsubwr叫 L飢1bnu皿 Ue血血tion:上tomle性torig血t U山ylndividual Bot血 Sut冶idi8rySales/pare皿tS山饉
Fixed R且Ildom Fixed 10ggdsu臨i血arys山饉
planatory/DummyVariables Si伊10fCoe仔. OggedsubsidiarysalesinU・S・doua
Constant loggednumberofsutBidiaryworkws
ratio−eXpOrtLLJ (−)一事 (−) − ) (+) 10ggednumberofsut讃idiari窃
Lparent」Mm ∫1 (+)■けl■ (+)■一事 (+)攣 (+)蝉岬■
OWnCa匹atio J2 II
intpymnt J2 l−)不一
Varatio J2 ▼
LvaperwTkr J2 (+)− (+)事 foodtlJ b
textile(2J b】
1urrber,pldp,paper(3)
Cbe血cal(4J lb」
generalmacbinery(7)
dec七山calmac山nery(8) 0.000
tranBpOrtationeqtlipm別止(9) a (+)−−−
preCi8ioninβtruJ刀亡nt【1UJ 8
petroleumandcoal(11) b lb】 (−)▼▼;lb】
miscell・mfg・Ll:り b
agricuItlⅡe,forestry(13)
miningt14J a
COn如mctiont15J lbl II
COmmerCe(16J
SerVICe(17J (+)−−;La」
Ⅱ止SCell・mn−m咤・tlB)
0.000
t+)
FY20叩(4) (+)一事■
上−Y2UUlし5J
3 Panel Data and Estimation
慧㌔認諾㌫監琵eL深沈怒冨?鑑
監f霊宝霊霊競悪箪笥と慧悪霊盟慧
Japan胱CorporateSta也stics,2001−2002FiBCalYbar.
Estimation The models a柁eStimated by RA2T
(=Regre感onAnalydsofTimeSeries)proamS・
4 Determinants of Japanese FDIin North America
Loca塩on−andinternaliza也oDtheore也chypotheses,L,
IlandIiinthefirst紀C也onweretestedandthetest results ba虎dontheglobanyfinalmodels ale being
putto辞tI肥ー血丁乱blel:
First,COndstentwithListheob紀rvednqgative彗一
bpaIentbtalsal錯).
Second,COnSistentwithIlisthepositiverela岨onr Ship of Lparent」nu皿(=logged number of parent
負・
withh8此thep。Sitiveefhct。f
LvaperwTkr(=loggedvalueaddedperworker)and
then印ativesignforintⅣⅡmt(=borrowingrateof
interestin%).
togetFDIbysuch紀CtOrSaS♪od,te虎ile,Chemi−
笥
α
peかpヱetJm肌d00αらm豆βα〃αneOMm ね再喝,
by the and constructionsectorslessactivethan
代
。_
cisioni7Wれment,minin9.andsenice(監e【a】inthe
paredtoFY1997.
References
【1】Hora即d止,H.,凡re亘れ加代Cf九帽亡meれf〆仇e九匹几eβe Fi7nS,Univ耶ib,OflbkyoPre$S(1992)(inJapaneSe)・
[2]Hsiao,C.,A7Wlysis ofPaneLDaぬ,2nded・,Cambridge
U山ver由tyPrt芳S(2003)・
【3】禁漁よ㌫慧慧宝器芸当だ恵三…㌫慧恕霊:
箭㍑監禁酢監怨㌫虎児忠邦詔7笥轟
(inJapanese)・
−289−
© 日本オペレーションズ・リサーチ学会. 無断複写・複製・転載を禁ず.