I Premire Partie : Courbes inflation-chmage

2

Elemente de corectri (Curba Phillips)

I-ma parte: Curbele inflaie-omaj

Reprezentarea graficului

Iat reprezentrile grafice ale relaiei inflaie-omaj pentru Frana i Statele Unite. Putem s ne gndim c este prea dificil de a gsi o relaie simplificativ pentru dou ri ntre omaj i inflaie la o dat deoarece cele dou curbe sunt prea instabile.

2. Estimaia

Frana

Estimm dup relaia unei constante ntre nivelul inflaiei i de cursul omajului ntrziat. E de ajuns s ptrundem n spaiul calcului din Eviews:

Ls inffra c chofra(-1)

Rezultatele obinute sunt urmtoarele(q2frar):

Tabelul 1

Dependent Variable: INFFRAMethod: Least SquaresDate: 05/07/00 Time: 11:39Sample(adjusted): 1965 1997Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C8.9183671.3048886.8345820.0000CHOFRA(-1)-0.4420730.173685-2.5452620.0161R-squared0.172856 Mean dependent var6.022121Adjusted R-squared0.146174 S.D. dependent var3.970653S.E. of regression3.668991 Akaike info criterion5.496402Sum squared resid417.3064 Schwarz criterion5.587100Log likelihood-88.69064 F-statistic6.478359Durbin-Watson stat0.233211 Prob(F-statistic)0.016113Obinem urmtoarea relaie: inffra = 8.918 0.442*choffra(-1)

Remarcm c probabilitatea critic a variabilei chaffra (-1) este inferioar de 5%, deci cursul omajului ntrziat este o variabil semnificativ. Mai constatm c regresia este global semnificativ de aceea probabilitatea critic a lui Fisher este inferioar de 5%.

Estimm acum relaia ntre inflaie i cursul omajului contemporan. Obinem ptrunznd n Eviews:Ls inffra c chofra (q2fra)

Tabelul 2Dependent Variable: INFFRAMethod: Least SquaresDate: 05/07/00 Time: 11:50Sample: 1964 1997Included observations: 34VariableCoefficientStd. Errort-StatisticProb. C8.2734001.3222306.2571580.0000CHOFRA-0.3451050.171588-2.0112500.0528R-squared0.112224 Mean dependent var5.952059Adjusted R-squared0.084481 S.D. dependent var3.931313S.E. of regression3.761589 Akaike info criterion5.544582Sum squared resid452.7857 Schwarz criterion5.634368Log likelihood-92.25790 F-statistic4.045125Durbin-Watson stat0.217186 Prob(F-statistic)0.052786

n acest caz, trebuia de remarcat c trebuia de luat cursul omajului contemporan nu este la fel semnificativ. In efect probabilitatea critic este uor mai superioar de 5%, dar rmne la fel acceptabil.

2.2. Statele UniteProcedm ca i n cazul Franei, ptrundem n Eviews:Ls infusa c chousa(-1) (q2usar)

Tabelul 3Dependent Variable: INFUSAMethod: Least SquaresDate: 05/07/00 Time: 12:01Sample(adjusted): 1965 1997Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C5.7181622.1473582.6628830.0122CHOUSA(-1)-0.0933470.343655-0.2716290.7877R-squared0.002374 Mean dependent var5.152424Adjusted R-squared-0.029807 S.D. dependent var2.959199S.E. of regression3.002977 Akaike info criterion5.095777Sum squared resid279.5540 Schwarz criterion5.186475Log likelihood-82.08032 F-statistic0.073783Durbin-Watson stat0.421963 Prob(F-statistic)0.787708

Et ls infusa c chousa (q2usa)

Tabelul 4Dependent Variable: INFUSAMethod: Least SquaresDate: 05/07/00 Time: 12:03Sample: 1964 1997Included observations: 34VariableCoefficientStd. Errort-StatisticProb. C2.3295242.0820071.1188840.2715CHOUSA0.4495180.3350991.3414460.1892R-squared0.053240 Mean dependent var5.038529Adjusted R-squared0.023654 S.D. dependent var2.988737S.E. of regression2.953178 Akaike info criterion5.060663Sum squared resid279.0803 Schwarz criterion5.150449Log likelihood-84.03127 F-statistic1.799477Durbin-Watson stat0.525849 Prob(F-statistic)0.189219

Remarcm c nici cursul omajului ntrziat, nici cursul omajului contemporan sunt semnificative, deoarece probabilitile lor critice sunt superioare de 5%.Noi putem, deci conclude c luarea omajului ntrziat sau omajului contemporan nu schimb semnificativ rezultatele.

3. Examinarea rezultatelorIat reprezentarea grafic a rezultatelor relaiei inflaie-omaj ntrziat pentru Frana i pentru Statele Unite. Plecnd de la aceste 2 grafice putem presupune prezena autocorelaiei pozitive a rezultatelor deoarece va prea c rezultatele .. de o dinamic ciclic.In ceea ce privete punctele .(aberants?) apar rezultatele ndeprtate pentru anii 1974 i 1981 pentru Frana ca i pentru Statele Unite, rezultatele sunt ndeprtate pentru anii 1974 i 1980.

Frana SUA

Residual Plot| * . | . |

Residual Plot| * . | . |

| *. | . || * . | . |

| .* | . || * | . |

| * | . || . * | . |

| . * | . || .* | . |

| . * . || . * | . |

| . * . || . * | . |

| . *| . || . * . |

| .* | . || . | . * |

| . |* . || . | * |

| . | . * || . | * . |

| . | . * || . | * . |

| . |* . || . | * . |

| . | * . || . | .* |

| . | *. || . | . *|

| . | . * || . | . *|

| . | . *|| . | . * |

| . | . * || . | .* |

| . | * . || . | * . |

| . * | . || . | * . |

| . *| . || . * | . |

| . * | . || . * | . |

| * | . || . * | . |

| . * | . || . * | . |

| . * | . || . * | . |

| . *| . || . * | . |

| . * . || . * | . |

| . *| . || . * | . |

| .* | . || . * | . |

| .* | . || . * | . |

| * | . || . * | . |

| .* | . || . * | . |

| .* | . |

| * | . |

4. Testele de autocorelaia erorilor

4.1. Frana

Este posibil de aplicat testul Durbin i Watson n relaia dintre inflaie i cursul omajului ntrziat deoarece numrul observaiilor este superior de 15, variabile a explicat nu figura printre variabile explicative i modelul este specificat n serii temporale.Statistica lui Durbin i Watson a tabelului 1 este egal cu 0,233211. Sau cnd citim pe tabl de Durbin i Watson pentru 33 observri i o variabil explicativ, gsim c d1=1,38 i d2=1,51. Statistica lui Durbin i Watson n acest caz este inferioar de d1, putem, deci conclude la o corelaie pozitiv de erori.Este egal posibil de a face un test reprogramat de Eviews. Asta avem fcnd testul Breusch-Godfrey.Ideea general a acestui test rezult din precutarea unei relaii semnificative ntre rezultatul i ceea ce a rezultat.

Efectum testul lui Breusch-Godfrey:

Breusch-Godfrey Serial Correlation LM Test:F-statistic73.24807 Probability0.000000Obs*R-squared23.41144 Probability0.000001

Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 05/09/00 Time: 22:06VariableCoefficientStd. Errort-StatisticProb. C0.3319770.7160630.4636130.6463CHOFRA(-1)-0.0595200.095424-0.6237390.5375RESID(-1)0.8498590.0993008.5585090.0000R-squared0.709438 Mean dependent var1.17E-15Adjusted R-squared0.690067 S.D. dependent var3.611208S.E. of regression2.010419 Akaike info criterion4.321071Sum squared resid121.2535 Schwarz criterion4.457117Log likelihood-68.29767 F-statistic36.62404Durbin-Watson stat1.497028 Prob(F-statistic)0.000000

Respingem ipoteza absenei a autocorelaiei, deoarece probabilitatea critic de fischer este inferioar de 5%.

4.2. Statele Unite

Statistica Durbin i Watson este aici egal cu 0,421963. Ins d1 i d2 au aceleai valori ca i pentru Frana, deci precum statistica lui Durbin i Watson este inferioar de d1, putem conclude la o autocorelaie de erori

Testul lui Breush-GodfreyBreusch-Godfrey Serial Correlation LM Test:F-statistic48.13823 Probability0.000000Obs*R-squared20.33014 Probability0.000007

Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 05/09/00 Time: 22:12VariableCoefficientStd. Errort-StatisticProb. C2.1041791.3861351.5180190.1395CHOUSA(-1)-0.3586290.222543-1.6115040.1175RESID(-1)0.8188330.1180196.9381720.0000R-squared0.616065 Mean dependent var7.27E-16Adjusted R-squared0.590469 S.D. dependent var2.955683S.E. of regression1.891477 Akaike info criterion4.199101Sum squared resid107.3306 Schwarz criterion4.335147Log likelihood-66.28517 F-statistic24.06911Durbin-Watson stat1.323774 Prob(F-statistic)0.000001

Respingem ipoteza absenei autocorelaiei deoarece probabilitatea critic de fisher este inferioar de 5%. Putem, deci conclude la o autocorelaie de erori pozitive.

II Curbele Phillips originale (fr anticipaia inflaiei) avnd i autocorelaia rezultatelor i ocul negativ a ofertei (ocuri petroliere 1974 i 1980)

6. Estimaia curbei Phillips fr autocorelaia reziduurilor

Sistemul urmtor:

t = .(-ut-1) (1)t = - -1. t (2) =c. t + (3)

Replasnd n expresia (3) t et t de relaiile (2) et (3), gsim c:

= c.( - -1. t) + .(-ut-1) = - -1.c. (ut-u t-1) + .(-ut-1) deci

Deci = - -1.c. ut ( + -1.c) u t-1 + .

Estimm aceast ecuaie cu ajutorul Eviews-lui.

6.1 Frana

Ptrundem n zona de comand: ls inffra c chofra chofra(-1) i obinem (q6fra):

Dependent Variable: INFFRAMethod: Least SquaresDate: 05/07/00 Time: 17:37Sample(adjusted): 1965 1997Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C8.1958131.3410906.1113050.0000CHOFRA1.9787001.1875701.6661760.1061CHOFRA(-1)-2.4120741.194353-2.0195650.0524R-squared0.242915 Mean dependent var6.022121Adjusted R-squared0.192443 S.D. dependent var3.970653S.E. of regression3.568195 Akaike info criterion5.468505Sum squared resid381.9605 Schwarz criterion5.604551Log likelihood-87.23033 F-statistic4.812839Durbin-Watson stat0.436616 Prob(F-statistic)0.015388

Obinem relaia urmtoare: pt= 1.978*ut- 2.412 *u t-1 + 8.195Remarcm c acest model este global specific deaorece probabilitatea critic de Fisher este inferioar de 5%. In timpul c cele 2 variabile nu snt semnificative deoarece probabilitile lor respective sunt superioare de 5%.

6.2.Statele Unite

Se apuc n zona de comand de acelai fel ls infusa c lucru (-1) i se obine (q6usa):

Dependent Variable: INFUSAMethod: Least SquaresDate: 05/07/00 Time: 17:45Sample(adjusted): 1965 1997Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C4.1237592.0739711.9883400.0560CHOUSA1.3632280.5332172.5566100.0159CHOUSA(-1)-1.1928170.533991-2.2337770.0331R-squared0.180847 Mean dependent var5.152424Adjusted R-squared0.126237 S.D. dependent var2.959199S.E. of regression2.766119 Akaike info criterion4.959276Sum squared resid229.5425 Schwarz criterion5.095322Log likelihood-78.82805 F-statistic3.311606Durbin-Watson stat0.683192 Prob(F-statistic)0.050173

Se obine ecuaia urmtoare: pt= 1.363*ut- 1.192 *u t-1 + 4.123

n ceea ce se refer Statelor Unite, acest model e mai adaptat dect cele 2 precedente, deoarece 2 variabile (omajul i omajul ntrziat) sunt semnificative. Probabilitatea critic Fisher este aproape egal la 5%, ceea ce arat c modelul e bine specificat

7. Luarea n numr autocorelaia reziduurilor.

7.1 Frana

Se ocup n zona de comand: ls inffra c chofra chofra(-1) dum ar(1) i obinem:Dependent Variable: INFFRAMethod: Least SquaresDate: 05/07/00 Time: 17:59Sample(adjusted): 1966 1997Included observations: 32 after adjusting endpointsConvergence achieved after 25 iterationsVariableCoefficientStd. Errort-StatisticProb. C15.140606.2694022.4149990.0228CHOFRA-0.2542290.522127-0.4869100.6303CHOFRA(-1)-0.7881790.495101-1.5919580.1230DUM2.6747370.7974773.3540000.0024AR(1)0.8665770.08184210.588420.0000R-squared0.888675 Mean dependent var6.137187Adjusted R-squared0.872182 S.D. dependent var3.977895S.E. of regression1.422161 Akaike info criterion3.684833Sum squared resid54.60861 Schwarz criterion3.913854Log likelihood-53.95732 F-statistic53.88326Durbin-Watson stat1.791241 Prob(F-statistic)0.000000Inverted AR Roots .87

Se constat n acest model c omajele ntrziate i contemporane nu sunt semnificative, dar regresia este global semnificativ

7.2. Statele Unite

Se ocupls infusa c chousa chousa(-1) dum ar(1) i obinem:Dependent Variable: INFUSAMethod: Least SquaresDate: 05/07/00 Time: 17:52Sample(adjusted): 1966 1997Included observations: 32 after adjusting endpointsConvergence achieved after 9 iterationsVariableCoefficientStd. Errort-StatisticProb. C12.668703.4750673.6455970.0011CHOUSA-0.2213110.292094-0.7576700.4552CHOUSA(-1)-0.9636670.296656-3.2484270.0031DUM2.2938750.7679182.9871370.0059AR(1)0.8951210.08637810.362880.0000R-squared0.801747 Mean dependent var5.260625Adjusted R-squared0.772377 S.D. dependent var2.939478S.E. of regression1.402421 Akaike info criterion3.656878Sum squared resid53.10321 Schwarz criterion3.885900Log likelihood-53.51005 F-statistic27.29748Durbin-Watson stat1.552390 Prob(F-statistic)0.000000Inverted AR Roots .90Se remarc c n acest caz, variabilele dummy i omajul contemporan nu sunt semnificative chiar dac omajul contemporan era n modelele precedente (nu ndreapt corelaiile reziduurilor i punctele aberante). Totui, relaia global semnificativ

8. Analiza regresiei cnd s-a ndreptat corelaia reziduurilor

Frana SUA

Residual PlotResidual Plot| . * | . || . * . || .* | . || * | . || . | * . || . | * . || . | .* || . | * . || * . | . || . * . || . * . || . * . || . *| . || . * | . || . | * . || . | . * || . | . * || . | . * || . | * . || . | . * || . * | . || . * | . || . * . || . * . || . * . || . | * . || . | . * || . | . *|| . | * . || . | * . || . | . *|| . | * . || . |* . || * . | . || . * | . || . * | . || .* | . || . | *. || . * . || * . | . || * . | . || * . | . || . |* . || . | *. || . * | . || . * | . || . * . || . * . || . * | . || . |* . || . * | . || . * | . || . * | . || . * . || . |* . || . |* . || . | * . || * | . || . |* . || .* | . || . *| . || . * | . || . *| . || .* | . |Se vede c reziduurile nu urmeaz un ciclu cum se fceau mai nainte, adic se poate rezolva prin a avea ridicat autocorelaia reziduurilor.

III: Curbele Phillips argumente anticipate inflaiei

9. Estimarea modelului

Se caut acum la estimarea modelului urmtor:

t = t + .(-ut-1) (1) t = - -1. t (2) =c. t + (3)t = (-1 -2 ) (4)

Procednd n acela fel pentru momentul precedent, se gsete relaia urmtoare::

deci = - -1.c. ut ( + -1.c) u t-1 + ((-1 -2 )+.9.1. Frana

Pentru a estima acest model fr ajustarea corelaia reziduurilor, se introfduce n zona de comand: ls infra c chofra chofra(-1) inffra(-1) dum (q9fra)

Dependent Variable: INFFRAMethod: Least SquaresDate: 05/07/00 Time: 19:41Sample(adjusted): 1965 1997Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C1.6633080.6259292.6573450.0129CHOFRA-0.3608130.439505-0.8209530.4186CHOFRA(-1)0.2252520.4487850.5019160.6197INFFRA(-1)0.8310800.06343313.101700.0000DUM4.9953350.8981385.5618810.0000R-squared0.921603 Mean dependent var6.022121Adjusted R-squared0.910404 S.D. dependent var3.970653S.E. of regression1.188522 Akaike info criterion3.322025Sum squared resid39.55237 Schwarz criterion3.548769Log likelihood-49.81342 F-statistic82.28930Durbin-Watson stat2.195938 Prob(F-statistic)0.000000

Se remarc c modelul e semnificativ deoarece probabilitatea Fisher e inferioar la 5%. n revan se remarc c variabilele choffra et choffra(-1) nu sunt semnificative.

Se poate calcula un test durbin-h la cazul prezent unei variabile ntrziate.

Se pune H0: h=0 H1: h diferit de 0sau =1- 0.5*DW deci = -0.095

Or

Deci h = 0.58Sau Student lu pentru 33 la pargul de 5% e egal cu 1.96 se deduce c se accept H0; deci nu are autocorelaia erorilor.

9.2. Statele Unite

Se introduce n zonele de calcul: ls c chousa chousa(-1) infusa(-1) dum (q9usa) i obinem:C3.7041370.9663803.8330040.0007CHOUSA-1.0251150.356171-2.8781590.0076CHOUSA(-1)0.4429100.2954411.4991510.1450INFUSA(-1)0.9169930.1158937.9124430.0000DUM4.6304721.0051354.6068150.0001R-squared0.836388 Mean dependent var5.152424Adjusted R-squared0.813015 S.D. dependent var2.959199S.E. of regression1.279608 Akaike info criterion3.469713Sum squared resid45.84714 Schwarz criterion3.696456Log likelihood-52.25026 F-statistic35.78426Durbin-Watson stat2.047469 Prob(F-statistic)0.000000

Se vede c modelul este semnificativ deoarece probabilitatea criteriului Fisher e inferioar la 5%. n acest model unic preul omajului contemporan nu e semnificativ.Se poate egala la efectuarea testului Durbin-h. Se gsete c h = 0.1798 care e foarte inferior la 1.96. Se poate deci rezolva care nu are autocorelaia erorilor.

10. Testul Chow

10.1 Frana

Se efectueaz testul Chow cu date cheie 1983. Deci se divizeaz n dou serii.Cum pentru a doua prioad matricea nlocuitoare este egal cu 0 i c n o poate inversa: se reine pentru dou perioade.. Se scrie n zona de comand:

Ls inffra c chofra chofra(-1) inffra(-1) (q10fra)

i obinem modelul global:

Dependent Variable: INFFRAMethod: Least SquaresDate: 05/16/00 Time: 22:56Sample: 1965 1997Included observations: 33VariableCoefficientStd. Errort-StatisticProb. C1.8358300.8912032.0599460.0485CHOFRA-0.6793970.621198-1.0936890.2831CHOFRA(-1)0.5131240.6355010.8074310.4260INFFRA(-1)0.9030390.08852710.200750.0000R-squared0.834990 Mean dependent var6.022121Adjusted R-squared0.817920 S.D. dependent var3.970653S.E. of regression1.694311 Akaike info criterion4.005642Sum squared resid83.25000 Schwarz criterion4.187037Log likelihood-62.09309 F-statistic48.91563Durbin-Watson stat1.756427 Prob(F-statistic)0.000000

Deci SCR=83.25

Se obine pentru prima subperioad: (q10fra1)

Dependent Variable: INFFRAMethod: Least SquaresDate: 05/16/00 Time: 22:52Sample: 1965 1983Included observations: 19VariableCoefficientStd. Errort-StatisticProb. C1.5554691.1728101.3262750.2046CHOFRA-1.1236421.710828-0.6567830.5213CHOFRA(-1)1.1871181.6335650.7267040.4786INFFRA(-1)0.8639200.2695103.2055170.0059R-squared0.726556 Mean dependent var8.192632Adjusted R-squared0.671867 S.D. dependent var3.772173S.E. of regression2.160811 Akaike info criterion4.563508Sum squared resid70.03655 Schwarz criterion4.762337Log likelihood-39.35332 F-statistic13.28525Durbin-Watson stat1.632930 Prob(F-statistic)0.000169

Deci SCR1= 70.03

Si pentru a doua subperioad:

Dependent Variable: INFFRAMethod: Least SquaresDate: 05/15/00 Time: 12:38Sample: 1984 1997Included observations: 14VariableCoefficientStd. Errort-StatisticProb. C4.6978072.5983611.8079890.1007CHOFRA-0.2386450.310433-0.7687470.4598CHOFRA(-1)-0.1092700.362999-0.3010200.7696INFFRA(-1)0.5598100.1184274.7270350.0008R-squared0.864791 Mean dependent var3.076429Adjusted R-squared0.824228 S.D. dependent var1.696080S.E. of regression0.711084 Akaike info criterion2.390905Sum squared resid5.056411 Schwarz criterion2.573493Log likelihood-12.73634 F-statistic21.31980Durbin-Watson stat2.554147 Prob(F-statistic)0.000115

Deci SCR2=5.056

F= cu ddln=k+1=4 et ddld =n-2(k+1)=25

Deci F=0.68 Sau Fisher lu la 5% pentru 4 et 25 este egal la 2.76 deci nu e semnificativ a diviza modelul n dou subperioade.

Se efectueaz acum testul cu funcia preprogramat n E-views. Se obine n q10fratest resultatul care confirm aceasta ca s gsim precednetul, deoarece Fisher e superior la 5%.

Chow Breakpoint Test: 1984 F-statistic0.678912 Probability0.6130200Log likelihood ratio3.403003 Probability0.492779

10.2 Statele Unite

Se obine regresia asupra perioadei complete, introducnd n zona de calcul:Ls infusa c chousa chousa(-1) infusa(-1) (q10usa)i se obine:

Dependent Variable: INFUSAMethod: Least SquaresDate: 05/16/00 Time: 23:06Sample: 1965 1997Included observations: 33VariableCoefficientStd. Errort-StatisticProb. C4.2347551.2500423.3876890.0020CHOUSA-1.0841970.463725-2.3380180.0265CHOUSA(-1)0.3373170.3837460.8790120.3866INFUSA(-1)1.0631220.1452217.3206950.0000R-squared0.712378 Mean dependent var5.152424Adjusted R-squared0.682624 S.D. dependent var2.959199S.E. of regression1.667098 Akaike info criterion3.973258Sum squared resid80.59722 Schwarz criterion4.154653Log likelihood-61.55875 F-statistic23.94228Durbin-Watson stat1.606710 Prob(F-statistic)0.000000

Deci SCR=80.59

Se efectueaz regresiile asupra dou subperioade 1965-1980 et 1981-1997.

i se obine (q10usa1) et (q10usa2) Prima subperioadDependent Variable: INFUSAMethod: Least SquaresDate: 05/16/00 Time: 23:09Sample: 1965 1980Included observations: 16VariableCoefficientStd. Errort-StatisticProb. C4.6182432.4365671.8953890.0824CHOUSA-1.6004651.099615-1.4554780.1712CHOUSA(-1)0.3858020.6559330.5881740.5673INFUSA(-1)1.4991160.4131923.6281360.0035R-squared0.699483 Mean dependent var6.360625Adjusted R-squared0.624353 S.D. dependent var3.397975S.E. of regression2.082619 Akaike info criterion4.517448Sum squared resid52.04763 Schwarz criterion4.710595Log likelihood-32.13958 F-statistic9.310384Durbin-Watson stat1.562543 Prob(F-statistic)0.001861

SCR1 = 52.04S-a nlat variabila dum deoarece ocul n 1980 i pentru a doua subperioad a matricei nu va fi inversibil.

Pentru a doua subperioad, se obine:Dependent Variable: INFUSAMethod: Least SquaresDate: 05/15/00 Time: 13:01Sample: 1981 1997Included observations: 17VariableCoefficientStd. Errort-StatisticProb. C3.4722561.0745563.2313410.0066CHOUSA-0.9741250.311533-3.1268790.0080CHOUSA(-1)0.4941710.2732911.8082240.0938INFUSA(-1)0.7861770.0926878.4820140.0000R-squared0.871446 Mean dependent var4.015294Adjusted R-squared0.841780 S.D. dependent var1.963540S.E. of regression0.781034 Akaike info criterion2.545928Sum squared resid7.930186 Schwarz criterion2.741979Log likelihood-17.64039 F-statistic29.37507Durbin-Watson stat1.986943 Prob(F-statistic)0.000005

Se are SCR2=7.93

Se calculeaz acelai fel Fisher i se obine o valoare de 2.14Sau Fisher lu n tabel e asemntor ca pentru Frana, deci se accept ipoteza c a diviza modelul n dou subperioade ne ameliorat semnificativ modelul.

Chow Breakpoint Test: 1981 F-statistic2.148648 Probability0.104452Log likelihood ratio9.751148 Probability0.044835

11. Phi

Se construiete seria:

Iat reprezentaiile grafice pentru Statele Unite i Frana a relaiei ntre omaj i phi. Totui se gsete o relaie mai semnificativ, dac se ia omajul ntrziat.Se remarc tendina curbei Franei la pant mai puin important ca acea a Statelor Unite.Deci ne permitem emiterea ipotezei unei politici dezinflaioniste mai puin negative la SUA ca n Frana. Se poate n rezolvarea unei politici dezinflaioniste e mai scump n termenii omajului n Frana ca la SUA.

12 Metoda TSLS

Se poate efectua metoda TSLS cu un omaj ntrziat de 4 perioade deoarece numrul regresiilor este egal cu 5 i se obine tanelele urmtoare:

Pour la FranceDependent Variable: INFFRAMethod: Least SquaresDate: 05/16/00 Time: 23:54Sample(adjusted): 1968 1997Included observations: 30 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C10.733461.3619887.8807240.0000CHOFRA(-1)0.5863661.1241840.5215920.6065CHOFRA(-2)-0.8108371.865376-0.4346780.6675CHOFRA(-3)0.5846172.0174940.2897740.7744CHOFRA(-4)-1.1334021.265536-0.8955910.3790R-squared0.451718 Mean dependent var6.358333Adjusted R-squared0.363992 S.D. dependent var4.013147S.E. of regression3.200488 Akaike info criterion5.315496Sum squared resid256.0781 Schwarz criterion5.549028Log likelihood-74.73243 F-statistic5.149236Durbin-Watson stat0.428976 Prob(F-statistic)0.003632

Se observ c relaia e global bine specificat i c variabilele explicative nu sunt semnificative.

Pentru SUAn acest caz se remarc c modelul e bine specificat i c variabilele explicative nu mai sunt semnificative.Dependent Variable: INFUSAMethod: Least SquaresDate: 05/16/00 Time: 23:56Sample(adjusted): 1968 1997Included observations: 30 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C7.0856212.9191752.4272680.0228CHOUSA(-1)-0.2563280.705819-0.3631650.7195CHOUSA(-2)-0.1962351.050882-0.1867340.8534CHOUSA(-3)0.1682831.0490580.1604130.8738CHOUSA(-4)0.0181380.6774530.0267730.9789R-squared0.030398 Mean dependent var5.420667Adjusted R-squared-0.124738 S.D. dependent var2.968564S.E. of regression3.148271 Akaike info criterion5.282595Sum squared resid247.7902 Schwarz criterion5.516128Log likelihood-74.23893 F-statistic0.195947Durbin-Watson stat0.409428 Prob(F-statistic)0.938187

Soluia ar fi estimat ecuaiilor diagramei de faze.