UNIVERSIDAD NACIONAL DE PIURA

FACULTAD DE ECONOMIA
DPTO. ACAD. DE ECONOMIA

SOLUCIÓN DE LA PRIMERA PRACTICA CALIFICADA DE ECONOMETRIA I

1º El investigador especifica los modelos siguientes:

MODELO 1: E(t) = a + b PBI(t) + c ITC(t) + d E(t-1) + u(t)
MODELO 2: E(t) = a + b PBI(t) + c(0) ITC(t) + c(1) ITC(t-1) + .... + c(K) ITC(t-7) + u(t)
MODELO 3: E(t) = a + b TI(t) + c(0) PBI(t) + c(1) PBI(t-1) + c(2) PBI(t-2) + .... + u(t)
se le pide:

1.1. Estimar el modelo 1 y realizar el test de Hausman y Wu. (5 puntos)

Dependent Variable: E
Method: Least Squares
Simple(adjusted): 1907 2005
Included observations: 99 after adjusting endpoints

Variable Coefficient Std. Error t-Statistic Prob.

C 19.54581 42.04352 0.464895 0.6431

PBI -0.011714 0.004159 -2.816522 0.0059
ITC -8.37E-07 2.95E-07 -2.840818 0.0055
E(-1) 0.476559 0.089115 5.347705 0.0000

R-squared 0.622925 Mean dependent var -164.4242

Adjusted R-squared 0.611018 S.D. dependent var 556.1871
S.E. of regression 346.8853 Akaike info criterion 14.57543
Sum squared resid 11431295 Schwarz criterion 14.68028
Log likelihood -717.4838 F-statistic 52.31313
Durbin-Watson stat 1.283669 Prob(F-statistic) 0.000000

MOD1T = @regobs = 99
MOD1VB4 = MOD1.@coefcov(4,4) = 0.007941
MOD1RHO = 1-(@DW/2) = 0.358166
MOD1H = MOD1RHO*SQR(MOD1T/(1-MOD1T*MOD1VB4)) = 7.707219

Sample: 1907 2005

Included observations: 99

Autocorrelation Partial Correlation AC PAC Q-Stat Prob

. |*** | . |*** | 1 0.332 0.332 11.241 0.001

***| . | ****| . | 2 -0.399 -0.572 27.629 0.000

mod2qbp1 = 99*(0.33192184296^2) = 10.90704

mod2qbp2 = 99*(0.33192184296^2+0.398721112116^2) = 26.64591

Breusch-Godfrey Serial Correlation LM Test:

F-statistic 65.58821 Probability 0.000000

Obs*R-squared 40.68742 Probability 0.000000

Method: Least Squares

Variable Coefficient Std. Error t-Statistic Prob.

 2

C 5.738039 32.44622 0.176848 0.8600

PBI -0.012757 0.003575 -3.568763 0.0006
ITC -9.17E-07 2.54E-07 -3.612604 0.0005
E(-1) -0.867859 0.127322 -6.816265 0.0000
RESID(-1) 1.238195 0.152889 8.098655 0.0000

F-statistic 40.83433 Probability 0.000000

Obs*R-squared 46.28877 Probability 0.000000

Method: Least Squares

Variable Coefficient Std. Error t-Statistic Prob.

C -14.55367 31.67852 -0.459418 0.6470

PBI 0.001215 0.005606 0.216667 0.8289
ITC -2.14E-07 3.30E-07 -0.648156 0.5185
E(-1) -0.080961 0.278330 -0.290883 0.7718
RESID(-1) 0.665539 0.233538 2.849811 0.0054
RESID(-2) -0.618619 0.196783 -3.143665 0.0022

Dependent Variable: E
Method: Two-Stage Least Squares
Sample(adjusted): 1907 2005
Included observations: 99 after adjusting endpoints
Instrument list: C PBI PBI(-1) ITC ITC(-1)

Variable Coefficient Std. Error t-Statistic Prob.

C 11.41820 49.17558 0.232193 0.8169

PBI -0.017875 0.005231 -3.417033 0.0009
ITC -1.24E-06 3.67E-07 -3.381644 0.0010
E(-1) -0.047722 0.195194 -0.244485 0.8074

R-squared 0.485542 Mean dependent var -164.4242

Adjusted R-squared 0.469296 S.D. dependent var 556.1871
S.E. of regression 405.1792 Sum squared resid 15596164
F-statistic 31.37612 Durbin-Watson stat 0.977493
Prob(F-statistic) 0.000000

BMCO VARMCO
R1 19.54581 R1 1767.658 -0.087653 3.10E-06 0.123111
R2 -0.011714 R2 -0.087653 1.73E-05 -7.56E-10 9.33E-05
R3 -8.37E-07 R3 3.10E-06 -7.56E-10 8.68E-14 6.11E-09
R4 0.476559 R4 0.123111 9.33E-05 6.11E-09 0.007941

BMC2E VARMC2E
R1 11.41820 R1 2418.237 -0.114621 4.56E-06 0.590650
R2 -0.017875 R2 -0.114621 2.74E-05 -7.85E-10 0.000448
R3 -1.24E-06 R3 4.56E-06 -7.85E-10 1.35E-13 2.93E-08
R4 -0.047722 R4 0.590650 0.000448 2.93E-08 0.038101
 3
1.2. Estimar el modelo 2 y obtener el retardo medio y el retardo mediano. (3 puntos)

2 3 4 5 6
Adjusted
R- 0.517918754766 0.529178926242 0.523910785375 0.530801797183 0.577602548415
squared
Akaike 14.8598393335 14.8462814916 14.8673531719 14.8625825326 14.767174864
Schwarz 14.9960005966 15.0096750073 15.0579789402 15.0804405535 15.0122651376
Prob
0.456920651149 0.0815835033398 0.847298899505 0.136198412704 0.00178299972056
Retardo

Dependent Variable: E
Method: Least Squares
Sample(adjusted): 1913 2005
Included observations: 93 after adjusting endpoints

Variable Coefficient Std. Error t-Statistic Prob.

C 40.73764 48.15298 0.846004 0.4000

PBI -0.035790 0.006830 -5.239941 0.0000
PDL01 0.000338 0.000194 1.738953 0.0857
PDL02 0.008555 0.002613 3.274647 0.0015
PDL03 0.001973 0.000574 3.439131 0.0009
PDL04 -0.011116 0.003480 -3.194337 0.0020
PDL05 -0.003574 0.001061 -3.369602 0.0011
PDL06 0.002117 0.000672 3.151992 0.0022
PDL07 0.000702 0.000217 3.227233 0.0018

R-squared 0.614333 Mean dependent var -175.3366

Adjusted R-squared 0.577603 S.D. dependent var 572.3040
S.E. of regression 371.9525 Akaike info criterion 14.76717
Sum squared resid 11621285 Schwarz criterion 15.01227
Log likelihood -677.6736 F-statistic 16.72554
Durbin-Watson stat 0.978924 Prob(F-statistic) 0.000000

Lag Distribution of
ITC i Coefficient Std. Error t-Statistic

* | 0 -2.1E-05 1.1E-05 -1.94311

* | 1 5.4E-06 2.3E-05 0.23416
* | 2 -0.00012 7.7E-05 -1.53352
* | 3 0.00034 0.00019 1.73895
* | 4 -0.00101 0.00054 -1.86740
* | 5 -0.00812 0.00232 -3.50545
.* | 6 0.48011 0.15428 3.11189
. *| 7 3.48170 1.10320 3.15601

Sum of
Lags 3.95289 1.25484 3.15011

mod2rm = @sum(@trend*mod2b)/@sum(mod2b) = 6.883213

mod2rmed = 6.501480748

1.3. Estimar el modelo 3 y obtener los multiplicadores. (5 puntos)

 4
E(t) = a + b TI(t) + c0 PBI(t) + c1 PBI(t-1) + c2 PBI(t-2)+ .... + u(t)
(1)

Asumimos ci = (1-d)di
E(t) = a +b TI(t) +(1-d) PBI(t) +(1-d)d PBI(t-1) + (1-d)d2 PBI(t-2)+ .... + u(t)

Retardando un periodo:
E(t-1) = a + b TI(t-1) + (1-d) PBI(t-1) + (1-d)d PBI(t-2) + (1-d)d2 PBI(t-
3)+ .... + u(t-1)

Multiplica por d:
d E(t-1) = ad + bd TI(t-1) - (1-d)d PBI(t-1) + (1-d)d2 PBI(t-2) + (1-d)d3
PBI(t-3)+ .... + d u(t-1) (2)

(1) - (2):
E(t) – d E(t-1) = a(1-d) + b TI(t) – bd TI(t-1) + (1-d) PBI(t) + u(t) – d
u(t-1)

E(t) = a(1-d) + b TI(t) – bd TI(t-1) + (1-d) PBI(t) + d E(t-1) + (u(t) - d

u(t-1))

Dependent Variable: E
Method: Least Squares
Sample(adjusted): 1907 2005
Included observations: 99 after adjusting endpoints

Variable Coefficient Std. Error t-Statistic Prob.

C 116.5552 112.7472 1.033775 0.3039

TI 263.1346 196.3648 1.340030 0.1835
TI(-1) -326.5420 192.8870 -1.692919 0.0938
PBI -0.019484 0.003440 -5.663792 0.0000
E(-1) 0.543078 0.089532 6.065711 0.0000

R-squared 0.603174 Mean dependent var -164.4242

Adjusted R-squared 0.586288 S.D. dependent var 556.1871
S.E. of regresión 357.7419 Akaike info criterion 14.64669
Sum squared resid 12030054 Schwarz criterion 14.77775
Log likelihood -720.0110 F-statistic 35.71997
Durbin-Watson stat 1.155248 Prob(F-statistic) 0.000000

mod3rho = 1-(mod3.@dw/2) = 0.422376

mod3vb5 = mod3.@coefcov(5,5) = 0.008016
mod3h = mod3rho*sqr(mod3t/(1-mod3t*mod3vb5)) = 9.250211

Sample: 1907 2005

Included observations: 99

Autocorrelation Partial Correlation AC PAC Q-Stat Prob

. |*** | . |*** | 1 0.372 0.372 14.116 0.000

**| . | ****| . | 2 -0.274 -0.479 21.866 0.000

mod3qbp1 = 99*(0.371951643241^2) = 13.69645

mod3qbp2 = 99*(0.371951643241^2+0.274205287725^2) = 21.14012
 5
Breusch-Godfrey Serial Correlation LM Test:

F-statistic 70.75174 Probability 0.000000

Obs*R-squared 42.77464 Probability 0.000000

Dependent Variable: RESID

Method: Least Squares

Variable Coefficient Std. Error t-Statistic Prob.

C 60.78727 85.72855 0.709067 0.4801

TI -193.0420 150.5363 -1.282362 0.2029
TI(-1) 167.4921 147.4919 1.135602 0.2590
PBI -0.020064 0.003533 -5.678788 0.0000
E(-1) -0.742978 0.111372 -6.671146 0.0000
RESID(-1) 1.161622 0.138101 8.411405 0.0000

Breusch-Godfrey Serial Correlation LM Test:

F-statistic 35.83480 Probability 0.000000

Obs*R-squared 43.35130 Probability 0.000000

Dependent Variable: RESID

Method: Least Squares

Variable Coefficient Std. Error t-Statistic Prob.

C 34.47789 89.88398 0.383582 0.7022

TI -175.6975 151.6183 -1.158815 0.2495
TI(-1) 162.5079 147.6172 1.100874 0.2738
PBI -0.014043 0.007107 -1.976042 0.0511
E(-1) -0.542186 0.233881 -2.318211 0.0227
RESID(-1) 1.035849 0.188876 5.484275 0.0000
RESID(-2) -0.191849 0.196488 -0.976395 0.3314

Dependent Variable: E
Method: Two-Stage Least Squares
Sample(adjusted): 1908 2005
Included observations: 98 after adjusting endpoints
Instrument list: C PBI PBI(-1) TI TI(-1) TI(-2)

Variable Coefficient Std. Error t-Statistic Prob.

C 78.16515 160.6264 0.486627 0.6277

TI 228.1495 276.3367 0.825622 0.4111
TI(-1) -242.7078 272.8327 -0.889585 0.3760
PBI -0.035650 0.006950 -5.129107 0.0000
E(-1) -0.304078 0.290417 -1.047038 0.2978

R-squared 0.224517 Mean dependent var -166.1490

Adjusted R-squared 0.191162 S.D. dependent var 558.7805
S.E. of regression 502.5415 Sum squared resid 23486958
F-statistic 13.68572 Durbin-Watson stat 0.753281
 6
Prob(F-statistic) 0.000000

MOD3MITI = C(2) = 228.1495

MOD3MIPBI = C(4) = -0.035650

MOD3D1RTI = C(3) + C(2)*C(5) = -183.639293352

MOD3D2RTI = C(3)*C(5) + C(2)*C(5)^2 = -99.7304841599

MOD3D1RPBI = C(4)*C(5) = -0.0105811374876

MOD3D2RPBI = C(4)*C(5)^2 = -0.00574638436765

MOD3MTTI = (C(3) + C(2))/(1-C(5)) = -138.770595905

MOD3MTPBI = C(4)/(1-C(5)) = -0.0426410704112

2º Comente y fundamente su respuesta. (7 puntos)

2.1. El modelo aleatorio uniecuacional carece de supuesto caeteris paribus.

2.2. Para que un modelo dinámico sea estable las raíces características del polinomio de retardo
deben ser en valor absoluto mayor que la unidad.