17/07/2024, 10:30 TestExercise 3.

ipynb - Colab

1
2 import pandas as pd
3 import numpy as np
4
5 df = pd.read_excel("TestExercise-3.xlsx")

1 df.head()

OBS INTRATE INFL PROD UNEMPL COMMPRI PCE PERSINC HOUST

0 1960:1 3.99 1.24095 10.03653 3.41845 7.95262 5.70962 1.68419 -11.88896

1 1960:2 3.97 1.41379 6.96248 3.46575 -8.55856 5.06452 1.33094 -9.83803

2 1960:3 3.84 1.51881 4.49681 2.71993 -16.83599 5.55733 0.89195 -31.54321

3 1960:4 3.92 1.93237 1.50624 2.79820 -5.03145 7.77351 0.67636 -18.93082

4 1960:5 3.85 1.82507 -0.11398 1.72552 -12.44240 4.39179 0.33667 -15.15354

a.) Solving for Part a

1 #Regressing INTRATE on all 7 variables and add a constant term too to X

2
3
4
5 y = df["INTRATE"]
6 X = df.drop(["INTRATE"], axis=1)
7 X = X.drop(["OBS"], axis=1)
8
9 #Append a constant term too to X
10 X = pd.DataFrame(np.ones(len(df)), columns=['const'])
11 X = pd.concat([X, df.drop(["INTRATE"], axis=1)], axis=1)
12 X = X.drop(["OBS"], axis=1)

1 X.head()

const INFL PROD UNEMPL COMMPRI PCE PERSINC HOUST

0 1.0 1.24095 10.03653 3.41845 7.95262 5.70962 1.68419 -11.88896

1 1.0 1.41379 6.96248 3.46575 -8.55856 5.06452 1.33094 -9.83803

2 1.0 1.51881 4.49681 2.71993 -16.83599 5.55733 0.89195 -31.54321

3 1.0 1.93237 1.50624 2.79820 -5.03145 7.77351 0.67636 -18.93082

4 1.0 1.82507 -0.11398 1.72552 -12.44240 4.39179 0.33667 -15.15354

1 # Using stats OLS method to regress

2
3 import statsmodels.api as sm
4 X = sm.add_constant(X)
5 est = sm.OLS(y, X).fit()
6 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.639
Model: OLS Adj. R-squared: 0.635
Method: Least Squares F-statistic: 164.5
Date: Wed, 17 Jul 2024 Prob (F-statistic): 1.64e-139
Time: 04:36:45 Log-Likelihood: -1449.2
No. Observations: 660 AIC: 2914.
Df Residuals: 652 BIC: 2950.
Df Model: 7
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const -0.2212 0.245 -0.903 0.367 -0.702 0.260
INFL 0.6961 0.062 11.185 0.000 0.574 0.818
PROD -0.0577 0.040 -1.447 0.148 -0.136 0.021
UNEMPL 0.1025 0.097 1.059 0.290 -0.088 0.292
COMMPRI -0.0055 0.003 -1.857 0.064 -0.011 0.000
PCE 0.3444 0.069 4.958 0.000 0.208 0.481
PERSINC 0.2470 0.061 4.077 0.000 0.128 0.366
HOUST -0.0194 0.005 -4.155 0.000 -0.029 -0.010
==============================================================================
Omnibus: 28.142 Durbin-Watson: 0.101
Prob(Omnibus): 0.000 Jarque-Bera (JB): 41.034
Skew: 0.365 Prob(JB): 1.23e-09

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 1/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
Kurtosis: 3.980 Cond. No. 102.
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 #Lets drop UNEMPL as it has the highest p-value among all the parameters and fit again
2
3 X = X.drop(["UNEMPL"], axis=1)
4 est = sm.OLS(y, X).fit()
5 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.638
Model: OLS Adj. R-squared: 0.635
Method: Least Squares F-statistic: 191.7
Date: Wed, 17 Jul 2024 Prob (F-statistic): 1.99e-140
Time: 04:36:59 Log-Likelihood: -1449.7
No. Observations: 660 AIC: 2913.
Df Residuals: 653 BIC: 2945.
Df Model: 6
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const -0.2909 0.236 -1.232 0.218 -0.754 0.173
INFL 0.6933 0.062 11.150 0.000 0.571 0.815
PROD -0.0255 0.026 -0.989 0.323 -0.076 0.025
COMMPRI -0.0065 0.003 -2.308 0.021 -0.012 -0.001
PCE 0.3686 0.066 5.618 0.000 0.240 0.497
PERSINC 0.2516 0.060 4.162 0.000 0.133 0.370
HOUST -0.0210 0.004 -4.760 0.000 -0.030 -0.012
==============================================================================
Omnibus: 21.820 Durbin-Watson: 0.104
Prob(Omnibus): 0.000 Jarque-Bera (JB): 30.851
Skew: 0.303 Prob(JB): 2.00e-07
Kurtosis: 3.868 Cond. No. 97.1
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Similarly lets drop PROD now and regress

2
3 X = X.drop(["PROD"], axis=1)
4 est = sm.OLS(y, X).fit()
5 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.637
Model: OLS Adj. R-squared: 0.635
Method: Least Squares F-statistic: 229.9
Date: Wed, 17 Jul 2024 Prob (F-statistic): 2.03e-141
Time: 04:37:04 Log-Likelihood: -1450.2
No. Observations: 660 AIC: 2912.
Df Residuals: 654 BIC: 2939.
Df Model: 5
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const -0.2401 0.230 -1.042 0.298 -0.692 0.212
INFL 0.7175 0.057 12.555 0.000 0.605 0.830
COMMPRI -0.0075 0.003 -2.841 0.005 -0.013 -0.002
PCE 0.3405 0.059 5.756 0.000 0.224 0.457
PERSINC 0.2402 0.059 4.048 0.000 0.124 0.357
HOUST -0.0205 0.004 -4.678 0.000 -0.029 -0.012
==============================================================================
Omnibus: 23.848 Durbin-Watson: 0.100
Prob(Omnibus): 0.000 Jarque-Bera (JB): 31.255
Skew: 0.354 Prob(JB): 1.63e-07
Kurtosis: 3.797 Cond. No. 94.1
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Now lets drop COMMPRI

2
3 X = X.drop(["COMMPRI"], axis=1)
4 est = sm.OLS(y, X).fit()
5 print(est.summary())

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 2/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
OLS Regression Results
==============================================================================
Dep. Variable: INTRATE R-squared: 0.633
Model: OLS Adj. R-squared: 0.631
Method: Least Squares F-statistic: 282.3
Date: Wed, 17 Jul 2024 Prob (F-statistic): 6.18e-141
Time: 04:37:10 Log-Likelihood: -1454.3
No. Observations: 660 AIC: 2919.
Df Residuals: 655 BIC: 2941.
Df Model: 4
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const -0.2136 0.231 -0.923 0.356 -0.668 0.241
INFL 0.7448 0.057 13.149 0.000 0.634 0.856
PCE 0.3110 0.059 5.311 0.000 0.196 0.426
PERSINC 0.2569 0.059 4.327 0.000 0.140 0.373
HOUST -0.0215 0.004 -4.893 0.000 -0.030 -0.013
==============================================================================
Omnibus: 27.399 Durbin-Watson: 0.100
Prob(Omnibus): 0.000 Jarque-Bera (JB): 33.853
Skew: 0.416 Prob(JB): 4.46e-08
Kurtosis: 3.733 Cond. No. 62.7
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

b.) Solving for Part b

1
2
3 # Start with just a constant
4 X = pd.DataFrame(np.ones(len(df)), columns=['const'])
5 est = sm.OLS(y, X).fit()
6 print(est.summary())
7
8 # Add INFLATION
9 X['INFL'] = df['INFL']
10 est = sm.OLS(y, X).fit()
11 print(est.summary())
12
13

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.000
Model: OLS Adj. R-squared: 0.000
Method: Least Squares F-statistic: nan
Date: Wed, 17 Jul 2024 Prob (F-statistic): nan
Time: 04:46:43 Log-Likelihood: -1784.9
No. Observations: 660 AIC: 3572.
Df Residuals: 659 BIC: 3576.
Df Model: 0
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 5.3476 0.141 37.958 0.000 5.071 5.624
==============================================================================
Omnibus: 90.501 Durbin-Watson: 0.022
Prob(Omnibus): 0.000 Jarque-Bera (JB): 141.949
Skew: 0.900 Prob(JB): 1.50e-31
Kurtosis: 4.388 Cond. No. 1.00
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
OLS Regression Results
==============================================================================
Dep. Variable: INTRATE R-squared: 0.560
Model: OLS Adj. R-squared: 0.559
Method: Least Squares F-statistic: 836.6
Date: Wed, 17 Jul 2024 Prob (F-statistic): 2.47e-119
Time: 04:46:43 Log-Likelihood: -1514.2
No. Observations: 660 AIC: 3032.
Df Residuals: 658 BIC: 3041.
Df Model: 1
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 1.6421 0.159 10.352 0.000 1.331 1.954
INFL 0.9453 0.033 28.925 0.000 0.881 1.010
==============================================================================

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 3/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
Omnibus: 5.019 Durbin-Watson: 0.063
Prob(Omnibus): 0.081 Jarque-Bera (JB): 4.841
Skew: 0.193 Prob(JB): 0.0889
Kurtosis: 3.166 Cond. No. 8.46
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # It has a good p-value so we keep it, now adding PROD variable column
2
3 X['PROD'] = df['PROD']
4 est = sm.OLS(y, X).fit()
5 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.575
Model: OLS Adj. R-squared: 0.573
Method: Least Squares F-statistic: 443.9
Date: Wed, 17 Jul 2024 Prob (F-statistic): 1.06e-122
Time: 04:46:45 Log-Likelihood: -1502.8
No. Observations: 660 AIC: 3012.
Df Residuals: 657 BIC: 3025.
Df Model: 2
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 1.2489 0.176 7.088 0.000 0.903 1.595
INFL 0.9750 0.033 29.785 0.000 0.911 1.039
PROD 0.0947 0.020 4.805 0.000 0.056 0.133
==============================================================================
Omnibus: 12.297 Durbin-Watson: 0.065
Prob(Omnibus): 0.002 Jarque-Bera (JB): 12.444
Skew: 0.326 Prob(JB): 0.00199
Kurtosis: 3.168 Cond. No. 11.9
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 #Even prod has a good p-value, we keep it too, now lets add UNEMPL
2
3 X['UNEMPL'] = df['UNEMPL']
4 est = sm.OLS(y, X).fit()
5 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.595
Model: OLS Adj. R-squared: 0.593
Method: Least Squares F-statistic: 321.7
Date: Wed, 17 Jul 2024 Prob (F-statistic): 2.16e-128
Time: 04:46:46 Log-Likelihood: -1486.4
No. Observations: 660 AIC: 2981.
Df Residuals: 656 BIC: 2999.
Df Model: 3
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 1.2041 0.172 6.994 0.000 0.866 1.542
INFL 0.8913 0.035 25.404 0.000 0.822 0.960
PROD -0.0799 0.036 -2.230 0.026 -0.150 -0.010
UNEMPL 0.4978 0.086 5.780 0.000 0.329 0.667
==============================================================================
Omnibus: 47.572 Durbin-Watson: 0.068
Prob(Omnibus): 0.000 Jarque-Bera (JB): 65.048
Skew: 0.580 Prob(JB): 7.50e-15
Kurtosis: 4.009 Cond. No. 12.8
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Even unempl has a good p-value, and prod has a p -value less than our threshold of 0.05, so lets keep it too
2 # Lets add COMMPRI
3
4 X['COMMPRI'] = df['COMMPRI']
5 est = sm.OLS(y, X).fit()
6 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.598

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 4/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
Model: OLS Adj. R-squared: 0.595
Method: Least Squares F-statistic: 243.3
Date: Wed, 17 Jul 2024 Prob (F-statistic): 6.16e-128
Time: 04:46:46 Log-Likelihood: -1484.5
No. Observations: 660 AIC: 2979.
Df Residuals: 655 BIC: 3001.
Df Model: 4
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 1.1943 0.172 6.949 0.000 0.857 1.532
INFL 0.9009 0.035 25.487 0.000 0.831 0.970
PROD -0.0415 0.041 -1.018 0.309 -0.121 0.039
UNEMPL 0.4349 0.092 4.742 0.000 0.255 0.615
COMMPRI -0.0061 0.003 -1.965 0.050 -0.012 -4.01e-06
==============================================================================
Omnibus: 42.015 Durbin-Watson: 0.066
Prob(Omnibus): 0.000 Jarque-Bera (JB): 56.210
Skew: 0.536 Prob(JB): 6.22e-13
Kurtosis: 3.946 Cond. No. 65.7
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Lets remove prod given its high p value and add PCE
2
3 X['PCE'] = df['PCE']
4 #Removing PROD
5 X = X.drop(["PROD"], axis=1)
6 est = sm.OLS(y, X).fit()
7 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.615
Model: OLS Adj. R-squared: 0.613
Method: Least Squares F-statistic: 262.0
Date: Wed, 17 Jul 2024 Prob (F-statistic): 2.53e-134
Time: 04:46:47 Log-Likelihood: -1469.6
No. Observations: 660 AIC: 2949.
Df Residuals: 655 BIC: 2972.
Df Model: 4
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 0.1872 0.240 0.781 0.435 -0.283 0.658
INFL 0.7297 0.046 16.010 0.000 0.640 0.819
UNEMPL 0.1391 0.060 2.329 0.020 0.022 0.256
COMMPRI -0.0100 0.003 -3.704 0.000 -0.015 -0.005
PCE 0.3064 0.055 5.588 0.000 0.199 0.414
==============================================================================
Omnibus: 18.195 Durbin-Watson: 0.074
Prob(Omnibus): 0.000 Jarque-Bera (JB): 20.267
Skew: 0.347 Prob(JB): 3.97e-05
Kurtosis: 3.504 Cond. No. 94.0
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Lets add persinc too

2
3 X['PERSINC'] = df['PERSINC']
4 est = sm.OLS(y, X).fit()
5 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.626
Model: OLS Adj. R-squared: 0.623
Method: Least Squares F-statistic: 219.1
Date: Wed, 17 Jul 2024 Prob (F-statistic): 4.01e-137
Time: 04:46:47 Log-Likelihood: -1460.2
No. Observations: 660 AIC: 2932.
Df Residuals: 654 BIC: 2959.
Df Model: 5
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 0.0635 0.238 0.267 0.790 -0.404 0.531
INFL 0.8534 0.053 16.040 0.000 0.749 0.958
UNEMPL 0.0787 0.061 1.300 0.194 -0.040 0.198
COMMPRI -0.0086 0.003 -3.217 0.001 -0.014 -0.003

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 5/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
PCE 0.1855 0.061 3.049 0.002 0.066 0.305
PERSINC 0.2658 0.061 4.349 0.000 0.146 0.386
==============================================================================
Omnibus: 12.857 Durbin-Watson: 0.083
Prob(Omnibus): 0.002 Jarque-Bera (JB): 13.800
Skew: 0.286 Prob(JB): 0.00101
Kurtosis: 3.418 Cond. No. 94.5
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Lets remove UNEMPL and add PERSINC

2
3 X = X.drop(["UNEMPL"], axis=1)
4 #Adding HOUST
5
6 X['HOUST'] = df['HOUST']
7 est = sm.OLS(y, X).fit()
8 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.637
Model: OLS Adj. R-squared: 0.635
Method: Least Squares F-statistic: 229.9
Date: Wed, 17 Jul 2024 Prob (F-statistic): 2.03e-141
Time: 04:46:47 Log-Likelihood: -1450.2
No. Observations: 660 AIC: 2912.
Df Residuals: 654 BIC: 2939.
Df Model: 5
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const -0.2401 0.230 -1.042 0.298 -0.692 0.212
INFL 0.7175 0.057 12.555 0.000 0.605 0.830
COMMPRI -0.0075 0.003 -2.841 0.005 -0.013 -0.002
PCE 0.3405 0.059 5.756 0.000 0.224 0.457
PERSINC 0.2402 0.059 4.048 0.000 0.124 0.357
HOUST -0.0205 0.004 -4.678 0.000 -0.029 -0.012
==============================================================================
Omnibus: 23.848 Durbin-Watson: 0.100
Prob(Omnibus): 0.000 Jarque-Bera (JB): 31.255
Skew: 0.354 Prob(JB): 1.63e-07
Kurtosis: 3.797 Cond. No. 94.1
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

b.) The models in a and b are quite different

c.) Solving for part c, computing the values for our models a and Taylors rule

1 # Getting the model from Part a and regressing it again, keeping onlt the INFL, PCE, PERSINC, HOUST parameters from the d
2
3 y = df["INTRATE"]
4 X = pd.DataFrame(np.ones(len(df)), columns=['const'])
5 X = pd.concat([X, df.drop(["INTRATE"], axis=1)], axis=1)
6 X = X.drop(["OBS"], axis=1)
7 X = X.drop(["PROD"], axis=1)
8 X = X.drop(["UNEMPL"], axis=1)
9 X = X.drop(["COMMPRI"], axis=1)
10
11
12 est = sm.OLS(y, X).fit()
13 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.633
Model: OLS Adj. R-squared: 0.631
Method: Least Squares F-statistic: 282.3
Date: Wed, 17 Jul 2024 Prob (F-statistic): 6.18e-141
Time: 04:51:52 Log-Likelihood: -1454.3
No. Observations: 660 AIC: 2919.
Df Residuals: 655 BIC: 2941.
Df Model: 4
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const -0.2136 0.231 -0.923 0.356 -0.668 0.241

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 6/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
INFL 0.7448 0.057 13.149 0.000 0.634 0.856
PCE 0.3110 0.059 5.311 0.000 0.196 0.426
PERSINC 0.2569 0.059 4.327 0.000 0.140 0.373
HOUST -0.0215 0.004 -4.893 0.000 -0.030 -0.013
==============================================================================
Omnibus: 27.399 Durbin-Watson: 0.100
Prob(Omnibus): 0.000 Jarque-Bera (JB): 33.853
Skew: 0.416 Prob(JB): 4.46e-08
Kurtosis: 3.733 Cond. No. 62.7
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

1 # Lets now regeress o the Taylor rule, which he chooses X as only a constant + INFL + PROD only
2
3 X = pd.DataFrame(np.ones(len(df)), columns=['const'])
4 X['INFL'] = df['INFL']
5 X['PROD'] = df['PROD']
6 est = sm.OLS(y, X).fit()
7 print(est.summary())

OLS Regression Results

==============================================================================
Dep. Variable: INTRATE R-squared: 0.575
Model: OLS Adj. R-squared: 0.573
Method: Least Squares F-statistic: 443.9
Date: Wed, 17 Jul 2024 Prob (F-statistic): 1.06e-122
Time: 04:52:57 Log-Likelihood: -1502.8
No. Observations: 660 AIC: 3012.
Df Residuals: 657 BIC: 3025.
Df Model: 2
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
const 1.2489 0.176 7.088 0.000 0.903 1.595
INFL 0.9750 0.033 29.785 0.000 0.911 1.039
PROD 0.0947 0.020 4.805 0.000 0.056 0.133
==============================================================================
Omnibus: 12.297 Durbin-Watson: 0.065
Prob(Omnibus): 0.002 Jarque-Bera (JB): 12.444
Skew: 0.326 Prob(JB): 0.00199
Kurtosis: 3.168 Cond. No. 11.9
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

c.) Hence the values of R squared, AIC and BIC are better for our model than the Taylors rule

d.) Solving for Part d

1
2
3 # RESET test
4 from statsmodels.stats.diagnostic import linear_reset
5 print("______________________________________________")
6 reset_test = linear_reset(est, power=2)
7 print(reset_test)
8 print("______________________________________________")
9
10 # Chow break test
11 from statsmodels.formula.api import ols
12 from statsmodels.stats.api import anova_lm
13 df['dummy'] = (df.index >= 20).astype(int)
14 model1 = ols('INTRATE ~ INFL + PROD', data=df).fit()
15 model2 = ols('INTRATE ~ INFL + PROD + dummy + INFL:dummy + PROD:dummy', data=df).fit()
16 chow_test = anova_lm(model1, model2)
17 print(chow_test)
18 print("______________________________________________")
19
20 # Forecast test
21 from statsmodels.tsa.stattools import adfuller
22 pre_break = df['INTRATE'][:20]
23 post_break = df['INTRATE'][20:]
24 adf_pre = adfuller(pre_break)
25 adf_post = adfuller(post_break)
26 print('Pre-break ADF statistic:', adf_pre[0])
27 print('Post-break ADF statistic:', adf_post[0])
28 print("______________________________________________")
29
30 # Jarque-Bera test
https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 7/8
17/07/2024, 10:30 TestExercise 3.ipynb - Colab
31 from statsmodels.stats.stattools import jarque_bera
32 jb_test = jarque_bera(est.resid)
33 print(jb_test)
34 print("______________________________________________")
35

______________________________________________
<Wald test (chi2): statistic=2.5371195394336548, p-value=0.11119747865833446, df_denom=1>
______________________________________________
df_resid ssr df_diff ss_diff F Pr(>F)
0 657.0 3671.394806 0.0 NaN NaN NaN
1 654.0 3669.745555 3.0 1.649251 0.097973 0.961138
______________________________________________
Pre-break ADF statistic: -0.89272514653163
Post-break ADF statistic: -2.603062206099372
______________________________________________
(12.444043308438582, 0.0019852277136483674, 0.32571817051428936, 3.1677538653763313)
______________________________________________

1 Start coding or generate with AI.

https://colab.research.google.com/drive/13b68CWylarwD28XhRqBUi5TJbyIUZQ-d?authuser=0#scrollTo=bT8oUQ0zCyBZ&printMode=true 8/8