Modeling and
forecasting on EVIEWS
TRAINING PROVIDED BY NASREDDINE DRIDI 2021
STATISTICAL ENGINEER
DIRECTOR OF MACROECONOMIC FORECASTS
Main forecasting steps
I. Preparing and processing the required data
II. Elaboration of linear equations (sectoral value added, production, intermediate
consumption) and exploiting them for forecasting
III. Estimate the pool data equations (IPI idustrual sectoriel production indicator)
IV. Estimate the equations system
V. Time Series: Univariate Modeling (CPI) and Forecasting
VI. Targeting ECModels for long-term and short-term prediction
NASREDDINE DRIDI : TRAINING 2021 2
I. Preparing and processing the required data
Date Variable1 Variable2
D1 V11 V21
D2 V12 V22
D3 V13 V31
D4 V14 V41
Date Individu Variable1 Variable2
D1 1 V111 V211
D2 1 V112 V212
D3 1 V113 V311
D1 2 V121 V221
D2 2 V122 V222
D3 2 V123 V223
D1 3 V131 V231
D2 3 V132 V232
D3 3 V133 V233
Data for estimate Pool Data
Data for estimate Time series
NASREDDINE DRIDI : TRAINING 2021 3
I. Preparing and processing the
required data
NASREDDINE DRIDI : TRAINING 2021 4
I. Preparing and processing the
required data
NASREDDINE DRIDI : TRAINING 2021 5
I. Preparing and processing the
required data
 For Annual Data : 1983 : 2021
 For semestriel Data : 1983S1 : 2021S2
 For trimestriel Data : 1983T1 : 2021T2
 For mensuel Data : start date 1983:1 End date 2021:12
 For weekly Data : nombre of month : nombre of day : year exp: 08:10:1983
NASREDDINE DRIDI : TRAINING 2021 6
I. Preparing and processing the
required data
 Exemple
NASREDDINE DRIDI : TRAINING 2021 7
I. Preparing and processing the
required data
 Import Data : Proc/Import/Read text-lotus-Excel
NASREDDINE DRIDI : TRAINING 2021 8
I. Preparing and processing the
required data
 Import Data : Proc/Import/Read text-lotus-Excel
NASREDDINE DRIDI : TRAINING 2021 9
I. Preparing and processing the
required data
 Convert data from annual to quarterly : TMM Money market rate
NASREDDINE DRIDI : TRAINING 2021 10
I. Preparing and processing the
required data
 Convert data from annual to quarterly
NASREDDINE DRIDI : TRAINING 2021 11
I. Preparing and processing the
required data
 Convert data from ammual to quarterly
NASREDDINE DRIDI : TRAINING 2021 12
I. Preparing and processing the
required data
 Convert data from ammual to quarterly
NASREDDINE DRIDI : TRAINING 2021 13
‫البيانات‬ ‫تحويل‬
‫إلى‬ ‫السنوية‬
‫أي‬ ‫ربعية‬
‫من‬ ‫تحويل‬
‫األقل‬ ‫التكرار‬
‫التكرار‬ ‫إلى‬
‫األعلى‬
I. Preparing and processing the
required data
 Convert data from ammual to quarterly
NASREDDINE DRIDI : TRAINING 2021 14
I. Preparing and processing the
required data
 Generate a new variable
NASREDDINE DRIDI : TRAINING 2021 15
II. Estmation of linear equations
 Estimate Equation: Quick/Estimate Equation
NASREDDINE DRIDI : TRAINING 2021 16
II. Estmation of linear equations : Multiple regression
 Estimate Equation: Resids
NASREDDINE DRIDI : TRAINING 2021 17
II. Estmation of linear equations : Multiple regression
 Residual Test: Pas d’autocorrelation des résidus
NASREDDINE DRIDI : TRAINING 2021 18
II. Estmation of linear equations : Multiple regression
 white Test: Pas d’autocorrelation des résidus P-value>5% we accepte the homoscedasticity
hypothesis H0
NASREDDINE DRIDI : TRAINING 2021 19
II. Estmation of linear equations : Multiple regression
 Breush-Godfrey Test: Pas d’autocorrelation des résidus
P-value>5% we don’t reject the no correlation
hypothesis H0
NASREDDINE DRIDI : TRAINING 2021 20
II. Estmation of linear equations : Multiple regression
 Test of Multicollinearity :
NASREDDINE DRIDI : TRAINING 2021 21
II. Estmation of linear equations : Multiple regression
 Test of Multicollinearity :
NASREDDINE DRIDI : TRAINING 2021 22
II. Estmation of linear equations : Multiple regression
 Test of Multicollinearity : No multicolinearity ‫منخفض‬ ‫المتعدد‬ ‫الخطي‬ ‫التداخل‬
NASREDDINE DRIDI : TRAINING 2021 23
II. Estmation of linear equations : Multiple regression
 Reset test of Ramsey: H0 :the model is well specified vs H1 : the model is incorrectly specified
P-value>5% we don’t reject the H0 hypothesis
NASREDDINE DRIDI : TRAINING 2021 24
II. Estmation of linear equations : Multiple regression
 Stability test of Chow: H0 :the model is stable vs H1 : the model is not stable
P-value<5% 2001 is a brekpoint
NASREDDINE DRIDI : TRAINING 2021 25
II. Estmation of linear equations : Multiple regression
 Forecast quality:
NASREDDINE DRIDI : TRAINING 2021 26
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Create a pool data project
NASREDDINE DRIDI : TRAINING 2021 27
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Create a pool data project
we must introduce the individual
codes
NASREDDINE DRIDI : TRAINING 2021 28
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Create a pool data project
NASREDDINE DRIDI : TRAINING 2021 29
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Import pool data
NASREDDINE DRIDI : TRAINING 2021 30
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Import pool data
NASREDDINE DRIDI : TRAINING 2021 31
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Import pool data
NASREDDINE DRIDI : TRAINING 2021 32
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Data representation
NASREDDINE DRIDI : TRAINING 2021 33
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Unit Root Test
NASREDDINE DRIDI : TRAINING 2021 34
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Descriptive Statistics
NASREDDINE DRIDI : TRAINING 2021 35
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 𝑦𝑖𝑡 = ∁ + 𝛼𝑖 + 𝛽𝑖𝑡𝑋𝑖𝑡 + 𝜀𝑖𝑡
cte Indivual effect Vector of estimators Vector of variables residue
 Random individual effect:
𝑦𝑖𝑡 = ∁ + 𝛽𝑖𝑡𝑋𝑖𝑡 + 𝛼𝑖 + 𝜀𝑖𝑡
𝑈𝑖𝑡 = 𝛼𝑖 + 𝜀𝑖𝑡
 Fixe individual effect:
𝑦𝑖𝑡 = ∁ + 𝛼𝑖 + 𝛽𝑖𝑡𝑋𝑖𝑡 + 𝜀𝑖𝑡
NASREDDINE DRIDI : TRAINING 2021 36
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Pool Estimation
Estimation
Pool
Random Individual effect
NASREDDINE DRIDI : TRAINING 2021 37
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Pool Estimation : Random Individual Effect
NASREDDINE DRIDI : TRAINING 2021 38
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Pool Estimation : Fixed Individual Effect
NASREDDINE DRIDI : TRAINING 2021 39
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Hausman Test: Random Individual Effect
NASREDDINE DRIDI : TRAINING 2021 40
We accept H0: Random
individuel effect
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Random Individual Effect
NASREDDINE DRIDI : TRAINING 2021 41
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Make system : Fixed Individual Effect
NASREDDINE DRIDI : TRAINING 2021 42
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 Pool Estimation : Fixe Individual Effect
NASREDDINE DRIDI : TRAINING 2021 43
III. Estimate the pool data equations (IPI idustrual
sectoriel production indicator)
 IPI forecasting
NASREDDINE DRIDI : TRAINING 2021 44
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 View/Correlogram
Existance of Unit root
NASREDDINE DRIDI : TRAINING 2021 45
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 View/Correlogram
NASREDDINE DRIDI : TRAINING 2021 46
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 View/Correlogram
Unit Root
NASREDDINE DRIDI : TRAINING 2021 47
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 View/Correlogram
NASREDDINE DRIDI : TRAINING 2021 48
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Process is identified from the functions of autocorrelation and partial autocorrelation
 AR(p) process : the partial
autocorrélation of ordre p+1 is null
 MA(q) process : The partial
autocorrélation of ordre q+1 is null
 The possible process are AR(1),
MA(1), ARMA(1,1)
NASREDDINE DRIDI : TRAINING 2021 49
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Estimation of ARMA(1,1) process
the
component
is not
significant
NASREDDINE DRIDI : TRAINING 2021 50
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Estimation of AR(1) process
NASREDDINE DRIDI : TRAINING 2021 51
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Make Risidual Test : unit Root Test
NASREDDINE DRIDI : TRAINING 2021 52
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Make Risidual Test : unit Root Test
NASREDDINE DRIDI : TRAINING 2021 53
We rejet H0 hypothesis:
Resid are stationnary
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Forecast : Box-jenkins
NASREDDINE DRIDI : TRAINING 2021 54
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Forecast
NASREDDINE DRIDI : TRAINING 2021 55
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Forecast : Non parametric method :Lissage expononciel Double: Holt –winters non saisonnière
NASREDDINE DRIDI : TRAINING 2021 56
IV. Time Series: Univariate Modeling (CPI) and
Forecasting
 Exemple : Consumption price Index (CPI)
 Forecast : Non parametric method :Lissage expononciel Double: Holt –winters non saisonnière
IPC 2020
ECModel 6,5
Box-Jenkis 4,48
Holt-winters 6,68
NASREDDINE DRIDI : TRAINING 2021 57
V. Time series: multivariate study
 Causality Test : CPI-M3-NEER
NASREDDINE DRIDI : TRAINING 2021 58
V. Time series: multivariate study
 Causality Test : View/Granger Causality
NASREDDINE DRIDI : TRAINING 2021 59
We accept the
hypothesis of
causality from
M3 to IPC
V. Time series: multivariate study
 Model VAR: Objects/New Objecs/VAR
NASREDDINE DRIDI : TRAINING 2021 60
V. Time series: multivariate study
 Vector Autoregression Estimate
NASREDDINE DRIDI : TRAINING 2021 61
V. Time series: multivariate study
 View/Lag Structure/Block Exogeneity Tests
NASREDDINE DRIDI : TRAINING 2021 62
NEER
Exogenous
variable
V. Time series: multivariate study
 Model VAR: VAR Specification
NASREDDINE DRIDI : TRAINING 2021 63
VI. Targeting ECModels for long-term and short-term
prediction
 Data Base
NASREDDINE DRIDI : TRAINING 2021 64
VI. Targeting ECModels for long-term and short-term
prediction
 Modeling 3 equations with ECModels : 1.producer price index 2.PCI 3.private consumption
NASREDDINE DRIDI : TRAINING 2021 65
VI. Targeting ECModels for long-term and short-term
prediction
 Cointegration Test : Test de johanson
NASREDDINE DRIDI : TRAINING 2021 66
VI. Targeting ECModels for long-term and short-term
prediction
 Cointegration Test : Test de johanson
Existence of at least two cointegration relations
NASREDDINE DRIDI : TRAINING 2021 67
VI. Targeting ECModels for long-term and short-term
prediction
 Cointegration Test : Estimation of VECM(1) Model
NASREDDINE DRIDI : TRAINING 2021 68
𝐷 𝑌𝑡 = 𝐷 𝑋𝑡 + σ(𝑌𝑡−1 − 𝑋𝑡−1)
𝜎 < 0
𝐷 𝑌𝑡 = 𝐷 𝑋𝑡 + σ𝜀𝑡−1
VI. Targeting ECModels for long-term and short-term
prediction
 Cointegration Test : Estimation of VECM(1) Model
NASREDDINE DRIDI : TRAINING 2021 69
the long-
term
relation
the Short-term
relation
the return to long-term
equilibrium takes 14 month
VI. Targeting ECModels for long-term and short-term
prediction
 Cointegration Test : Test of Ljung-Box
NASREDDINE DRIDI : TRAINING 2021 70
VI. Targeting ECModels for long-term and short-term
prediction
 Cointegration Test : Test of Ljung-Box
NASREDDINE DRIDI : TRAINING 2021 71
VI. Targeting ECModels for long-term and short-term
prediction
 Estimate ECModel : Cointegration Price Idex equation
Force de rappelle
NASREDDINE DRIDI : TRAINING 2021 72
VI. Targeting ECModels for long-term and short-term
prediction
 Estimate ECModel : Cointegration real wage equation
Force de rappelle
NASREDDINE DRIDI : TRAINING 2021 73
VI. Targeting ECModels for long-term and short-term
prediction
 simultaneous equation system
NASREDDINE DRIDI : TRAINING 2021 74
VI. Targeting ECModels for long-term and short-term
prediction
 simultaneous equation system
NASREDDINE DRIDI : TRAINING 2021 75
VI. Targeting ECModels for long-term and short-term
prediction
 simultaneous equation system : exp Production Function CES
NASREDDINE DRIDI : TRAINING 2021 76
Bloc à estimer pour chaque secteur.
Stock de capital du secteur i:
ki
t = yi
t ― s (cki
t ― pyi
t) + (s―1) gki.t + k0
Emploi total dans le secteur i:
li
t = yi
t + s (wi
t ― pyi
t) + (s ― 1)gli.t + l0
Prix de la valeur ajoutée du secteur i:
pyi
t = π0 (cki
t ― gki.t) + (1― π0) (wi
t ― gli.t) + py0
VI. Targeting ECModels for long-term and short-term
prediction
 simultaneous equation system : The estimators
NASREDDINE DRIDI : TRAINING 2021 77
VI. Targeting ECModels for long-term and short-term
prediction
 simultaneous equation system : All variables of Model
NASREDDINE DRIDI : TRAINING 2021 78
Inflation
Forecast
VI. Targeting ECModels for long-term and short-term
prediction
 choice of hypothesis
NASREDDINE DRIDI : TRAINING 2021 79
Thanks For Attention
nasreddine.dridi@yahoo.fr
NASREDDINE DRIDI : TRAINING 2021 80

EVIEWS_1418112021.pdf

  • 1.
    Modeling and forecasting onEVIEWS TRAINING PROVIDED BY NASREDDINE DRIDI 2021 STATISTICAL ENGINEER DIRECTOR OF MACROECONOMIC FORECASTS
  • 2.
    Main forecasting steps I.Preparing and processing the required data II. Elaboration of linear equations (sectoral value added, production, intermediate consumption) and exploiting them for forecasting III. Estimate the pool data equations (IPI idustrual sectoriel production indicator) IV. Estimate the equations system V. Time Series: Univariate Modeling (CPI) and Forecasting VI. Targeting ECModels for long-term and short-term prediction NASREDDINE DRIDI : TRAINING 2021 2
  • 3.
    I. Preparing andprocessing the required data Date Variable1 Variable2 D1 V11 V21 D2 V12 V22 D3 V13 V31 D4 V14 V41 Date Individu Variable1 Variable2 D1 1 V111 V211 D2 1 V112 V212 D3 1 V113 V311 D1 2 V121 V221 D2 2 V122 V222 D3 2 V123 V223 D1 3 V131 V231 D2 3 V132 V232 D3 3 V133 V233 Data for estimate Pool Data Data for estimate Time series NASREDDINE DRIDI : TRAINING 2021 3
  • 4.
    I. Preparing andprocessing the required data NASREDDINE DRIDI : TRAINING 2021 4
  • 5.
    I. Preparing andprocessing the required data NASREDDINE DRIDI : TRAINING 2021 5
  • 6.
    I. Preparing andprocessing the required data  For Annual Data : 1983 : 2021  For semestriel Data : 1983S1 : 2021S2  For trimestriel Data : 1983T1 : 2021T2  For mensuel Data : start date 1983:1 End date 2021:12  For weekly Data : nombre of month : nombre of day : year exp: 08:10:1983 NASREDDINE DRIDI : TRAINING 2021 6
  • 7.
    I. Preparing andprocessing the required data  Exemple NASREDDINE DRIDI : TRAINING 2021 7
  • 8.
    I. Preparing andprocessing the required data  Import Data : Proc/Import/Read text-lotus-Excel NASREDDINE DRIDI : TRAINING 2021 8
  • 9.
    I. Preparing andprocessing the required data  Import Data : Proc/Import/Read text-lotus-Excel NASREDDINE DRIDI : TRAINING 2021 9
  • 10.
    I. Preparing andprocessing the required data  Convert data from annual to quarterly : TMM Money market rate NASREDDINE DRIDI : TRAINING 2021 10
  • 11.
    I. Preparing andprocessing the required data  Convert data from annual to quarterly NASREDDINE DRIDI : TRAINING 2021 11
  • 12.
    I. Preparing andprocessing the required data  Convert data from ammual to quarterly NASREDDINE DRIDI : TRAINING 2021 12
  • 13.
    I. Preparing andprocessing the required data  Convert data from ammual to quarterly NASREDDINE DRIDI : TRAINING 2021 13 ‫البيانات‬ ‫تحويل‬ ‫إلى‬ ‫السنوية‬ ‫أي‬ ‫ربعية‬ ‫من‬ ‫تحويل‬ ‫األقل‬ ‫التكرار‬ ‫التكرار‬ ‫إلى‬ ‫األعلى‬
  • 14.
    I. Preparing andprocessing the required data  Convert data from ammual to quarterly NASREDDINE DRIDI : TRAINING 2021 14
  • 15.
    I. Preparing andprocessing the required data  Generate a new variable NASREDDINE DRIDI : TRAINING 2021 15
  • 16.
    II. Estmation oflinear equations  Estimate Equation: Quick/Estimate Equation NASREDDINE DRIDI : TRAINING 2021 16
  • 17.
    II. Estmation oflinear equations : Multiple regression  Estimate Equation: Resids NASREDDINE DRIDI : TRAINING 2021 17
  • 18.
    II. Estmation oflinear equations : Multiple regression  Residual Test: Pas d’autocorrelation des résidus NASREDDINE DRIDI : TRAINING 2021 18
  • 19.
    II. Estmation oflinear equations : Multiple regression  white Test: Pas d’autocorrelation des résidus P-value>5% we accepte the homoscedasticity hypothesis H0 NASREDDINE DRIDI : TRAINING 2021 19
  • 20.
    II. Estmation oflinear equations : Multiple regression  Breush-Godfrey Test: Pas d’autocorrelation des résidus P-value>5% we don’t reject the no correlation hypothesis H0 NASREDDINE DRIDI : TRAINING 2021 20
  • 21.
    II. Estmation oflinear equations : Multiple regression  Test of Multicollinearity : NASREDDINE DRIDI : TRAINING 2021 21
  • 22.
    II. Estmation oflinear equations : Multiple regression  Test of Multicollinearity : NASREDDINE DRIDI : TRAINING 2021 22
  • 23.
    II. Estmation oflinear equations : Multiple regression  Test of Multicollinearity : No multicolinearity ‫منخفض‬ ‫المتعدد‬ ‫الخطي‬ ‫التداخل‬ NASREDDINE DRIDI : TRAINING 2021 23
  • 24.
    II. Estmation oflinear equations : Multiple regression  Reset test of Ramsey: H0 :the model is well specified vs H1 : the model is incorrectly specified P-value>5% we don’t reject the H0 hypothesis NASREDDINE DRIDI : TRAINING 2021 24
  • 25.
    II. Estmation oflinear equations : Multiple regression  Stability test of Chow: H0 :the model is stable vs H1 : the model is not stable P-value<5% 2001 is a brekpoint NASREDDINE DRIDI : TRAINING 2021 25
  • 26.
    II. Estmation oflinear equations : Multiple regression  Forecast quality: NASREDDINE DRIDI : TRAINING 2021 26
  • 27.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Create a pool data project NASREDDINE DRIDI : TRAINING 2021 27
  • 28.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Create a pool data project we must introduce the individual codes NASREDDINE DRIDI : TRAINING 2021 28
  • 29.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Create a pool data project NASREDDINE DRIDI : TRAINING 2021 29
  • 30.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Import pool data NASREDDINE DRIDI : TRAINING 2021 30
  • 31.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Import pool data NASREDDINE DRIDI : TRAINING 2021 31
  • 32.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Import pool data NASREDDINE DRIDI : TRAINING 2021 32
  • 33.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Data representation NASREDDINE DRIDI : TRAINING 2021 33
  • 34.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Unit Root Test NASREDDINE DRIDI : TRAINING 2021 34
  • 35.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Descriptive Statistics NASREDDINE DRIDI : TRAINING 2021 35
  • 36.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  𝑦𝑖𝑡 = ∁ + 𝛼𝑖 + 𝛽𝑖𝑡𝑋𝑖𝑡 + 𝜀𝑖𝑡 cte Indivual effect Vector of estimators Vector of variables residue  Random individual effect: 𝑦𝑖𝑡 = ∁ + 𝛽𝑖𝑡𝑋𝑖𝑡 + 𝛼𝑖 + 𝜀𝑖𝑡 𝑈𝑖𝑡 = 𝛼𝑖 + 𝜀𝑖𝑡  Fixe individual effect: 𝑦𝑖𝑡 = ∁ + 𝛼𝑖 + 𝛽𝑖𝑡𝑋𝑖𝑡 + 𝜀𝑖𝑡 NASREDDINE DRIDI : TRAINING 2021 36
  • 37.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Pool Estimation Estimation Pool Random Individual effect NASREDDINE DRIDI : TRAINING 2021 37
  • 38.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Pool Estimation : Random Individual Effect NASREDDINE DRIDI : TRAINING 2021 38
  • 39.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Pool Estimation : Fixed Individual Effect NASREDDINE DRIDI : TRAINING 2021 39
  • 40.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Hausman Test: Random Individual Effect NASREDDINE DRIDI : TRAINING 2021 40 We accept H0: Random individuel effect
  • 41.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Random Individual Effect NASREDDINE DRIDI : TRAINING 2021 41
  • 42.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Make system : Fixed Individual Effect NASREDDINE DRIDI : TRAINING 2021 42
  • 43.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  Pool Estimation : Fixe Individual Effect NASREDDINE DRIDI : TRAINING 2021 43
  • 44.
    III. Estimate thepool data equations (IPI idustrual sectoriel production indicator)  IPI forecasting NASREDDINE DRIDI : TRAINING 2021 44
  • 45.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  View/Correlogram Existance of Unit root NASREDDINE DRIDI : TRAINING 2021 45
  • 46.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  View/Correlogram NASREDDINE DRIDI : TRAINING 2021 46
  • 47.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  View/Correlogram Unit Root NASREDDINE DRIDI : TRAINING 2021 47
  • 48.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  View/Correlogram NASREDDINE DRIDI : TRAINING 2021 48
  • 49.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Process is identified from the functions of autocorrelation and partial autocorrelation  AR(p) process : the partial autocorrélation of ordre p+1 is null  MA(q) process : The partial autocorrélation of ordre q+1 is null  The possible process are AR(1), MA(1), ARMA(1,1) NASREDDINE DRIDI : TRAINING 2021 49
  • 50.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Estimation of ARMA(1,1) process the component is not significant NASREDDINE DRIDI : TRAINING 2021 50
  • 51.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Estimation of AR(1) process NASREDDINE DRIDI : TRAINING 2021 51
  • 52.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Make Risidual Test : unit Root Test NASREDDINE DRIDI : TRAINING 2021 52
  • 53.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Make Risidual Test : unit Root Test NASREDDINE DRIDI : TRAINING 2021 53 We rejet H0 hypothesis: Resid are stationnary
  • 54.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Forecast : Box-jenkins NASREDDINE DRIDI : TRAINING 2021 54
  • 55.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Forecast NASREDDINE DRIDI : TRAINING 2021 55
  • 56.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Forecast : Non parametric method :Lissage expononciel Double: Holt –winters non saisonnière NASREDDINE DRIDI : TRAINING 2021 56
  • 57.
    IV. Time Series:Univariate Modeling (CPI) and Forecasting  Exemple : Consumption price Index (CPI)  Forecast : Non parametric method :Lissage expononciel Double: Holt –winters non saisonnière IPC 2020 ECModel 6,5 Box-Jenkis 4,48 Holt-winters 6,68 NASREDDINE DRIDI : TRAINING 2021 57
  • 58.
    V. Time series:multivariate study  Causality Test : CPI-M3-NEER NASREDDINE DRIDI : TRAINING 2021 58
  • 59.
    V. Time series:multivariate study  Causality Test : View/Granger Causality NASREDDINE DRIDI : TRAINING 2021 59 We accept the hypothesis of causality from M3 to IPC
  • 60.
    V. Time series:multivariate study  Model VAR: Objects/New Objecs/VAR NASREDDINE DRIDI : TRAINING 2021 60
  • 61.
    V. Time series:multivariate study  Vector Autoregression Estimate NASREDDINE DRIDI : TRAINING 2021 61
  • 62.
    V. Time series:multivariate study  View/Lag Structure/Block Exogeneity Tests NASREDDINE DRIDI : TRAINING 2021 62 NEER Exogenous variable
  • 63.
    V. Time series:multivariate study  Model VAR: VAR Specification NASREDDINE DRIDI : TRAINING 2021 63
  • 64.
    VI. Targeting ECModelsfor long-term and short-term prediction  Data Base NASREDDINE DRIDI : TRAINING 2021 64
  • 65.
    VI. Targeting ECModelsfor long-term and short-term prediction  Modeling 3 equations with ECModels : 1.producer price index 2.PCI 3.private consumption NASREDDINE DRIDI : TRAINING 2021 65
  • 66.
    VI. Targeting ECModelsfor long-term and short-term prediction  Cointegration Test : Test de johanson NASREDDINE DRIDI : TRAINING 2021 66
  • 67.
    VI. Targeting ECModelsfor long-term and short-term prediction  Cointegration Test : Test de johanson Existence of at least two cointegration relations NASREDDINE DRIDI : TRAINING 2021 67
  • 68.
    VI. Targeting ECModelsfor long-term and short-term prediction  Cointegration Test : Estimation of VECM(1) Model NASREDDINE DRIDI : TRAINING 2021 68 𝐷 𝑌𝑡 = 𝐷 𝑋𝑡 + σ(𝑌𝑡−1 − 𝑋𝑡−1) 𝜎 < 0 𝐷 𝑌𝑡 = 𝐷 𝑋𝑡 + σ𝜀𝑡−1
  • 69.
    VI. Targeting ECModelsfor long-term and short-term prediction  Cointegration Test : Estimation of VECM(1) Model NASREDDINE DRIDI : TRAINING 2021 69 the long- term relation the Short-term relation the return to long-term equilibrium takes 14 month
  • 70.
    VI. Targeting ECModelsfor long-term and short-term prediction  Cointegration Test : Test of Ljung-Box NASREDDINE DRIDI : TRAINING 2021 70
  • 71.
    VI. Targeting ECModelsfor long-term and short-term prediction  Cointegration Test : Test of Ljung-Box NASREDDINE DRIDI : TRAINING 2021 71
  • 72.
    VI. Targeting ECModelsfor long-term and short-term prediction  Estimate ECModel : Cointegration Price Idex equation Force de rappelle NASREDDINE DRIDI : TRAINING 2021 72
  • 73.
    VI. Targeting ECModelsfor long-term and short-term prediction  Estimate ECModel : Cointegration real wage equation Force de rappelle NASREDDINE DRIDI : TRAINING 2021 73
  • 74.
    VI. Targeting ECModelsfor long-term and short-term prediction  simultaneous equation system NASREDDINE DRIDI : TRAINING 2021 74
  • 75.
    VI. Targeting ECModelsfor long-term and short-term prediction  simultaneous equation system NASREDDINE DRIDI : TRAINING 2021 75
  • 76.
    VI. Targeting ECModelsfor long-term and short-term prediction  simultaneous equation system : exp Production Function CES NASREDDINE DRIDI : TRAINING 2021 76 Bloc à estimer pour chaque secteur. Stock de capital du secteur i: ki t = yi t ― s (cki t ― pyi t) + (s―1) gki.t + k0 Emploi total dans le secteur i: li t = yi t + s (wi t ― pyi t) + (s ― 1)gli.t + l0 Prix de la valeur ajoutée du secteur i: pyi t = π0 (cki t ― gki.t) + (1― π0) (wi t ― gli.t) + py0
  • 77.
    VI. Targeting ECModelsfor long-term and short-term prediction  simultaneous equation system : The estimators NASREDDINE DRIDI : TRAINING 2021 77
  • 78.
    VI. Targeting ECModelsfor long-term and short-term prediction  simultaneous equation system : All variables of Model NASREDDINE DRIDI : TRAINING 2021 78 Inflation Forecast
  • 79.
    VI. Targeting ECModelsfor long-term and short-term prediction  choice of hypothesis NASREDDINE DRIDI : TRAINING 2021 79
  • 80.