Scribd Logo
Upload

Welcome to Scribd!

  • 7 views

    Use of LOG in Regression Model

    Uploaded by

    Aiman Mufeez
    The document discusses four regression models used to analyze the relationship between GDP and several independent variables: FDI, FLF, and INF. The four models are: 1) a standard linear model with untransformed variables, 2) a double log model with all variables in log form, 3) a semi-log model with the dependent variable in log form and independent variables untransformed, and 4) a semi-log model with the dependent variable untransformed and independent variables in log form. Each model provides coefficients that indicate the effect of a one unit or percentage change in the independent variables on GDP.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    Use of LOG in Regression Model

    Uploaded by

    Aiman Mufeez
    7 views5 pages
    The document discusses four regression models used to analyze the relationship between GDP and several independent variables: FDI, FLF, and INF. The four models are: 1) a standard linear model with untransformed variables, 2) a double log model with all variables in log form, 3) a semi-log model with the dependent variable in log form and independent variables untransformed, and 4) a semi-log model with the dependent variable untransformed and independent variables in log form. Each model provides coefficients that indicate the effect of a one unit or percentage change in the independent variables on GDP.

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses four regression models used to analyze the relationship between GDP and several independent variables: FDI, FLF, and INF. The four models are: 1) a standard linear model with untransformed variables, 2) a double log model with all variables in log form, 3) a semi-log model with the dependent variable in log form and independent variables untransformed, and 4) a semi-log model with the dependent variable untransformed and independent variables in log form. Each model provides coefficients that indicate the effect of a one unit or percentage change in the independent variables on GDP.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    7 views5 pages

    Use of LOG in Regression Model

    Uploaded by

    Aiman Mufeez
    The document discusses four regression models used to analyze the relationship between GDP and several independent variables: FDI, FLF, and INF. The four models are: 1) a standard linear model with untransformed variables, 2) a double log model with all variables in log form, 3) a semi-log model with the dependent variable in log form and independent variables untransformed, and 4) a semi-log model with the dependent variable untransformed and independent variables in log form. Each model provides coefficients that indicate the effect of a one unit or percentage change in the independent variables on GDP.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    of 5

    Use of LOG in regression Model

    Conditions for use of log


    1) No Log Model:
    GDP = α + β FDI +β FLF +β INF + e
    Variable Coefficient
    C 1.252290
    FDI 0.682627
    FLF 0.131388
    INF 0.077046

    Interpretation:
    GDP = 1.252290 + 0.682627 FDI + 0.131388 FLF + 0.077046 INF
     If FDI increase by 1 unit then GDP will increase by 0.682627 unit while
    keeping other factor constant.
     If FLF increase by 1 unit then GDP will increase by 0.131388 unit while
    keeping other factor constant.
     If INF increase by 1 unit then GDP will increase by 0.077046 unit while
    keeping other factor constant.

    2) Double Log / Log- Log Model:


    Log (GDP) = α + β log (FDI) +β log
    (FLF) +β log Variable Coefficient (INF) + e
    C 0.624806
    log (FDI) 0.115864
    log (FLF) 0.367544
    log (INF) -0.019632

    Interpretation:
    Log (GDP) = 0.624806 + 0.115864 log (FDI) + 0.367544 log (FLF) - 0.019632 log
    (INF)
     If FDI increase by 1% then GDP will increase by 0.115864 % while keeping
    other factor constant.
     If FLF increase by 1% then GDP will increase by 0.367544 % while keeping
    other factor constant.
     If INF increase by 1% then GDP will decrease by 0.019632 % while keeping
    other factor constant.

    3) Semi Log / Log-Lin Model:


    Log (GDP) = α + β FDI +β FLF +β
    INF + e Variable Coefficient
    C 1.047362
    FDI 0.158824
    FLF 0.018038
    INF 0.013042

    Interpretation:
    Log (GDP) = 1.047362 + 0.158824 FDI + 0.018038 FLF + 0.013042 INF
     If FDI increase by 1 unit then GDP will increase by 0.158824 % while
    keeping other factor constant.
     If FLF increase by 1unit then GDP will increase by 0.018038 % while
    keeping other factor constant.
     If INF increase by 1unit then GDP will decrease by 0.013042 % while
    keeping other factor constant.

    4) Semi Log / Lin-Log Model:


    GDP = α + β log (FDI) +β log
    (FLF) +β log Variable Coefficient (INF) + e
    C -3.584964
    log (FDI) 0.523321
    log (FLF) 2.954766
    log (INF) -0.029517

    Interpretation:
    GDP = -3.584964 + 0.523321 log (FDI) + 2.954766 log (FLF) - 0.029517 log
    (INF)
     If FDI increase by 1% then GDP will increase by 0.523321 unit while
    keeping other factor constant.
     If FLF increase by 1% then GDP will increase by 2.954766 unit while
    keeping other factor constant.
     If INF increase by 1% then GDP will decrease by 0.029517unit while
    keeping other factor constant.

    Appendix
    1) No Log Model:
    Dependent Variable: GDP
    Method: Least Squares
    Date: 12/01/23 Time: 19:32
    Sample: 1996 2020
    Included observations: 25

    Variable Coefficient Std. Error t-Statistic Prob.

    C 1.252290 1.786665 0.700910 0.4911


    FDI 0.682627 0.549309 1.242702 0.2277
    FLF 0.131388 0.067309 1.951997 0.0644
    INF 0.077046 0.107344 0.717750 0.4808

    R-squared 0.387069 Mean dependent var 5.756247


    Adjusted R-squared 0.299507 S.D. dependent var 1.132301
    S.E. of regression 0.947685 Akaike info criterion 2.876056
    Sum squared resid 18.86023 Schwarz criterion 3.071077
    Log likelihood -31.95071 Hannan-Quinn criter. 2.930147
    F-statistic 4.420526 Durbin-Watson stat 1.573751
    Prob(F-statistic) 0.014696
    2) Double Log / Log- Log Model:
    Dependent Variable: LOG(GDP)
    Method: Least Squares
    Date: 12/01/23 Time: 20:38
    Sample: 1996 2020
    Included observations: 25

    Variable Coefficient Std. Error t-Statistic Prob.

    C 0.624806 1.212342 0.515371 0.6117


    LOG(FDI) 0.115864 0.059080 1.961155 0.0633
    LOG(FLF) 0.367544 0.352675 1.042159 0.3092
    LOG(INF) -0.019632 0.115613 -0.169809 0.8668

    R-squared 0.393977 Mean dependent var 1.730433


    Adjusted R-squared 0.307402 S.D. dependent var 0.207391
    S.E. of regression 0.172596 Akaike info criterion -0.530076
    Sum squared resid 0.625579 Schwarz criterion -0.335056
    Log likelihood 10.62595 Hannan-Quinn criter. -0.475986
    F-statistic 4.550714 Durbin-Watson stat 1.581114
    Prob(F-statistic) 0.013141

    3) Semi Log / Log-Lin Model:


    Dependent Variable: LOG(GDP)
    Method: Least Squares
    Date: 12/01/23 Time: 21:30
    Sample: 1996 2020
    Included observations: 25

    Variable Coefficient Std. Error t-Statistic Prob.

    C 1.047362 0.331371 3.160696 0.0047


    FDI 0.158824 0.101880 1.558939 0.1340
    FLF 0.018038 0.012484 1.444893 0.1632
    INF 0.013042 0.019909 0.655080 0.5195

    R-squared 0.371513 Mean dependent var 1.730433


    Adjusted R-squared 0.281730 S.D. dependent var 0.207391
    S.E. of regression 0.175766 Akaike info criterion -0.493680
    Sum squared resid 0.648767 Schwarz criterion -0.298660
    Log likelihood 10.17100 Hannan-Quinn criter. -0.439589
    F-statistic 4.137866 Durbin-Watson stat 1.565507
    Prob(F-statistic) 0.018805
    4) Semi Log / Lin-Log Model:
    Dependent Variable: GDP
    Method: Least Squares
    Date: 12/01/23 Time: 21:12
    Sample: 1996 2020
    Included observations: 25

    Variable Coefficient Std. Error t-Statistic Prob.

    C -3.584964 6.555737 -0.546844 0.5903


    LOG(FDI) 0.523321 0.319473 1.638079 0.1163
    LOG(FLF) 2.954766 1.907090 1.549359 0.1362
    LOG(INF) -0.029517 0.625178 -0.047214 0.9628

    R-squared 0.405516 Mean dependent var 5.756247


    Adjusted R-squared 0.320590 S.D. dependent var 1.132301
    S.E. of regression 0.933314 Akaike info criterion 2.845497
    Sum squared resid 18.29258 Schwarz criterion 3.040517
    Log likelihood -31.56871 Hannan-Quinn criter. 2.899587
    F-statistic 4.774922 Durbin-Watson stat 1.593322
    Prob(F-statistic) 0.010866

    You might also like