Greetings,
I am Eun-Son Lim, currently employed as an assistant professor in South Korea. My research focuses on investigating the relationship between energy consumption and GDP, specifically examining how it's influenced by the percentage of internet users. I employed a Fixed Panel Threshold regression model for this analysis.
However, I've encountered a perplexing issue with the results. The values of the threshold points appear to be descending instead of ascending, which is counterintuitive. For example, Th-21 is at 75, while Th-22 is at 44. It seems to me that the order should be the opposite, with Th-21 at 44 and Th-22 at 75.
This descending sequence of threshold values has left me uncertain about how to interpret the regression results. The panel regression results are presented below:
xthreg GDP , rx(ENGC) qx(INUR) thnum(2) grid(400) trim(0.01, 0.01) bs(0 300)
Threshold estimator (level = 95):
-----------------------------------------------------
model | Threshold Lower Upper
-----------+-----------------------------------------
Th-1 | 75.0000 73.5100 75.0900
Th-21 | 75.0000 73.5100 75.0900
Th-22 | 44.0100 43.1500 44.1300
-----------------------------------------------------
Threshold effect test (bootstrap = 0 300):
-------------------------------------------------------------------------------
Threshold | RSS MSE Fstat Prob Crit10 Crit5 Crit1
-----------+-------------------------------------------------------------------
Single | . . . . . . .
Double | 3.14e+12 8.54e+09 42.04 0.0200 29.6744 34.4889 51.2095
-------------------------------------------------------------------------------
Fixed-effects (within) regression Number of obs = 384
Group variable: country Number of groups = 24
R-sq: within = 0.3544 Obs per group: min = 16
between = 0.9007 avg = 16.0
overall = 0.8838 max = 16
F(3,357) = 65.32
corr(u_i, Xb) = 0.8585 Prob > F = 0.0000
------------------------------------------------------------------------------
GDP | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_cat#c.ENGC |
0 | 185.6713 67.52431 2.75 0.006 52.87587 318.4667
1 | 277.6014 68.59508 4.05 0.000 142.7002 412.5026
2 | 368.6861 69.27706 5.32 0.000 232.4437 504.9285
|
_cons | 391619.5 69216.9 5.66 0.000 255495.4 527743.6
-------------+----------------------------------------------------------------
sigma_u | 635886.15
sigma_e | 93839.56
rho | .9786864 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(23, 357) = 143.59 Prob > F = 0.0000
*Variables: GDP, ENGC(Energy Consumption), INUR( % of Internet users)
Now, I'm trying to determine which interpretation is correct:
Interpretation 1: If the percentage of internet users is less than 44%, the coefficient for the effects of energy consumption on GDP is 186.67. If it falls between 44% and 75%, the coefficient is 277.60, and if it exceeds 75%, the coefficient is 368.68.
Interpretation 2: If the percentage of internet users exceeds 75%, the coefficient for the effects of energy consumption on GDP is 186.67. If it falls between 44% and 75%, the coefficient is 277.60, and if it's less than 44%, the coefficient is 368.68.
If both of these interpretations are incorrect, I would greatly appreciate your assistance in understanding how to correctly interpret these results.
Sincerely,
Eun-Son
I am Eun-Son Lim, currently employed as an assistant professor in South Korea. My research focuses on investigating the relationship between energy consumption and GDP, specifically examining how it's influenced by the percentage of internet users. I employed a Fixed Panel Threshold regression model for this analysis.
However, I've encountered a perplexing issue with the results. The values of the threshold points appear to be descending instead of ascending, which is counterintuitive. For example, Th-21 is at 75, while Th-22 is at 44. It seems to me that the order should be the opposite, with Th-21 at 44 and Th-22 at 75.
This descending sequence of threshold values has left me uncertain about how to interpret the regression results. The panel regression results are presented below:
xthreg GDP , rx(ENGC) qx(INUR) thnum(2) grid(400) trim(0.01, 0.01) bs(0 300)
Threshold estimator (level = 95):
-----------------------------------------------------
model | Threshold Lower Upper
-----------+-----------------------------------------
Th-1 | 75.0000 73.5100 75.0900
Th-21 | 75.0000 73.5100 75.0900
Th-22 | 44.0100 43.1500 44.1300
-----------------------------------------------------
Threshold effect test (bootstrap = 0 300):
-------------------------------------------------------------------------------
Threshold | RSS MSE Fstat Prob Crit10 Crit5 Crit1
-----------+-------------------------------------------------------------------
Single | . . . . . . .
Double | 3.14e+12 8.54e+09 42.04 0.0200 29.6744 34.4889 51.2095
-------------------------------------------------------------------------------
Fixed-effects (within) regression Number of obs = 384
Group variable: country Number of groups = 24
R-sq: within = 0.3544 Obs per group: min = 16
between = 0.9007 avg = 16.0
overall = 0.8838 max = 16
F(3,357) = 65.32
corr(u_i, Xb) = 0.8585 Prob > F = 0.0000
------------------------------------------------------------------------------
GDP | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_cat#c.ENGC |
0 | 185.6713 67.52431 2.75 0.006 52.87587 318.4667
1 | 277.6014 68.59508 4.05 0.000 142.7002 412.5026
2 | 368.6861 69.27706 5.32 0.000 232.4437 504.9285
|
_cons | 391619.5 69216.9 5.66 0.000 255495.4 527743.6
-------------+----------------------------------------------------------------
sigma_u | 635886.15
sigma_e | 93839.56
rho | .9786864 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(23, 357) = 143.59 Prob > F = 0.0000
*Variables: GDP, ENGC(Energy Consumption), INUR( % of Internet users)
Now, I'm trying to determine which interpretation is correct:
Interpretation 1: If the percentage of internet users is less than 44%, the coefficient for the effects of energy consumption on GDP is 186.67. If it falls between 44% and 75%, the coefficient is 277.60, and if it exceeds 75%, the coefficient is 368.68.
Interpretation 2: If the percentage of internet users exceeds 75%, the coefficient for the effects of energy consumption on GDP is 186.67. If it falls between 44% and 75%, the coefficient is 277.60, and if it's less than 44%, the coefficient is 368.68.
If both of these interpretations are incorrect, I would greatly appreciate your assistance in understanding how to correctly interpret these results.
Sincerely,
Eun-Son
