Below are the results from the same linear error model described in Part I of this series of blogs, but this time the model takes more quintile variables into account:
n
20
R2 1.00
Adjusted R2 0.98
SE 3.366E+04
Term Coefficient 95% CI SE t statistic DF p
Intercept 21059554 -30137429 to 72256537 18439763 1.14 4 0.3171
Q1 Tax -37479 -155898 to 80939 42651 -0.88 4 0.4292
Q2 Tax 139401 6977 to 271824 47695 2.92 4 0.0431
Q3 Tax -81591 -440482 to 277299 129263 -0.63 4 0.5622
Q4 Tax -104546 -826096 to 617005 259883 -0.40 4 0.7080
Q5 Tax 33972 -204834 to 272779 86012 0.39 4 0.7130
Q1 ISB 14.28 -85.82 to 114.38 36.053 0.40 4 0.7122
Q2 ISB -9.003 -115.469 to 97.462 38.3461 -0.23 4 0.8259
Q3 ISB -61.86 -204.68 to 80.96 51.441 -1.20 4 0.2955
Q4 ISB -6.985 -64.307 to 50.338 20.6459 -0.34 4 0.7521
Q5 ISB -1.831 -50.935 to 47.273 17.6860 -0.10 4 0.9225
Q1 Spend -184300 -737154 to 368554 199123 -0.93 4 0.4071
Q2 Spend -310334 -1047012 to 426344 265331 -1.17 4 0.3071
Q3 Spend -172793 -735579 to 389993 202700 -0.85 4 0.4420
Q4 Spend -119840 -630579 to 390898 183954 -0.65 4 0.5503
Q5 Spend -226944 -722706 to 268818 178560 -1.27 4 0.2726
Source of variation Sum squares DF Mean square F statistic p
Model 1.120E+12 15 7.465E+10 65.88 0.0005
Residual 4.533E+09 4 1.133E+09
Total 1.124E+12 19
Coefficients Coefficient Value Value Trade Deficit (Oil) Ave Trade Deficit (Oil)
Intercept 21059554 1 2.11E+07 1 21059554
Q1 Tax -37479 -6.8 2.55E+05 -6.8 254857.2
Q2 Tax 139401 -0.04 -5.58E+03 -0.04 -5576.04
Q3 Tax -81591 3.3 -2.69E+05 3.3 -269250.3
Q4 Tax -104546 6.2 -6.48E+05 6.2 -648185.2
Q5 Tax 33972 17.5 5.95E+05 13 441636
Q1 ISB 14.28 1.98E+04 2.83E+05 1.98E+04 282744
Q2 ISB 9 1.96E+04 1.76E+05 1.96E+04 176458.8
Q3 ISB -61.86 2.10E+04 -1.30E+06 2.10E+04 -1299060
Q4 ISB -6.985 2.88E+04 -2.01E+05 2.88E+04 -201168
Q5 ISB -1.831 4.68E+04 -8.57E+04 5.08E+04 -93014.8
Q1 Spend -184300 9.00E+00 -1.66E+06 8.50E+00 -1566550
Q2 Spend -310334 13 -4.03E+06 12 -3724008
Q3 Spend -172793 17.5 -3.02E+06 16.5 -2851084.5
Q4 Spend -119840 23.5 -2.82E+06 23 -2756320
Q5 Spend -226944 36 -8.17E+06 40 -9077760
1.56E+05 -276726.84
Result 1.56E+11 -2.76727E+11
The above model has perfect correlation as shown by the R² variable equal to 1. The results are vastly different to yesterday’s result with the annual national trade deficit varying from a gain of 156 billion to a loss of 277 billion dollars annually. These result variations are dependent on changing the fifth quintiles (highlighted in red) effective tax rate from 13 to 17.5%, its income from 46,800 to 50,800 dollars, and its consumer spending rate from 36 to 40%. Remember, spending between all the quintiles must add up to 100%. Hence, if spending for the fifth quintile is lowered, then spending must go up for the other quintiles by the same percentage and vice versa. These values were updated accordingly in the above table. For instance, when spending for the fifth quintile was reduced from 38 to 36%, spending for the first quintile was increased from 8.8 to 9%, spending for the second quintile was increased from 12.6 to 13%, spending for the third quintile was increased from 16.9 to 17.5%, and spending for the fourth quintile was increased from 23.2% to 23.5%. The results are highly dependent on the fifth quintiles spending percentage. The bottom line from this analysis: if the fifth quintile has less money to spend, it does not import as many products thus the trade deficit goes down.
In summary, it is interesting how two similar models show vastly different results: one showing that increasing taxes increases the national trade deficit level and another one showing increasing taxes lowers the national trade deficit level. The models obtain these same trends if the trade deficit is adjusted for oil imports. Both results make sense. On one hand, higher taxes could lower business productivity and they therefore have few products to export but on the other hand, higher taxes also means businesses and individuals have less money to buy foreign products. Thus, these two analysis probably offset meaning that United States’ tax levels, income levels, national debt, social benefits, and spending levels do not affect the trade deficit as much as one may suspect.
My Book: Is America Dying? (Amazon.com, Barnes and Noble)
No comments:
Post a Comment