Remembering that:
ln T = ln b0 + b11 ln(77 )
(2)
ln TI = ln b0 + b12ln(y )
(3)
ε = b11 ■ b12 = 0,959262 * 0,980447 = 0, 9405055
Our results show that the long run elasticity of income tax is 0. 94, which is slightly less than unity. Our next step is
to estimate equation (4) and investigate whether there is a long run relationship for this equation in which income
tax elasticity could be also obtained.
The results of the cointegration analysis for equation (4) are given in Table 6.
Table 6: Johansen Trace Test and Maximum Eigenvalue Test Results ( For Ln(realT) and Ln(realY) )
|
Trace test | |||
|
Null (H0) Hypothesis |
Alternative (H1) Hypothesis |
Test Statistics |
%5 Critical Value |
|
r=0 |
r≥1 |
22,68904* |
20,261 |
|
r≤1 |
r≥2 |
2,3484527 |
9,1645 |
|
Maximum Eigenvalue Test | |||
|
Null (H0) Hypothesis |
Alternative (H1) Hypothesis |
Test Statistics |
%5 Critical Value |
|
r=0 |
r=1 |
19,2045* |
15,8921 |
|
r≤1 |
r=2 |
2,3484527 |
9,1645 |
* Denotes the rejection of the null hypothesis at the 0,05 level
Johansen trace and maximum eigenvalue tests indicate that real income tax revenue and real income are
cointegrated with one cointegrating vector. The optimal lag number is also one in this system.
The normalized cointegrating vector for equation (4) is found as :
Ln Real T = -1,8632 + 0,953928 Ln Real Y
(1,8333) (0,16254)
The coefficient of real GNP, which is the long run elasticity of income tax, is statistically significant at %5 level.
Comparing this elasticity to the one we have computed earlier, we see that:
ln 7 = ln b 0 + b13 ln(Γ )
(4)
10