Econometrics Case

1.1) INCOME, ASSET, WAGE, LABOR have much explanation on GO in HCMC, while in Hanoi, if the variables affect individually on GO, the level of explaining will be slighter. For instance, 16.19% of INCOME explain GO in HCMC, meanwhile in Hanoi, the figure is 2.53%. WAGE and LABOR are the same with 16.19% and 18.69% respectively in HCMC, when 2.51% is the figure for both in Hanoi. In the capital, it seems that the 4 variables do not have so much effect on GO: very little percent in the variables determine the output.

1.2) Much of value added determines output. In Hanoi, 97% VA explain for GO, while in HCMC, the figure is nearly 100%. Gross Output seems to be determined much by Value Added and these 2 figures have very tight relationship.

1.3) In HCMC, when VA combines with other variables, it usually gives high coefficient of determination. Most of the figures are between 99% and 99.5%, only RATIOKL is the variable that cannot create a significant and no mistake model when going with VA. VA can go together with 4 in 5 other variables to create models explain much for GO in HCMC, however, in Hanoi, everything is opposite. Although VA still has much determination on GO, it cannot incorporate with other variables to explain more for GO. In the other way of talking, VA determine much for GO only when individually working, and the variable cannot work collectively with others (always creating overall insignificant model).

1.4) When it comes to 2 variables go with one another, in HCMC, except VA, INCOME, WAGE, ASSET and LABOR cannot incorporate, always creating insignificant model or showing Multicolinearity at a very high level (P-value of t-test always stands at the level nearby 1, much larger than 0.05). Therefore, to explain GO, only 1 independent variable can work effectively and should not combine with each other. The case is similar in Hanoi, but P-value of t-test of the same models in Hanoi just fluctuate between 0.1 to 0.7 or 0.8, which means that Multicolinearity exists in Hanoi's figures at high level, but still lower than HCMC. Exceptionally, Hanoi has one pair that can combine to create significant and "clean" model (INCOME and WAGE) which have 7% of independent variables explain GO. 7% is not a high number, but at least it shows the capability of incorporating of variables in Hanoi, a little bit better than HCMC.

1.5) RATIOKL does not explain GO much, or we can even say that it has no determination in the output. In HCMC, only 0.025% of RATIOKL has the strength of determination, while in Hanoi, model Ls GO c RATIOKL is even insignificant.

From 7 quantitative variables, we can build a lot of regressions with different results which are useful for your research. However, problems are the things we always have to face and to ensure that we can avoid meeting them, maybe using small models with few variables is the smart choice, at least with the given information.