Notesale: Turn your study into money

Document Preview

Extracts from the notes are below, to see the PDF you'll receive please use the links above

---

Stata Commands
1
...

2
...
sum y x1 x2, detail: summarize in the detail, show the summary, and also the percentiles, kurtosis,
skewness, and variance of the data
...
gen lny=log(y): this command is used to make values small, by taking log, and then again summarize
them
...


5
...
corr y x1 x2: to see the correlation of the variables
...


7
...


8
...


9
...


Interpretation of y variable for β2X1
Our regression analyses explaining that, independent variable x1 is showing significant positive impact on
dependent variable
...
By assuming other things constant,
a one unit increase in X1 will lead to increase 92741
...


Interpretation of y variable for β2X2
Our regression analysis explain that, the independent variable x2 is showing significant negative impact

on dependent variable
...
002 which is less than 0
...
A one unit increase in x2 will lead to decrease -46359
...

Rule of Thumb
Rule of thumb of t statistics is when t value is 2 or more than 2 then we can say that it is significant at 5%
level
...
It also explain that total variation in dependent
variable Y explained by the regression model
...
90, 0
...

Adjusted R square value is always less than R square value
...

As increase in the variables of model the value of R square increases
...
dufller y: this command is used for to take dicky-fuller test
...
It is one of the most commonly used statistical test when it comes to
analyzing the stationary of a series
...
00 compare with the significance level
at 1%=0
...
05, and 10%=0
...

Make the hypothesis, null hypothesis is that variable is not stationary, If p-value is less than the
significance level of 1,5,10 percent, then we will reject the null the null hypothesis which means variable
is stationary
...
But in Y variable p-value is greater than
the significance level, which means it is not stationary
...
dfuller D
...


As we can see that the p-value is now smaller than the significance level 10%
...
0944<0
...
If p-value is stationary without difference then
we will say that order of Integration is zero I(0)
...

12
...
1 If there is no difference taken in the model so we will use the simple OLS model for the variables

(reg y x1 x2)
...
2 If all variables are starting at I(1), then we will use the Johanson cointegration approach
...
If cointegration exists than we use the VECM model/approach, and cointegration do not exist than
we use the VAR model/approach
...
If order of integration is mixed then we use the ARDL model
...
predict resid: command is used for the assume residual
...

For Residual Error First we take Regression of dependent or independent variable
...
dfuller resid: command is used for the see the residual is stationary or not
...

And if not which means residual is not stationary and have no long-run relationship
...
For Graph: kdensity y, normal

18
...

We can use the vif command after the regression to check for multicollinearity
...
First we use regression and then vif
...
Tolerance, defined as 1/VIF, is used by many
researchers to check on the degree of collinearity
...
1 is comparable to a
VIF of 10
...


VIF value is less than 10 so its independent variables are not correlated
...

19
...
The most common method of test
autocorrelation is the Durbin-Watson test
...
The Durbin-Watson always produces a
test number range from 0 to 2
...

To correct the autocorrelation problem and for Durbin-Watson test prais y x1 x2, corc

D-W value is 0
...
There is no autocorrelation now
...
Then we will use
Vector Error Correction Model (VECM)
...


Varsoc y x1 x2: To see the lags

Vecrank y x1 x2, lag(4):

In this we will see the trace statistic value which should be greater than 5% critical value
...
11 > 5% critical value 29
...

Vec y x1 x2, rank(1): In this we put all our variable as dependent variable
...




---