代写ECN5221 Final MOCK SET B帮做R编程

Question 1

The following estimated ARDL (2,1,0,1) model describes the determinants of financial development (LFD), where three important determinants are trade openness (LTO), Interest Rate (LR), Real Income (LY). The estimated result is as follows:

a. Based on the above result, compute the long-run coefficients of LTOt, LRt and LYt.

b. What are the advantages of ARDL model?

Question 2

a. Explain the possible biasness of an estimate that omits one true variable from the model? (Hint: suppose that the omitted variable’s coefficient is B3 )

b. In your opinion, which one has more severe negative implications to the econometric modelling, i. Omitting true variable from the model or ii. Adding unnecessary variables to a model? Explain.

Question 3

a. Explain what is unobserved heterogeneity in panel data analysis and show how to fixed the problems.

d. How to determine whether the fixed effect model or random effect model is more appropriate?

Question 4

Questions (a) to (d) refer to the following statement. We fitted a model (Model 1) to explain the (log) markup (L_MARKUP) of 554 bank branches as a function of:

i. the (log) turnover of the branch (L_TURNOVER)

ii. the (log) number of clients of the branch (L_CLIENTS)

iii. the salesforce, measured by the number of employees in the branch (SALES_FORCE)

iv. its SQUARE (SQ_SALES_FORCE)

v. five dummy variables, denoted by Cj ( j = 1, 2, 3, 4, 5), so the value of Cj is one if the ith branch has the complexity level j, and zero otherwise. Bear in mind that C1 and C5 correspond to the highest and smallest degrees of complexity.

Model 1

Dependent variable: L  MARKUP OLS using 554 observations

	____________ Coefficient	Standard error	t-statistic	p-value
Constant	2.1149	0.3302	6.4043	<0.0001
L_TURNOVER	0.8198	0.0392	20.8942	<0.0001
L_CLIENTS	0.3144	0.0409	7.6795	<0.0001
SALES_FORCE	0.0161	0.0333	0.4829	0.6294
SQ_SALES_ FORCE	−0.0024	0.0023	−1.0330	0.3021
C2	0.1145	0.0352	3.2482	0.0012
C3	0.0831	0.0442	1.8816	0.0604
C4	0.0929	0.0530	1.7533	0.0801
C5	0.0812	0.0636	1.2772	0.2021

Mean dep. variable	11.7720	S.D. dependent var.	0.7946
Sum squared resid.	35.4814	S.E. of regression	0.2552
R-squared	0.8984	Adjusted R-squared	0.8969
F(8, 545)	602.23	P-value(F)	0.0000

a. The errors in Model 1 could be heteroskedastic, so we run an auxiliary regression (White’s test) of the squared residuals from Model 1. If the R2 of this regression is 0.393 and Prob (X332 > 217.685) = 0, can we reject the null hypothesis? Why?

b. Do you expect a multicollinearity problem as shown in Model 1? Why?

c. What are the determinants of the Markup based on the estimation result? How do you know?