Scribd
Upload
1 views

Assignments

Uploaded by

Sidra Nazir
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
1 views

Assignments

Uploaded by

Sidra Nazir
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Econometrics

Project Assignment

Part 1

Data Description
The provided data set is taken from The World Bank Enterprise Surveys. The sample in-
cludes 377 observations of active Albanian enterprises with at least five employees in the
year 2019.
The variables are:

1. industry: 1 for manufacturing, 2 for retail, 3 for others,

2. size: 1 for small (5–19 employees), 2 for medium (20–99), 3 for large (100 and more)

3. sales: total annual sales in the last fiscal year,

4. n_employees: 1 for small (5–19 employees), 2 for medium (20–99), 3 for large (100
and more),

5. female_top_manager: 1 if company is run by female manager, 2 if not,

6. rd_spendings: 1 if company did invest in R&D last fiscal year, 2 if not,

7. equipment_spendings: sum of spendings on purchasing equipment during last fiscal


year,

8. land_spendings: sum of spendings on purchasing land during last fiscal year,

9. w: observation weights.

There are two special values: NA means that the observation is missing, −9 means "don’t
know". Sample weights w are usually used to make inference of the population while having
stratified sample, which often violates the assumption of random sampling.

Main task
Analyse the effect of various firm characteristics on sales.

1. Provide descriptives for the data.

(a) Summary statistics (mean, median, minimum, maximum etc.)


(b) Histograms for main variables of interest
(c) Scatter plots

2. Estimate and select your model

1
(a) Estimate a linear model:

salesi = β0 + β1 n_employeesi + εi

(b) Estimate the log-log and log-lin specifications of the model:

ln salesi = β0 + β1 ln n_employeesi + εi

ln salesi = β0 + β1 n_employeesi + εi

(c) Provide interpretations for estimated models.


(d) Include a dummy variable rd_spendings as an intercept dummy, as a slope dummy
and simultaneously (as intercept and slope dummy)
(e) Conduct a Chow test. Which specification do you prefer?
(f) Estimate other model specifications using other variables in the dataset. Choose
final specification.
(g) Provide interpretations for the chosen specification.

3. Evaluate your model

(a) Conduct the RESET-Ramsey test. What is the conclusion?


(b) Investigate residuals for normality: draw Q-Q plot, histogram with the kernel
density function, conduct Kolmogorov-Smirnov test, Shapiro-Wilk test.
(c) Check whether your model suffers from multicollinearity: calculate VIF and cor-
relations.
(d) Test your model for heteroskedasticity. Conduct White test, Goldfeld-Quandt
test and Breusch-Pagan test. Provide corrected standard errors if necessary.

Bonus task
Estimate your final model using weighted least squares and the weighting variable w.

2
Part 2

Data Description
The provided data set includes daily closing price of a high-tech portfolio constructed as
25% of Apple shares, 15% Tesla, 20% Yandex, 20% Google, 20% Boeing. The prices are in
the column portfolio. The column sp500 is a daily value of S&P index, and r_free is a
return of three-month US Treasury bills (expressed as a share of 1).

Main Task
1. Prepare necessary variables:
Dependent variable: excess returns of an asset rt − rf
Independent variable: excess market returns rmt − rf

2. Estimate a linear model for an asset/portfolio

rt − rf = β(rmt − rf ) + εt

3. Investigate residuals for autocorrelation:

(a) Plot the graph for residuals. Can you notice signs of autocorrelation?
(b) Plot the graph for residuals and their lagged values. Can you notice signs of
autocorrelation?
(c) Conduct Durbin-Watson test. Try two-sided and one sided hypothesis. Explain
the result.
(d) Conduct the h-Durbin test. Make the conclusion.
(e) Conduct the Breusch-Godfrey test. Interpret the results.

4. Correct standard errors if necessary.

5. Test linear hypothesis on β:

(a) One-sided hypothesis that β is greater than 1.


(b) One-sided hypothesis that β is less than 1.
(c) Provide economic interpretation the results of the tests.

6. Provide interpretation for the resulting CAPM model.

You might also like