Karachi Institute of Economics and Technology

Course: Quantitative Techniques for Management Sciences

Faculty: Dr. Arsalan Hashmi
Class ID: 103527
Final Assessment Date: 11th May, 2020
Total Marks: 50

INSTRUCTIONS

1. All questions must be answered. Answers should be written in your own words.
2. Identical answers of two or more students will be considered as a cheating case and not be graded.
3. All forms of academic dishonesty, such as plagiarism, are prohibited and will be strictly penalized.
4. Save your answers in an MS Word file with your full name and student ID. Email the MS Word file at
([email protected]) and upload the same on LMS within a maximum time of 12 hours from
the start of the assessment. All late submissions will not be graded.

SECTION A – DISCUSSION QUESTIONS

(16 marks)

1. Explain the purpose of conducting regression analysis. How does it compare with correlations analysis?
2. Differentiate between a null hypothesis and alternative hypothesis. Explain how can we can test a hypothesis?
3. Explain how t-tests, F-tests and confidence intervals can be used by a researcher?
4. What is a dummy variable? Explain when and how dummy variables are used in research?
5. Explain why we should analyse the regression residuals for the normal distribution assumption.
6. What is multicollinearity and how should a researcher detect it? Explain ways of addressing the
multicollinearity problem.
7. Explain the term heteroscedasticity? Discuss ways of detecting and correcting the problem?
8. Why is autocorrelation considered to be a statistical problem? How can a researcher detect autocorrelation and
address the problem?

SECTION B – ANALYTICAL QUESTIONS

9. You estimated a regression of the form given by (Model A) below in order to evaluate the effect of various
firm-specific factors on the returns of a sample of firms. You run a cross-sectional regression of 200 firms (6
marks)

(Model A)

Where, ri is the percentage annual return for the stock

Si is the size of firm i measured in terms of sales revenue

MBi is the market to book ratio of the firm

PEi is the price/earning (P/E) ratio of the firm

BETAi is the stock’s CAPM beta coefficient

You obtain the following results (with standard errors in parentheses)

Page 1 of 6
a. What do you conclude about the effect of each variable on the returns of the security? Explain in detail.

b. On the basis of your results, what variables would you consider deleting from the regression? (Hint: Calculate the t-
statistics and delete coefficients that are insignificant at 5% level of significance).

c. If a stock’s beta increased from 1 to 1.2, what would be the expected effect on the stock’s return? Is the sign on beta
as you would have expected?

10. The following results were obtained by using several different functional forms of regression models.
(6 marks)

Log(EXDUR) = -7.5417 + 1.6266Ln(PCEX) -------- (1)

R-squared=0.9695

Log(EXS) = 8.3226 + 0.00705t ------------ (2)

R-squared=0.9919

EXS = 4111.545 + 30.674t ----------- (3)

R-squared=0.9935

Where,

PCEX= Personal consumption expenditure (in billions)

EXDUR = expenditure on durable goods (in billions)

EXS = expenditure on services (in billions)

t = time trend (2003-Q1 TO 2006-Q3)

N = 15

a. Interpret the foregoing results.

b. Calculate the value of adjusted r-squared for each model and interpret it.
c. Perform an F-test to test the overall significance of the models at the 5% level of significance.

Page 2 of 6
11. A researcher obtained the statistical results after estimating the following double log model (10 marks)

Log(Output) = log(β1) +β2log(Capital) + β3log(Labour) + μ

Descriptive Statistics
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
output | 27 97.88367 37.18507 35.858 138.842
capital | 27 196.8046 99.3422 59.6 358.231
labor | 27 862.9407 133.1631 637 1020.9

Correlations
| output capital labor
-------------+---------------------------
output | 1.0000
capital | 0.9474 1.0000
labor | 0.9895 0.9547 1.0000

Regression Results
Source | SS df MS Number of obs = 27
-------------+------------------------------ F( 2, 24) = 407.50
Model | 5.37753949 2 2.68876975 Prob > F = 0.0000
Residual | .158356562 24 .00659819 R-squared = 0.9714
-------------+------------------------------ Adj R-squared = 0.9690
Total | 5.53589605 26 .212919079 Root MSE = .08123

------------------------------------------------------------------------------
logoutput | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
loglabor | 2.328398 .5994894 3.88 0.001 1.091112 3.565683
logcapital | .1398108 .1653906 0.85 0.406 -.2015386 .4811603
_cons | -11.93657 3.211061 -3.72 0.001 -18.56388 -5.30927
------------------------------------------------------------------------------

Breusch-Godfrey LM test for autocorrelation

---------------------------------------------------------------------------
lags(p) | chi2 df Prob > chi2
-------------+-------------------------------------------------------------
1 | 18.055 1 0.0000
---------------------------------------------------------------------------
H0: no serial correlation

a. Briefly explain the descriptive statistics of the variables.

b. Interpret the correlations between the variables.
c. Interpret the regression coefficients. Explain whether they make logical sense or not.
d. Which regression coefficients are statistically significant at the 1% level?
e. Comment about the goodness of fit of the regression models.
f. Is the model statistically significant at the 1% level? Why?
g. Explain the role of the Breusch-Godfrey LM test? What do the test results indicate and how should the
researcher respond?

Page 3 of 6
 12. The following results were obtained by regression smoking (X) on cancer mortality (Y). (6 marks)

150
Cancer Mortality (Y)

100
50
60 80 100 120 140
Smoking(X)

Source | SS df MS Number of obs = 25

-------------+------------------------------ F( 1, 23) = 24.23
Model | 8395.74904 1 8395.74904 Prob > F = 0.0001
Residual | 7970.25096 23 346.53265 R-squared = 0.5130
-------------+------------------------------ Adj R-squared = 0.4918
Total | 16366 24 681.916667 Root MSE = 18.615

------------------------------------------------------------------------------
cancermort~y | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
smokingx | 1.087532 .2209452 4.92 0.000 .6304724 1.544592
_cons | -2.885319 23.03372 -0.13 0.901 -50.5342 44.76356
------------------------------------------------------------------------------

Shapiro-Wilk W test for normal data

Variable | Obs W V z Prob>z

-------------+--------------------------------------------------
v | 25 0.96038 1.101 0.197 0.42203

a. What do you conclude from the scatterplot?

b. Interpret the regression coefficients and comment whether the relationship is consistent with your
expectations.
c. Test the null hypothesis (i.e. H0: β2=0) using the t-test and confidence interval approach. Explain whether you
would reject the null hypothesis.
d. Test whether there is a one-to-one relationship between Y and X (i.e. H0: β2=1).
e. Explain whether the residuals from the above regression are normally distributed at the 5% level of
significance. Explain your answer.

13. A researcher estimated the following dummy variable model and obtained the results (6 marks)

DUR = β1 +β2D2 + β3D3 + β4D4 + μ

Where, DUR = Expenditure on durable goods in millions of dollars

D2= Dummy variable for second quarter of the year
D3 = Dummy variable for third quarter of the year
D4 = Dummy variable for fourth quarter of the year

Page 4 of 6
 Source | SS df MS Number of obs = 32
-------------+------------------------------ F( 3, 28) = 0.10
Model | 496.255936 3 165.418645 Prob > F = 0.9568
Residual | 44339.9643 28 1583.57016 R-squared = 0.0111
-------------+------------------------------ Adj R-squared = -0.0949
Total | 44836.2203 31 1446.32969 Root MSE = 39.794

------------------------------------------------------------------------------
dur | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
d2 | 1.037497 19.89705 0.05 0.959 -39.71976 41.79475
d3 | 8.062498 19.89705 0.41 0.688 -32.69476 48.81975
d4 | 8.637499 19.89705 0.43 0.668 -32.11976 49.39476
_cons | 276.15 14.06934 19.63 0.000 247.3303 304.9697
------------------------------------------------------------------------------

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity

Ho: Constant variance
Variables: fitted values of dur

chi2(1) = 0.27
Prob > chi2 = 0.6054

a. Interpret the regression parameters in detail.

b. What is the omitted or reference category in the model. Explain.
c. What do the Breusch-Pagan test results indicate? Do we need to apply any corrective approach for addressing
heteroscedasticity? Explain.

Selected Formulas

Page 5 of 6
Page 6 of 6