Name: Kamshagini Nallainathan

CPM: 19601
MC: 94269

Question 01

Thomas is a Senior Manager of Hairgrowth Company, and Thomas expects to

estimate the demand for the product Hairgrowth Double X hair oil. Thomas’s work
colleague suggests the use of regression analysis. He has run a regression model for
Hairgrowth Double X and an extract of the summary output of the regression model
is given below.

SUMMARY
OUTPUT

Regression Statistics
Multiple R 0.893435
R Square 0.817753
Adjusted R Square 0.789481
Standard Error 650.978288
Observations 40

Standard P-
Coefficients Error t Stat value
Intercept 22230.85 4417.857 2.316 0.014
PHDX -120.09 20.989 -4.902 0.000
PRG1 50.43 21.440 2.309 0.032
ADEX 0.76 0.378 2.021 0.045
INC 0.16 0.117 1.363 0.192
PRG2 -25.30 14.079 -1.797 0.079

The dependent variable is sales quantity of

Hairgrowth Double X. The Independent variables are
defined below:
PHDX: Price of Hairgrowth Double X
PRG1: Price of a related good 1
ADEX: Advertising expenses
INC: Average income of the customers
PRG2: Price of a related good 2

You are required to answer the following questions related to the above regression output.
 a) Construct the estimated demand equation for this regression model and interpret the
estimated coefficient of PHDX.
 Estimated Demand Equation:
Y=22230.85-120.09PHDX+50.43PRG1+0.76ADEX+0.16INC-25.30PRG

 Interpretation:
PHDX= Sales quantity will decrease by 120.09 units when price of hair growth double X is increased
by 1 rupee.

b) Comment on the individual significance of each of the independent variable and

overall explanatory power of this model (Copy the table below to the answer sheet).
Based on P value approach
Variable Comment on individual significance
Price of hair growth P value < Significance value
double X 0.0 < 0.05
So there is a statistically significant relationship between Price of
hair growth double X and sales quantity of Hair Growth Double X,
at a 5% significant level.
Price of a related good 1 P value < Significance value
0.032 < 0.05
So there is a statistically significant relationship between Price of a
related good 1 and sales quantity of Hair Growth Double X, at a 5%
significant level.
Advertising expenses P value < Significance value
0.045 < 0.05
So there is a statistically significant relationship between
Advertising expenses and sales quantity of Hair Growth Double X,
at a 5% significant level.
Average income of the P value >Significance value
customers 0.192 > 0.05
So there is no statistically significant relationship between Average
income of the customers and sales quantity of Hair Growth Double
X, at a 5% significant level.
Price of a related good 2 P value >Significance value
0.079 > 0.05
So there is no statistically significant relationship between Price of
a related good 2and sales quantity of Hair Growth Double X, at a
5% significant level.

Overall explanatory It is high goodness of fit. Because of it’s R 2 value is very close to 1.
power [R2 = 0.817753 ]

The explanatory power of this model is determined by R 2 value. .In

our case 81.77% of total variance in sales quantity of Hair Growth
Double X [y] is explained by variance in Price of hair growth double
X, Price of a related good 1, Advertising expenses, Average income of the

2
 customers and Price of a related good 2.
c) What observation or suggestions can you propose for the competitive strategy for
Hairgrowth Double X.
 Product differentiation.
 Market niche focus strategy.
Eg – when estimating sales of Hair growth Double X should proceed true income and Price
of a related good 2.

d) Assuming that initial average income of customers is Rs. 10,000 and the quantity of sales at
that point is 16,000. Calculate Income Elasticity of demand for Hairgrowth Double X and
interpret the answer (Clearly show the function used to arrive at the answer).
 Income Elasticity = -b4 * I
QX
= 0.16 * 10000
16000
= 0.1

 Interpretation: One rupee additional income will increase the sales quantity of Hair Growth
Double X by 10%.

Question 02

Unilever Sri Lanka, formerly known as Lever Brothers Ceylon Limited was incorporated in 1938
with brands such as Sunlight, Lux and Pears Rose. In 1958 they established their own selling and
distribution organization. Over the next three decades the company diversified and expanded their
operations which were managed by several corporate entities such as Lipton, Brooke Bond and
Premier Foods. In the nineties, they consolidated their businesses under the umbrella of Unilever
Ceylon Limited. In 2004 they became known as Unilever Sri Lanka.

Brands
 Home Care: Sunlight, Vim, Rin, Surf Excel, Comfort, Wonderlight

 Personal Care: Lux, Lifebuoy, Signal, Rexona, Ponds, Pears, Fair & Lovely, Sunsilk,
Dove, Axe, Clear, Vaseline
 Food: Ceylonta, Lipton, Laojee, Bru, Astra, Flora, Marmite and Knorr

3
 Summary Findings of the Consumer Survey
Qdx = 38.597 -0.071 Px + 0.243 Cx + 0.104 Sx - 0.652 I + 0.005 A

t ( 9.158 ) ( -0.723 ) ( 2.716 ) ( 1.477 ) ( -3.738 ) ( 0.985 )

Prob. 0.000 0.125 0.004 0.0956 0.001 0.105

Where, Px = Price of the product, Cx = Price of an ayeuruwedic soup, Sx = Price of a herbal

soup, I = Per capita income, A = Advertising expenses, Qdx = Lux sales volume
(t-statistics are shown in parentheses.)
R2 = 0.944
Adjusted R2 = 0.902
Sample Size = 23
Standard Error (Regression) = 1.980
F statistic = 57.633
Prob (F-statistic) = 0.000
Significance Level = 5%

a) Recently sales reports of the personal care indicated that there is a sales drop in LUX compared
to the other brands. Therefore the brand manager of the LUX has conducted a consumer survey
and the results are stated in the table above.

1. Describe the impact of each independent variable on quantity demanded, relative

importance of each independent variable, and overall explanatory power of this model
and overall significance of the model.

 Px = Quantity demanded of Lux will decrease by 0.071million units when price of the
product is increased by Rs. 1 million. Quantity demanded of Lux will increase by
0.071million units when price of the product is decreased by Rs. 1 million. - Negative
relationship

 Cx = Quantity demanded of Lux will increase by 0.243 million units when price of an
ayuerwedic soap is increased by Rs. 1million. Quantity demanded of Lux will decrease by
0.243million units when price of an ayuerwedic soap is decreased by Rs. 1million.-Postive
relationship

 Sx = Quantity demanded of Lux will increase by 0.104 million units when price of an herbal
soup is increased by Rs. 1million. Quantity demanded of Lux will decrease by 0.104 million
units when price of an herbal soup is decreased by Rs. 1million.-Postive relationship
4
  I= Quantity demanded of Lux will decrease by 0.652 million units when per capita Income is
increased by Rs. 1million. Quantity demanded of Lux will increase by 0.652 million units
when per capita income is decreased by Rs. 1million.-negative relationship

 A= Quantity demanded of Lux volume will increase by 0.005 million units when advertising
expenses is increased by Rs. 1million. Quantity demanded of Lux will decrease by 0.005
million units when advertising expenses is decreased by Rs. 1million.-Postive relationship
 Relative importance of each independent variables

 Testing significance of 1st independent variable – Price of the product [lux]-(PX)

1. Establish Hypothesis.

H0: PX = 0 (There is no statistically significant relationship between Sales quantity of Lux soup
and price of the Lux soup.)

H1: PX = 0 (There is a statistically significant relationship between Sales quantity of Lux soup
and price of the Lux soup.)

2. Select a level of significance & establish Decision Rule.

Level of significance – 5%

Decision Rule – If the calculated t statistic lies in a critical region, we reject the null
hypothesis.

-Critical value +Critical value

3. Calculate t statistic
t = -0.723

4. Compare t statistic with critical value.

Degree of freedom = n-k
= 23-6 =17
Two tailed test-5%
Critical values = -2.110 or 2.110
5
 t value < critical value
-0.723 < -2.110
0.723 < 2.110

5. Arrived at the conclusion

Since the calculated t value (-0.723) does not exceed the critical value ( -2.110) at 5%
significance level we do not reject the null hypothesis that there is no statistically relationship
between the sales quantity of Lux soup and the price of lux soup.

We accept that. Therefore there is no statistically relationship between the sales quantity of Lux
soup and the price of lux soup at a 5% level of significance.

 Testing significance of 2nd independent variable – Price of an ayeuruwedic soup (CX)

1. Establish Hypothesis.

H0: PX = 0 (There is no statistically significant relationship between Sales quantity of Lux soup
and price of an ayeuruwedic soup)

H1: PX = 0 (There is a statistically significant relationship between Sales quantity of Lux soup
and price of an ayeuruwedic soup.)

2. Select a level of significance & establish Decision Rule.

Level of significance – 5%

Decision Rule – If the calculated t statistic lies in a critical region, we reject the null
hypothesis.

-Critical value +Critical value

3. Calculate t statistic
t = 2.716

4. Compare t statistic with critical value.

Degree of freedom = n-k
6
 = 23-6 =17
Two tailed test-5%
Critical values = -2.110 or 2.110

t value > critical value

2.716 > 2.110
5. Arrived at the conclusion
Since the calculated t value (2.716) exceed the critical value (2.110) at 5% significance level we
reject the null hypothesis that there is no statistically relationship between the sales quantity of
Lux soup and price of an ayeuruwedic soup.

We accept alternative hypothesis. Therefore there is statistically relationship between the sales
quantity of Lux soup and the price of an ayeuruwedic soup at a 5% level of significance.

 Testing significance of 3rd independent variable –Price of a herbal soup (Sx)

1. Establish Hypothesis.

H0: PX = 0 (There is no statistically significant relationship between Sales quantity of Lux soup
and price of a herbal soup.)

H1: PX = 0 (There is a statistically significant relationship between Sales quantity of Lux soup
and price of a herbal soup.)

2. Select a level of significance & establish Decision Rule.

Level of significance – 5%

Decision Rule – If the calculated t statistic lies in a critical region, we reject the null
hypothesis.

-Critical value +Critical value

3. Calculate t statistic
t = 1.477

7
 4. Compare t statistic with critical value.
Degree of freedom = n-k
= 23-6 =17
Two tailed test-5%
Critical values = -2.110 or 2.110

t value < critical value

1.477 < 2.110

5. Arrived at the conclusion

Since the calculated t value (1.477) does not exceed the critical value (2.110) at 5% significance
level we do not reject the null hypothesis that there is no statistically relationship between the
sales quantity of Lux soup and the price of a herbal soup.

We accept that. Therefore there is no statistically relationship between the sales quantity of Lux
soup and the price of a herbal soup at a 5% level of significance.

 Testing significance of 4th independent variable – Per capita income (I)

1. Establish Hypothesis.

H0: PX = 0 (There is no statistically significant relationship between Sales quantity of Lux soup
and per capita income.)

H1: PX = 0 (There is a statistically significant relationship between Sales quantity of Lux soup
and per capita income.)

2. Select a level of significance & establish Decision Rule.

Level of significance – 5%

Decision Rule – If the calculated t statistic lies in a critical region, we reject the null
hypothesis.

-Critical value +Critical value

8
 3. Calculate t statistic
t = -3.738

4. Compare t statistic with critical value.

Degree of freedom = n-k
= 23-6 =17
Two tailed test-5%
Critical values = -2.110 or 2.110

t value >critical value

3.738 > 2.110

5. Arrived at the conclusion

Since the calculated t value (-3.738) exceed the critical value (-2.110) at 5% significance level
we reject the null hypothesis that is there is no statistically relationship between the sales quantity
of Lux soup and the per capita income.

We accept alternative hypothesis. Therefore there is a statistically relationship between the sales
quantity of Lux soup and the per capita income at a 5% level of significance

 Testing significance of 5th independent variable – Advertising expenses (A)

1. Establish Hypothesis.

H0: PX = 0 (There is no statistically significant relationship between Sales quantity of Lux soup
and advertising expenses.)

H1: PX = 0 (There is a statistically significant relationship between Sales quantity of Lux soup
advertising expenses.)

2. Select a level of significance & establish Decision Rule.

Level of significance – 5%

Decision Rule – If the calculated t statistic lies in a critical region, we reject the null
hypothesis.

9
 -Critical value +Critical value

3. Calculate t statistic
t =0.985

4. Compare t statistic with critical value.

Degree of freedom = n-k
= 23-6 =17
Two tailed test-5%
Critical values = -2.110 or 2.110

t value < critical value

0.985< 2.110

5. Arrived at the conclusion

Since the calculated t value (0.985)does not exceed the critical value ( 2.110) at 5% significance
level we do not reject the null hypothesis that is there is no statistically relationship between the
sales quantity of Lux soup and the advertising expenses.

We accept that. Therefore there is no statistically relationship between the sales quantity of Lux
soup and the advertising expenses at a 5% level of significance.

 Overall explanatory power of the model

The explanatory power of the model is determined by the R 2 value. This shows how much of the
variation in explained variable is reflected in the variance of estimated explanatory variables. In our case
the R2value is 0.944, meaning 94.4% of total variance in quantity demanded of Lux(Y) is explained by
the variances in price of the product(Px), price of an ayeurwedic soap(Cx), price of the herbal soup(Sx),
per capita income(I) and advertising expenses(A).

However, we have to consider the problem of declining of freedom. That is, as more explanatory
variables are introduced, degree of freedom declines. Thus R2is adjusted for this and is modified as
adjusted R squared, denoted as R̅2. The R̅2 of this model is 0.902 meaning price of the product(Px), price
of an ayeurwedic soap(Cx), price of the herbal soup(Sx), per capita income(I) and advertising
expenses(A) explain 90.2% of total variance in the model.

 Overall significance of the model.

The overall significance of the model can be determined by the variable known as the F statistic or the F ratio,
which is calculated as so:
2
R
[ ]
k −1
F=
1−R2
[ ]
n−k

1. Finding F statistic
F= 57.633

10
 2. Establish Hypothesis.
H0: BK 0 (No independent variables affects on the dependent variables.)

H1: At least one Bi = 0 (At least one independent variables affects on the dependent
variables.)

3. Select a level of significance & establish Decision Rule.

Level of significance = 0.05

Decision Rule – If the calculated F statistic lies in a critical region, we reject the null hypothesis.

Critical value

4. Compare the calculated F statistic with the critical value.

Numerator= k-1 = 6-1 =5
Denominator = n-k = 23-6 =17

Critical value = 2.8100

F statistic > critical value
57.633 > 2.8100

5. Conclusion.
Since calculated F statistic (57.633) exceeds the critical value (2.8100) , at 5% significance
level, we reject the null hypothesis.
Not all coefficients are equal to zero at 5% level of significance. Therefore , the overall model is
significant at 5% level of significance.

2. What factors might LUX brand manager consider in developing an effective competitive
strategy?

b) The Brand Manager of Astra is 90 % confident that the true value of Astra sales volume in next
year will be 10.5 million (Lowest value). Also the manager expects changes in the independent
variables of Astra as follows.

Increase Curren
Variable Change Elasticity
11
 or t
Decrease Values
Price (PY) 10% Increase ………………….. Rs. 200 /=
Income ( I ) 20% Increase -0.50 Rs.150 /=
Price of Flora ( PF) 25% Decrease 2.50 Rs. 250 /=
Price of Pam Oil
15% Decrease -3.00 Rs. 100 /=
(PP)
Advertising (A) 10% Increase 2.00 Rs.125 /=

Standard error (SE) of the regression is 2.575.

Estimated Sales (t) = 20.194
t α/2,n-k = 2

1. Find price elasticity of demand.

Confidence interval = point estimator +_ t 10%2(n-k)*

Standard error = 20.194+_ 2*2.575= 20.194+_ 5.15
Here we take lowest value which is 20.194-5.15= 15.044
St= 15.044 St+1= 10.5
Price elasticity = ∂Q%/∂P%*∂Q= (10.5- 15.044)/ 15.044= -0.3020
= 30.20%/10% = -3.020

2. Find Marginal Revenue of Astra and interpret

Qd= a-bp
15.044= a-b*200
15.044= a- 0.2272p
a= 60.484

Price elasticity = (∂ Q / ∂ p) * (P / Q)
Price elasticity = b * (P / Q)
-3.020 = b * (200/15.044)

Qd= 60.484- 0.2272 p

P= 266.21- 4.40Q
TR = P*Q
TR= (266.21- 4.40Q)Q
TR= (266.21- 4.40Q)Q2
MR= 266.21-8.8Q
MR= 266.21-8.8* 15.044= 133.82
When astra increase by one unit, it drives additional Rs. 133.82 sales value to the total revenue.

12
 3.Construct Regression line for Astra
Yt=60.484-0.2272Px-(0.50*15.044/150)I+(2.50*15.044/250)Pf–(-3*15.044/100)Pp+(200*15.044/125) A
Yt=60.484-0.2272Px- 0.0501467I+0.15044Pf- -0.45132Pp+ 0.240

13