1) Company Black&White produces a special black-white coffee.

Consider the following table in

which the dependent variable is the quantity demanded for the special coffee. The independent
variables are the price of the special coffee and consumer’s income.
Company Black&White

Quantity demanded (kg) Price of coffee (EUR/kg) Consumer’s income (EUR)

1 8640 13,5 3600
2 8532 18 3750
3 8478 22,5 4050
4 8460 27 4500
5 8424 31,5 4950
6 8388 36 5250
7 8370 40,5 5430
8 8334 45 5835
9 8298 49,5 6000
10 8244 54 6300
11 8244 58,5 6480
12 8208 63 6600
13 8172 67,5 6720
14 8154 72 7650
15 8118 76,5 7800
16 8100 81 8100
17 8064 85,5 8250
18 8028 90 8400
19 8010 94,5 9000
20 7974 99 9300

a) Provide a regression estimation of the quantity demanded for the special green tea produced
by Company Green&Tea on the price of the special green tea using by a reciprocal regression
̂1
β
̂𝟎 +
model (𝜷 ). ̂ 𝟎 and β̂1 parameters, and give the estimated regression equation.
Compute the 𝜷
X

1
𝑦 = 𝛽0 + 𝛽1 ∙ .
𝑥

𝑛 𝑛
1
∑ 𝑦̂𝑖 = 𝛽̂0 ∙ 𝑛 + 𝛽̂1 ∑
𝑥𝑖
𝑖=1 𝑖=1

1 1 1
∑𝑛𝑖=1 𝑦̂𝑖 ∙ = 𝛽̂0 ∙ ∑𝑛𝑖=1 𝑥 + 𝛽̂1 ∑𝑛𝑖=1(𝑥 )2 ,
𝑥𝑖 𝑖 𝑖

165240 = 20 ∙ 𝛽̂0 + 0.486847 ∙ 𝛽̂1

4077.198 = 0.486847 ∙ 𝛽̂0 + 0.017308 ∙ 𝛽̂1

𝛽̂1 = 10053,14

𝛽̂0 = 8017.283

The estimated model:

 1
𝑦 = 8017.283 + 10053.14 ∙
𝑥
b) Calculate the SSE (Residual Sum of Squares). Give the interpretation of your result.

𝑆𝑆𝐸 = ∑(𝑦𝑖 − 𝑦̂í )2 = 121061.2

SSE shows the deviations of estimated values from actual empirical values of data.
c) Calculate the SSR (Regression Sum of Squares). Give the interpretation of your result.

𝑆𝑆𝑅 = ∑(𝑦̂í − 𝑦̅)2 = 551562.8196

SSR shows the deviations of the estimated values around the sample mean.
d) Test the null hypothesis that the coefficient β1 is zero against the alternative hypothesis that it is
not, at the 5% level of significance. Give the interpretation of your result.

The null hypothesis states that the 𝛽1 regression parameter is equal to zero, and the alternative
hypothesis states the opposite of the null hypothesis:

𝐻0 : 𝛽1 = 0,

𝐻𝑎 : 𝛽1 ≠ 0.

The value of t-calculated is the ratio of the estimated parameters and its standard error:
̂
𝛽
𝑡 =𝜎1.
̂1
𝛽

2
∑(𝑦𝑖 −𝑦̂í ) 121061.2
𝜎𝑒 = √ =√ = 82.0099.
𝑛−2 18

𝜎𝑒 82.0099
𝜎𝛽̂1 = ̅
= = 1110.12.
1 1
√∑(𝑥 −(𝑥))2 √0.0054575
𝑖
 𝛽̂1 10053,14
𝑡𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 = = = 9.0559
𝜎𝛽̂1 1110.12

t table value: 2.101

|𝑡| > 𝑡1−𝛼 ,

2

9.0559 >2.101,

We reject the null hypothesis (that 𝛽1 = 0) and accept the alternative hypothesis that 𝛽1 ≠ 0.
In this case, we can state that there is a statistically significant relationship between the quantity
demanded and the price of coffee.
e) Create a 95% confidence interval for the coefficient β1 . Give the interpretation of your result.

To calculate the confidence interval, we need the two-sided critical t-value and the standard
error of the estimated parameter:

𝛽̂𝑗 − 𝑡𝑐 ∙ σβ̂j ≤ 𝛽𝑗 ≤ 𝛽̂𝑗 + 𝑡𝑐 ∙ σβ̂j .

𝛽1 = 𝛽̂1 ± 𝑡1−𝛼 ∙ 𝜎𝛽̂1 ,

2

𝑡1−𝛼 = 𝑡 0,05 = 𝑡1−0,025 = 𝑡0,975 = 2.101

2 1−
2

Degrees of freedom n-2=18.

10053.1405 + 2.101 ∙ 1110.12 = 12385.5,
10053.1405 − 2.101 ∙ 1110.12 = 7720.8

confidence interval:

7720.8 < 𝛽1 < 12385.5

The confidence interval does not include the zero value, and it means that 95% is the probability
that 𝛽1 is different from zero, as we could have seen from the t-test. Confidence interval means
that 95% is the probability that the regression parameter can be found in the interval:

7720.8 < 𝛽1 < 12385.5

f) Provide a regression estimation of the quantity demanded for the special coffee produced by
Company Black&White on the price of the special coffee and the consumer’s income data
using by a multiple regression model. Give the estimated multiple regression equation. Give the
interpretation for the slopes of the estimated regression equation.
 𝑦 = 8636.95 − 7.497 ∙ 𝑥1 + 0.0073 ∙ 𝑥2 ,
Interpretation of the coefficients:

- If the coffee price increases by 1 EUR, the quantity demanded decreases by 7.497 kg, all other
things being equal, i.e. with the same consumers’ income.

- If the consumers’ income increases by 100 EUR, the quantity demanded for coffee increases
by 0.073kg with all other things unchanged, i.e. with the same coffee price.

g) Conduct an F-statistic for the model and interpret your results. (significance level is 5%)
SUMMARY OUTPUT

Regression Statistics
Multiple R 0,993242134
R Square 0,986529937
Adjusted R Square 0,984945224
Standard Error 23,08588522
Observations 20

ANOVA
Significance
df SS MS F F
Regression 2 663563,7124 331781,8562 622,529 1,26E-16
Residual 17 9060,287638 532,9580963
Total 19 672624

Standard Upper
Coefficients Error t Stat P-value Lower 95% 95%
Intercept 8636,94854 89,57258314 96,4240199 1E-24 8447,967 8825,93
Price of coffee (EUR/kg) - - 0,00271
(x) 7,497209854 2,138726901 3,505454507 2 -12,0095 -2,98489
0,82518
Consumer’s income (EUR) 0,007309735 0,032587437 0,224311455 9 -0,06144 0,076063

Calculate the sta Implement the F test at 95% confidence level for Exercise 1:

𝐻0 : 𝛽1 = 𝛽2 = 0

𝐻1 : at least one of 𝛽1 , 𝛽2 is different from zero. 𝛽k (𝑘 = 1, 2) ≠ 0

𝑆𝑆𝑅 663563.7124
MSR = = = 331781.8532
𝑘 2

𝑆𝑆𝐸 9060.287638
MSE = = = 532.958
𝑛−𝑘−1 17

SSR/k MSR 331781.8532

F𝑘𝑛−𝑘−1 = = = = 622.529
SSE/(n − k − 1) MSE 532.958

F𝑘𝑛−𝑘−1 ≥ 𝐹𝑐𝑟𝑖𝑡𝑖𝑐𝑎𝑙 (2,17)

622.529 > 3,59.

We reject the null hypothesis, which means that there exists at least one explanatory variable
whose coefficient is different from zero, i.e., it has explanatory power in the model concerning
the dependent variable.

h) The standard error of the β1 regression parameter (coefficient of the consumer’s income).

𝜎𝑒 23.085885
𝜎𝛽1 = = = 2.13873
√∑(𝑋1 − 𝑋̅1 )2 ∙ (1 − 2
𝑟𝑥1,𝑥2 ) √13466.25 ∙ (1-0,99556642 )
i) Calculate the VIF (Variance Inflation Factor) indicator. Give the interpretation of your result.

SUMMARY OUTPUT

Regression Statistics
Multiple R 0,99566441
R Square 0,991347617
Adjusted R Square 0,990866929
Standard Error 2,544221671
Observations 20

ANOVA
Significance
df SS MS F F
Regression 1 13349,73485 13349,73 2062,352 5,06E-20
Residual 18 116,5151504 6,473064
Total 19 13466,25

Upper
Coefficients Standard Error t Stat P-value Lower 95% 95%
-
Intercept 40,81643143 2,211825408 -18,4537 3,85E-13 -45,4633 -36,1696
Consumer’s income (EUR) 0,015170778 0,000334062 45,41312 5,06E-20 0,014469 0,015873

1 1
𝑉𝐼𝐹 = 2 = 1 − 0,991347 = 115.57
1 − 𝑅𝑗

There is a strong multicollinearity in the model, we can mitigate it if we omit one

independent variable from the model, or insert a new independent variable into the model.