EXERCISES UNIT 2

- First two exercises were solved.

- The solutions of the rest of the exercises appear at the end of the document.

-------------------------------------------------------------------------------------------

EXERCISE 2.1

Given the following sales data for different industries, determine the degree of
concentration:

a) According to the concentration coefficients CR2 and CR4

b) According to the Herfindahl index

Company Industry 1 Industry 2 Industry 3 Industry 4

1 320 450 230 160
2 280 300 190 160
3 50 300 190 160
4 50 300 120 160
5 50 90 120 160
6 50 60 70 160
7 60 160
8 50 160
9 40 160
10 30 160
Total Sales 800 1500 1100 1600

RESOLUTION

a) Concentration coefficients CR2 and CR4

|
Sum of market shares between 2 and 4 largest firms in each industry. Being:

CR m = � ∝i
i=1

where ∝i is the market share of a company i, and

1
 ∝1 ≥∝2 ≥ ⋯ ≥ ∝m ≥ ⋯ ∝1 ≥∝T

Industry 1 Industry 2 Industry 3 Industry 4

0.75 0.50
CR 2 0.38 0.20
(320/800+280/800) (0.30+0.20)
0.87 0.90
CR 4 0.73 0.40
(320/800+280/800+50/800+50/800) (0.30+0.20+0.20+0.20)

The index is easy to compute but does not take into account the existing asymmetries
between major companies considered in the calculation, in this case for example by the
CR4 Industry 2 is more concentrated than the industry 1, where the latter has 2 companies
that dominate the market.

b) Herfindahl Index

Sum of the squares of the market shares of all companies in the sector.

H = � ∝2𝑖𝑖
i=1

Company Industry 1 Industry 2 Industry 3 Industry 4

1 0.1600 (0.42) 0.0900 0.0437 0.0100
2 0.1225 (0.352) 0.0400 0.0298 0.0100
3 0.0039 (0.06252) 0.0400 0.0298 0.0100
4 0.0039 (0.6252) 0.0400 0.0298 0.0100
5 0.0039 (0.6252) 0.0036 0.0119 0.0100
6 0.0039 (0.6252) 0.0016 0.0040 0.0100
7 0.0000 0.0000 0.0030 0.0100
8 0.0000 0.0000 0.0021 0.0100
9 0.0000 0.0000 0.0013 0.0100
10 0.0000 0.0000 0.0007 0.0100
0.2981
Herfindahl 0.2152 0.1563 0.1000
(0.16+0.1225+0.0039+0.0039+0.0039+0.0039)
When squaring the market shares, gives greater weight to higher fees. Therefore, the
index takes into account the asymmetries within the industry.

2
As can be seen, the industry market is a concentrated one; industries 2 and 3 are
moderated concentrated, while industry 4 is not concentrated.

3
EXERCISE 2.2

Given a series of 20 quarterly data calculate the following provisions:

a) Forecast for the quarter 21 by the method of Moving Averages. Choose the best value
of n using the Mean Square Error.

b) Forecast for the quarter 21 according to the method of Exponential Smoothing. You
should choose the best value using the mean square error method.

c) Which of the two methods is more accurate?

Quarter Sales Quarter Sales

1 21335 11 86130

2 45360 12 60260

3 75900 13 27615

4 55880 14 50160

5 29325 15 67575

6 53760 16 47880

7 73245 17 26700

8 43860 18 58960

9 23310 19 92805

10 52000 20 56700

RESOLUTION:

a) Method of moving averages:

The provision by the method of moving averages is calculated as the average value of
the data of the last n periods. For example, being PS the predicted Sales, and S the real
sales, if we take n = 5 the forecast for period t + 1 is:

4
 st + st−1 + ⋯ + st−4
PSt+1 =
n
In this case, the first period for which an estimate is for the period 6 (mean periods 1-5)
is obtained. For the next forecasting, the period 7, the average is moving one step below
(periods 2-6).

Period (t) Sales (S t ) Predicted Sales PS t = MA(5)

1 21335
2 45360
3 75900
4 55880
5 29325
21335 + 45360 + 75900 + 55880 + 29325
6 53760 PS6 = = 45560
5
45360 + 75900 + 55880 + 29325 + 53760
7 73245 PS7 = = 52045
5
8 43860 57622
9 23310 51214
10 52000 44700
11 86130 49235
12 60260 55709
13 27615 53112
14 50160 49863
15 67575 55233
16 47880 58348
17 26700 50698
18 58960 43986
19 92805 50255
20 56700 58784
47880 + 26700 + 58960 + 92805 + 56700
21 - PS21 = = 56609
5

The forecast value for period 21 is 56609.0. Now it is necessary to experiment with
different values of n and choose one that allows a more accurate prediction. To this end,

5
the mean square error (MSE) for each is calculated. In the case of the moving average
with n = 5, the Mean Square Error is calculated as follows:

(53760 − 45560)2 + (73245 − 52045)2 + ⋯ + (58784 − 56700)2

MSE (MA5) = = 429,800,529
15

The process is repeated with different n. In particular, we have tested MA3, i.e. moving
average with n = 3, MA4, MA5, MA7 and MA10. The lowest mean square error is
obtained to implement Moving Average with n = 4. (With MSE 405,487,441).

c) Exponential smoothing method:

In this method, projected sales for the period t + 1 (PST + 1) are calculated by the
following expression:

PSt+1 =∝ St + (1−∝)PSt
Where are the expected sales PS, S are the actual sales and the weighting being. In the
following table, the forecast for α = 0.1 is performed.

In the first period is used as the sales data provided actual sales data of the same period;
therefore, the first period for which an estimate is obtained for the second.

Predicted Sales PS t α= 0.1

Period (t) Sales (S t )
21335
1 21335
PS2 = 0.1 ∗ 21335 + (1 − 0.1)21335=21335
2 45360
3 75900 PS3 = 0.1 ∗ 45360 + (1 − 0.1)21335 = 23737.5
28953.8
4 55880
31646.4
5 29325
31414.2
6 53760
33648.8
7 73245
37608.4
8 43860
38233.6
9 23310
36741.2
10 52000
38267.1
11 86130
43053.4
12 60260
44774.1
13 27615
6
 43058.2
14 50160
43768.3
15 67575
46149.0
16 47880
46322.1
17 26700
44359.9
18 58960
45819.9
19 92805
50518.4
20 56700
21 - PS21 = 0.1 ∗ 56700 + (1 − 0.1)50518.4 = 51136.6

To measure the goodness of the forecast, we calculate the Mean Square Error for the
periods 2 to 20:

∑20
t=2(St − PSt )
2
𝐌𝐌𝐌𝐌𝐌𝐌 (∝= 0.1) =
19

(45360 − 21335)2 + (75900 − 27737.5)2 + ⋯ + (56700 − 50518.4)2

=
19

= 680,837,037.9

The whole process by giving different values must be repeated. Of those who were tested
(= 0.1, 0.3, 0.5, 0.7 and 0.9), is chosen to be a = 0.3 which has a lower Mean Square Error.

d) Calculation of the goodness of the forecast

Once the forecasts made by the two models: Moving Average with n = 4 and exponential
smoothing, one model of a lower mean square error is chosen.

We choose the moving average n=4 because it has lower MSE as it can be seen in the
following table.

Mean Square Error (MSE) for each model

MSE. Moving average MSE. Exponential Smoothig

(n=4) (α= 0.3)

405,487,441 577,723,868

7
EXERCISE 2.3

The Herfindahl index in an industry made up of identical firms that produce and sell a
homogeneous product is H=0.025.

a) Calculate the number of firms in the industry

b) If after a few years the Herfindahl index of the same industry changes from H=0.04.
What could be the causes of this increase?

EXERCISE 2.4

An industry is made up of seven companies whose market shares are as follows, from 1
to 7: 0.12; 0.22; 0.26; 0.10; 0.06; 0.20 and 0.04.

a) Draw the concentration curve for this industry.

b) Calculate the concentration coefficients CR3 and CR5, as well as the Herfindahl index.

EXERCISE 2.5

According to the information provided by the Naturgy Foundation, there are more than
300 electricity marketing companies in the domestic segment (households) in Spain.
However, there are five large companies that control this market: Endesa, Naturgy,
Iberdrola, Viesgo-Repsol and EDP. Their respective market shares are: 35%, 14%, 34%,
3% and 4%. The rest of the companies in the sector reach a combined share of 10%.

a) Calculate the concentration coefficient of the three largest companies in the sector,
that of the four largest and the Herfindahl index.

b) Assume that there is a merger of the companies Viesgo-Repsol and EDP. Would
this merger affect the concentration ratio of the three largest companies in the sector?
What about the concentration ratio of the four largest companies? And to the
Herfindahl index?

8
 EXERCISE 2.6

CUBAS SA the company dedicated to the manufacture of steel tanks has decided to make
a forecast for sales that will calculate the next month of December. Sales data for the first
11 months of the year are:

Period Demand (units)

1 3600
2 2430
3 3510
4 3555
5 5580
6 3150
7 2790
8 2340
9 3960
10 4986
11 4230

Calculate the forecast by:

a) 3 months moving averages

b) 5 months moving averages
c) Exponential Smoothing for the following values α=0,1; α=0,5; α=0,9
d) What is the best value of α?

9
EXERCISE 2.7

Given the weekly demand data of a tennis Squash presented in the table below, calculate
the forecast for weeks 1 to 13 by moving averages with n = 3 weeks.

Week Demand Week Demand

1 56 7 53
2 48 8 60
3 39 9 56
4 71 10 51
5 42 11 45
6 51 12 65

EXERCISE 2.8

The company Entrevías, SA, has had over the last 24 months the ticket sales that show
the attached table. Perform forecast for January 2023, using the method of moving
averages of four months and exponential smoothing method taking α = 0.2. Which of the
two models has a lower error?

PERIOD SALES PERIOD SALES PERIOD SALES

January 21 280 September 21 225 May 22 211
February 21 305 October 21 251 Jun 22 211
March 21 251 November 21 243 July 22 200
April 21 279 December 21 207 August 22 225
May 21 234 January 22 227 September 22 208
Jun 21 215 February 22 230 October 22 202
July 21 279 March 22 258 November 22 186
August 21 285 April 22 208 December 22 210

10
 SOLUTIONS TO EXERCISES

Solution to exercise 2.3

a) 40 companies

Solution to exercise 2.4

a)

Concentration curve
1.2

1
Accumuoated market share

0.8

0.6

0.4

0.2

0
0 1 2 3 4 5 6 7 8
Companies (ordered from the biggest to the smallest one)

b) CR3= 0.68; CR5= 0.9; H= 0.1856

Solution to exercise 2.5

a) CR3= 0.83; CR4= 0.87; H=0.2602

b) Does not affect CR3, but does affect CR4 (0.9) and H (0.2626)

11
Solution to exercise 2.6

Forecast Method | Forecast period 12 MSE

3 months moving average 4392.0 2068132
5 months moving average 3661.2 972484
Smoothing α = 0.1 3699.2 1111426
Smoothing α = 0.5 4209.4 1405661
Smoothing α = 0.9 4293.8 1629578

The best method is the moving average of 5 months, in relation to smoothing: α = 0.1.

Solution to exercise 2.7

Week Forecast
4 47.7
5 52.7
6 50.7
7 54.7
8 48.7
9 54.7
10 56.3
11 55.7
12 50.7
13 53.7

12
Solution to exercise 2.8

Smoothing
Period MA4
α=0.2
1 280.0
2 280.0
3 285.0
4 278.2
5 278.8 278.4
6 267.3 269.5
7 244.8 258.6
8 251.8 262.7
9 253.3 267.1
10 251.0 258.7
11 260.0 257.2
12 251.0 254.3
13 231.5 244.9
14 232.0 241.3
15 226.8 239.0
16 230.5 242.8
17 230.8 235.9
18 226.8 230.9
19 222.0 226.9
20 207.5 221.5
21 211.8 222.2
22 211.0 219.4
23 208.8 215.9
24 205.3 209.9
25 201.5 209.9
MSE 680.2 742.52

The best method is moving averages, as is the one with a lower MSE. Our forecast is
201.5.

13