Time Series:

Set of numerical data that is obtained at regular intervals of time.

Objective:
To identify and isolate influencing factors for predictive (forecasting) purposes.
Factors influencing a time series:
Trend – Overall tendency for the curve to rise or fall – inflation, technological change, productivity
increases. Denoted by T.
Cyclical – It’s a series of wave-like fluctuations – cycles of more than one year duration – changing
economic conditions generally produce cycles. Denoted by C. In practice cycles are often difficult to
identify and are frequently regarded as part of the trend.
Seasonal variations – Regular up-and-down pattern that repeats ANNUALLY, repeats itself year after year.
These fluctuations are found in quarterly, monthly or weekly data – influence of the weather, or calendar
related events such as school vacations and national holidays. Denoted by S.
When a cyclical pattern in our data has a period of 1 year, we usually call the pattern seasonal variation.
When a cyclical pattern has a period other than 1 year, we refer to it as cyclical variation.
Irregular/Random variations – These are the unpredictable/random fluctuations – no pattern at all.
Strikes, fires, sudden changes in taxes. Denoted by R or I.

Decomposition:
Separation of time series in to its components part. A complete decomposition separates the time series in
to four components: seasonality, trend, cycle and randomness.
Analysis of a time series:
Additive Model Actual data  Y  T + S + I
Multiplicative model: Actual data  Y  T × S × I  TSI

1|Page
De-seasonalised/Seasonally adjusted data:
Removing the seasonal pattern from the data is called seasonally adjusted data. Such data still include
trend, cyclical and irregular movements.
To remove seasonable influence in an actual seasonal data, the following formulae are used:
(a) Additive method:
De-seasonalised data  actual data – seasonal variation  Y − S
(b) Multiplicative method:

De-seasonalised data  × 100  × 100

actual sales

deasesonalised

Trend:
The trend can be obtained by using regression analysis to obtain the line of best fit, OR to smooth out the
fluctuations by the moving averages.

Moving average:
Moving average is simply a moving total divided by the number of periods comprising the total. Each
average eliminates seasonal influence and is located at the center of the period to which it relates. [
Note: The series of moving averages is often referred to as the trend. Moving averages have the property
that they tend to reduce the amount of variation present in a set of data. In the case of time series this
property is often used to eliminate unwanted fluctuations and the process is called smoothing of time
series. The resulting smoothed series will contain trend and cycle but not seasonality or the irregular
component. If we then divide each observation (Y) by the corresponding value of the moving-average
(MA) series, we will have isolated the seasonal and irregular components, i.e   ]

Q1. Following are the annual production of a commodity. Estimate the output in 2023. Assuming a
linear trend.

Year 2016 2017 2018 2019 2020 2021 2022

Production 38 41 45 48 52 56 63

2|Page
Q2. Fit a straight line trend by applying the method of least square to the following data, and estimate
the price of for the year 2023.

Year 2014 2015 2016 2017 2018 2019 2020 2021 2022
Price 2 6 7 8 11 11 10 12 17
Seasonal index:
Seasonal indices are calculated so that there average is 1. This means that the sum of the seasonal indices
equals the number of seasons.
If the seasons are months, the seasonal indices add to 12. If the seasons are quarters, then the seasonal
indices would add to 4, and so on.
Seasonal indices tell us how a particular season (generally a day, month or quarter) compares
to the average season.
Consider the monthly seasonal indices for unemployment in the following table:

Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
1.1 1.2 1.1 1.0 0.95 0.95 0.9 0.9 0.85 0.85 1.1 1.1
 For example, the seasonal index for unemployment for the month of February is 1.2 or 120%.
This tells us that February unemployment figures tend to be 20% higher than the monthly
average.
 Similarly, the seasonal index for unemployment for the month of August is 0.9 or 90%.
Then in August unemployment figures tend to be 10% lower than the monthly
average.
 In multiplicative model, the seasonal component is expressed as percentages, so they are centered
about 100.
 In the additive model, all three components are measured in the same units as the original data, so
that the seasonal and irregular components vary about zero.
Q3. Which of the following characteristic movement of a time series would you mainly associate?

seasonal cyclical long-term irregular

a fire in a factory delaying production for
3 weeks.
an era of prosperity.
an after Eid sale in a department store.
a need for increased wheat production due
to a constant increase in population.
the monthly number of millimeters of
rainfall in Karachi over a 5 year period.

Q4. For the observations 2, 6, 1, 5, 3, 7, 2 ; compute the moving average of order 3.

3|Page
Q5. Parts (a), (b), and (c) refer to the following moving average (MA) data. Complete the table prior to answering
the questions.
Year ⁄ 4-Quarter 4-Quarter Centered Ratio to
Quarter Value Total MA MA MA

2020 1 22
2 23
101 25.25
3 26 25.000 ?
99 24.75
4 30 24.875 ?
100 25.00
2021 1 20 25.250 ?
102 25.50
2 24 25.625 ?
103 25.75
3 28 26.000 ?
105 26.25
4 31 26.375 ?
106 26.50
2022 1 22 26.625 ?
107 26.75
2 25 27.250 ?
111 27.75
3 29
4 35

(a) What is the ratio to MA for the second quarter of 2021?

1) 0.79
2) 0.94
3) 1.08
4) 1.18
(b) What is the adjusted seasonal index for the first quarter?
1) 79.3
2) 81.3
3) 83.1
4) 106.1
(c) What is the seasonally adjusted value for the third quarter of 2020?
1) 24.5
2) 27.9
3) 32.1
4) 408.1

4|Page
Q6. Parts (a), (b), and (c) refer to the following moving average (MA) data. Complete the table prior to answering
the questions.
Year ⁄ 4-Quarter 4-Quarter Centered Ratio to
Quarter Value Total MA MA MA

2020 1 25
2 28
115 28.75
3 28 28.375 ?
112 28
4 34 28 ?
112 28
2021 1 22 28.5 ?
116 29
2 28 29.25 ?
118 29.5
3 32 29.75 ?
120 30
4 36 30.125 ?
121 30.25
2022 1 24 30.375 ?
122 30.5
2 29 30.75 ?
124 31
3 33
4 38

(a) What is the ratio to MA for the second quarter of 2021?

1) 0.241
2) 0.949
3) 0.957
4) 0.966
(b) What is the adjusted seasonal index for the first quarter?
1) 77.2
2) 78.1
3) 78.9
4) 79.0
(c) What is the seasonally adjusted value for the third quarter of 2022?
1) 27.17
2) 31.75
3) 31.99
4) 34.32

5|Page
7. Gradual shifting of a time series over a long period of time is called:
(a) periodicity (b) cycle (c) regression (d) trend

8. After detrending, the time series multiplicative model is represented as:

(a) Y=T×S×C×I (b) Y=S×C×I
(c) Y=T×S×I (d) none

9. The unadjusted mean values for the seasonal indexes are as follows:

Q1 Q2 Q3 Q4
92.8 100.4 99.8 105.6

Adjust the mean values to create seasonal indexes to 2 decimal places and deseasonalize a third
quarter value of 5,000.
10. The trend equation, y = 20 + 0.5x relates the quarterly sales for 20152016, with origin at the
first quarter of 2015, where x = time unit (one quarter) and y = sales (Rs 000). The seasonal
variations are:
Q1 Q2 Q3 Q4
80 90 120 110

Required:
Estimate the quarterly sales for 2016.
11. Actual sales for June and July were, respectively, 1680 and 1520 units. Respective seasonal
indexes for June and July are 105 and 95.
(a) Compute the percent change in actual sales.
(b) Compute the percent change in seasonally adjusted sales from June to July.

12. A deseasonalized series contains only which components?

(a) all four components (b) T and C

(c) T and I (d) T, C and I

13. The data given below are quarterly sales for a large computer firm, in Rs 100,000s.

Quarter 2022 - Q1 2022 - Q2 2022 - Q3 2022 - Q4 2023 - Q1 2023 - Q2

Sales 105 110 122 120 125 135

Compute the four-quarter moving average for the first four quarters and center it at the third
quarter.
(a) 114.25 (b) 119.25 (c) 116.75 (d) 119.5

6|Page
14-15.

14. The linear trend equation yˆ  500  60t (in Rs 000) was developed for annual sales from 2013 to
2021 with 2013 the base (zero) year. What are the estimated sales for 2023 (in Rs 000)?
(a) Rs 1160 (b) Rs 560 (c) Rs 1040 (d) Rs 1100

15. How much sales are increasing by?

(a) Rs 6000 per year (b) Rs 6000 per month
(c) Rs 5,60,000 per year (d) Rs 60000 per year

16-19.
A plastic manufacturing company performed quarterly time series analysis for demands over the
last 5 years (periods 1 through 20). The analysis resulted in the following trend equation and
seasonal indexes:
Tˆ  920.0  22.6t

Q1 Q2 Q3 Q4
0.75 1.04 1.21 1.00

16. The seasonal indexes :

(a) have not been corrected (b) have been corrected.
(c) incorrect. (d) cannot be interpreted
17. Based on the seasonal indexes, which quarter is expect to have 21% more demand than predicted
by the trend line?
(a) Q1 (b) Q2 (c) Q3 (d) Q4
18. Based on the seasonal indexes, which quarter is expect to have 25% less demand than predicted
by the trend line?
(a) Q1 (b) Q2 (c) Q3 (d) Q4
19. Using the trend line equation and the seasonal indexes, predict demand for the third period of the
next year.
(a) 1439.8 (b) 1190 (c) 1742.16 (d) 1195.24
20-26.
The following table shows the monthly sales figures (in Rs’ 000s) and seasonal indices (for
January to November) for a product produced by the CMA Company:

7|Page
 Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Seasonal index 1.2 1.3 1.1 1.0 1.0 0.9 0.8 0.7 0.9 1.0 1.1 -
Sales (Rs’ 000s) 9.6 10.5 8.6 - 7.1 6.0 5.4 - 6.4 7.2 8.3 7.4

20. The seasonal index for December is:

(a) 0.8 (b) 0.9 (c) 1.0 (d) 1.1
21. The deseasonalised sales (in Rs’ 000s) for March is closest to:
(a) 7.7 (b) 7.8 (c) 8.6 (d) 9.5
22. The deseasonalised sales (in Rs’000s) for June is closest to:
(a) 5.4 (b) 5.9 (c) 6.0 (d) 6.6
23. The deseasonalised sales (in Rs’000s) for August are 5.6. The actual sales are closest to:
(a) 2.7 (b) 3.9 (c) 5.6 (d) 5.9

24. The deseasonalised sales (in Rs’000s) for April are 6.9. The actual sales are closest to:
(a) 5.4 (b) 6.3 (c) 6.9 (d) 7.6
25. The seasonal index for February tells us that, over time, February sales tend to be
greater than the average monthly sales by:
(a) 0% (b) 10% (c) 20% (d) 30%
26. The seasonal index for September tells us that, over time, September sales tend to be less than the
average monthly sales by:
(a) 90% (b) 20% (c) 0% (d) 10%
27. The sales of kitchen units (in 000s) for CMA Products Ltd. for the years 2020-2022 are set out in
the following table:
Quarters

1 2 3 4
2020 3 5 8 4
2021 4 6 10 5
2022 5 8 11 7

An additive time series model is suggested.

Required:
(a) the trend in sales, using a centered moving average.
(b) compute the seasonal variations for each quarter and adjust the data for 2022.

8|Page
Year ⁄ Sales 4-Quarter 4-Quarter Centered Seasonal variations
Quarter (000) Total MA MA y−t

2020 1 3
2 5
20 5
3 8 5.125 2.875
21 5.25
4 4 5.375 −1.375
22 5.5
2021 1 4 5.75 −1.75
24 6
2 6 6.125 −0.125
25 6.25
3 10 6.375 3.625
26 6.5
4 5 6.75 −1.75
28 7
2022 1 5 7.125 −2.125
29 7.25
2 8 7.5 0.5
31 7.75
3 11
4 7

Q1 Q2 Q3 Q4
2020 2.875 −1.375
2021 −1.75 −0.125 3.625 −1.75
2022 −2.125 0.5
Total −3.875 0.375 6.5 −3.125
Mean −1.938 0.188 3.25 −1.563

Sum of mean  − 0.063

.
Required adjustment (quantity to be added)   0.01575
Adjusted seasonal variations:
Q1 Q2 Q3 Q4
−1.938 + 0.01575 0.188 + 0.1575 3.25 + 0.01575 −1.563 + 0.01575
 −1.92225  0.20375  3.26575  −1.54725

9|Page
The adjusted data for 2022 will be:

Q1 Q2 Q3 Q4

Actual Data 5 8 11 7
Seasonal Variation −1.92225 0.20375 3.26575 −1.54725
5 – (−1.92225)  8 – 0.20375  11 − 3.26575  7 – (−1.54725) 
Adjusted Data
6.922 7.796 7.734 8.547

28. A manufacturer computes the seasonal indexes for its monthly sales. Which of the following
statements about the index are correct?
(a) The sum of the 12 monthly index numbers, expressed as percentages, should be 1200.
(b) An index of 85 for May indicates that sales are 15% lower than the average monthly
sales.
(c) An index of 130 for January indicates that sales are 30% above the average monthly
sales.
(d) The index for any month must be between zero and 200.
(e) The average percent index for each of the 12 months should be 100.
29. The time series component which reflects a regular, multi-year pattern of being above and below
the trend line is:
(a) a trend (b) seasonal (c) cyclical (d) irregular
30. The time series component that reflects variability during a single year is called:
(a) a trend (b) seasonal (c) cyclical (d) irregular
31. The time series component that reflects variability due to natural disasters is called:
(a) a trend (b) seasonal (c) cyclical (d) irregular
32. The time series component that reflects gradual variability over a long time period is called:
(a) a trend (b) seasonal (c) cyclical (d) irregular
33. Visual inspection of the data will help the forecaster identify:
(a) trend (b) seasonality (c) linearity (d) nonlinearity (e) all the above
34. Which time-series component is said to fluctuate around the long-term trend and is fairly irregular
in appearance?
(a) trend (b) cyclical (c) seasonal (d) irregular.
35. When a time series contains no trend, it is said to be
(a) nonstationary (b) seasonal (c) nonseasonal (d) stationary

10 | P a g e
36. The sales of a product are recorded monthly for 24 months. The four-point (centered) moving
averages are calculated and plotted on a graph. How many moving average points are plotted?
(a) 20 (b) 21 (c) 22 (d) 24

37. The following data represents a time series for the last seven days:
2 6 10 6 10 14 10
Which ONE of the following moving averages would result in a straight-line graph?
(a) 2 point (b) 3 point (c) 4 point (d) none of these

38. The underlying trend in the demand for a particular product is constant (flat), and is subject to
quarterly seasonal variations as follows:

Q1 Q2 Q3 Q4
1.5 1.5 0.5 0.5

Assume a multiplicative model is appropriate.

If the demand for the last quarter, Q2 was 240 units, then the forecasted demand for the next
quarter, Q3 is:
(a) 80 units (b) 100 units (c) 120 units (d) 140 units
39-40. Following are the first five values of a quarterly time series. The multiplicative model is
appropriate and a four-quarter (centered) moving average will be used:

Year Quarter Time Series Value (Yt)

1 1 36
2 24
3 16
4 20
2 1 44

39. An estimate of the trend-cycle component (TC) for Quarter 3 of Year 1, when a four-quarter
moving average is used, is
(a) 24 (b) 25 (c) 26 (d) 28

40. An estimate of the seasonal-irregular (SI) component for Quarter 3 of Year 1 is

(a) 0.64 (b) 1.5625 (c) 5.333 (d) 30

11 | P a g e
41-44. Consider the following time series:

Year (t) Yt
1 7
2 5
3 4
4 2
5 1

41. The slope of linear trend equation is:

(a) − 1.5 (b) 8.3 (c) 1.5 (d) − 8.3

42. The intercept is:

(a) − 1.5 (b) 8.3 (c) 1.5 (d) − 8.3

43. In which time period, does the value of Yt reach zero?

(a) 0 (b) 0.181 (c) 5.53 (d) 4.21

44. The forecast for period 10 is:

(a) − 6.7 (b) 23.3 (c) 15 (d) 6.7

Answers:
5. (a) 2 (b) 2 c) 1 6. (a) 3 (b) 3 (c) 2 7. (d) 8. (b)
9. 4992.5 10. 17.6 , 20.25 , 27.6 , 25.85 11. (a) decreased by 9.52% (b) no change

12. (d) 13. (c) 14. (d) 15. (d) 16. (b) 17. (c)
18. (a) 19. (c) 20. (c) 21. (b) 22. (d) 23. (b)
24. (c) 25. (d) 26. (d) 28. (a), (b), (c), (e) are correct. 29. (c)
30. (b) 31. (d) 32. (a) 33. (e) 34. (b) 35. (d)
36. (a) 37. (a) 38. (a) 39. (b) 40. (a) 41. (a)
42. (b) 43. (c) 44. (a)

12 | P a g e