1. Consider the following time series data.

Week 1 2 3 4 5 6
Value 18 13 16 11 17 14
Using the naïve method (most recent value) as the forecast for the next week, compute the
following measures of forecast accuracy.
a. Mean absolute error
b. Mean squared error
c. Mean absolute percentage error
d. What is the forecast for week 7?
2. Refer to the time series data in Exercise 1. Using the average of all the historical data as a
forecast for the next period, compute the following measures of forecast accuracy:
a. Mean absolute error

b. Mean squared error test

c. Mean absolute percentage error
d. What is the forecast for week 7?
3. Exercises 1 and 2 used different forecasting methods. Which method appears to provide
the more accurate forecasts for the historical data? Explain.
4. Consider the following time series data.
Month 1 2 3 4 5 6 7
Value 24 13 20 12 19 23 15
a. Compute MSE using the most recent value as the forecast for the next period. What is
the forecast for month 8?
b. Compute MSE using the average of all the data available as the forecast for the next
period. What is the forecast for month 8?
c. Which method appears to provide the better forecast?
5. Consider the following time series data.
Week 1 2 3 4 5 6
Value 18 13 16 11 17 14
a. Construct a time series plot. What type of pattern exists in the data?
b. Develop a three-week moving average for this time series. Compute MSE and a forecast
for week 7.
c. Use _ _ 0.2 to compute the exponential smoothing values for the time series. Compute
MSE and a forecast for week 7.
d. Compare the three-week moving average forecast with the exponential smoothing
forecast using _ _ 0.2. Which appears to provide the better forecast based on MSE?
Explain.
e. Use trial and error to find a value of the exponential smoothing coefficient _ that
results in a smaller MSE than what you calculated for _ _ 0.2.
6. Consider the following time series data.
Month 1 2 3 4 5 6 7
Value 24 13 20 12 19 23 15
a. Construct a time series plot. What type of pattern exists in the data?
b. Develop a three-week moving average for this time series. Compute MSE and a forecast
for week 8.
c. Use _ _ 0.2 to compute the exponential smoothing values for the time series. Compute
MSE and a forecast for week 8.
d. Compare the three-week moving average forecast with the exponential smoothing
forecast using _ _ 0.2. Which appears to provide the better forecast based on
MSE?
e. Use trial and error to find a value of the exponential smoothing coefficient _ that
results in a smaller MSE than what you calculated for _ _ 0.2.
7. Refer to the gasoline sales time series data in Table 6.1.
a. Compute four-week and five-week moving averages for the time series.
b. Compute the MSE for the four-week and five-week moving average forecasts.
c. What appears to be the best number of weeks of past data (three, four, or five) to use
in the moving average computation? Recall that MSE for the three-week moving
average is 10.22.
224 Chapter 6 Time Series Analysis and Forecasting
SELF test
SELF test
WEB file
Gasoline
8. Refer again to the gasoline sales time series data in Table 6.1.
a. Using a weight of 1/2 for the most recent observation, 1/3 for the second most recent,
and 1/6 for third most recent, compute a three-week weighted moving average for the
time series.
b. Compute the MSE for the weighted moving average in part (a). Do you prefer this
weighted moving average to the unweighted moving average? Remember that the
MSE for the unweighted moving average is 10.22.
c. Suppose you are allowed to choose any weights as long as they sum to 1. Could you
always find a set of weights that would make the MSE smaller for a weighted moving
average than for an unweighted moving average? Why or why not?
9. With the gasoline time series data from Table 6.1, show the exponential smoothing forecasts
using _ _ 0.1.
a. Applying the MSE measure of forecast accuracy, would you prefer a smoothing
constant of _ _ 0.1 or _ _ 0.2 for the gasoline sales time series?
b. Are the results the same if you apply MAE as the measure of accuracy?
c. What are the results if MAPE is used?
10. With a smoothing constant of _ _ 0.2, equation (6.8) shows that the forecast for week 13
of the gasoline sales data from Table 6.1 is given by F13 _ 0.2Y12 _ 0.8F12. However, the
forecast for week 12 is given by F12 _ 0.2Y11 _ 0.8F11. Thus, we could combine these two
results to show that the forecast for week 13 can be written
F13 _ 0.2Y12 _ 0.8(0.2Y11 _ 0.8F11) _ 0.2Y12 _ 0.16Y11 _ 0.64F11
a. Making use of the fact that F11 _ 0.2Y10 _ 0.8F10 (and similarly for F10 and F9),
continue to expand the expression for F13 until it is written in terms of the past data
values Y12, Y11, Y10, Y9, Y8, and the forecast for period 8.
b. Refer to the coefficients or weights for the past values Y12, Y11, Y10, Y9, and Y8. What
observation can you make about how exponential smoothing weights past data values
in arriving at new forecasts? Compare this weighting pattern with the weighting
pattern of the moving averages method.
11. For the Hawkins Company, the monthly percentages of all shipments received on time over
the past 12 months are 80, 82, 84, 83, 83, 84, 85, 84, 82, 83, 84, and 83.
a. Construct a time series plot. What type of pattern exists in the data?
b. Compare a three-month moving average forecast with an exponential smoothing forecast
for _ _ 0.2. Which provides the better forecasts using MSE as the measure of
model accuracy?
c. What is the forecast for next month?
12. Corporate triple A bond interest rates for 12 consecutive months follow.
9.5 9.3 9.4 9.6 9.8 9.7 9.8 10.5 9.9 9.7 9.6 9.6
a. Construct a time series plot. What type of pattern exists in the data?
b. Develop three-month and four-month moving averages for this time series. Does the
three-month or four-month moving average provide the better forecasts based on
MSE? Explain.
c. What is the moving average forecast for the next month?
13. The values of Alabama building contracts (in millions of dollars) for a 12-month period
follow.
240 350 230 260 280 320 220 310 240 310 240 230
a. Construct a time series plot. What type of pattern exists in the data?
Problems 225

SELF test
SELF test
SELF test
b. Compare a three-month moving average forecast with an exponential smoothing forecast.
Use _ _ 0.2. Which provides the better forecasts based on MSE?
c. What is the forecast for the next month?
14. The following time series shows the sales of a particular product over the past 12 months.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use _ _ 0.3 to compute the exponential smoothing values for the time series.
c. Use trial and error to find a value of the exponential smoothing coefficient _ that
results in a relatively small MSE.
15. Ten weeks of data on the Commodity Futures Index are 7.35, 7.40, 7.55, 7.56, 7.60, 7.52,
7.52, 7.70, 7.62, and 7.55.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use trial and error to find a value of the exponential smoothing coefficient _ that results
in a relatively small MSE.
16. The Nielsen ratings (percentage of U.S. households that tuned in) for the Masters golf tournament
from 1997 through 2008 follow (Golf Magazine, January 2009).
The rating of 11.2 in 1997 indicates that 11.2% of U.S. households tuned in to watch Tiger
Woods win his first major golf tournament and become the first African American to win
the Masters. Tiger Woods also won the Masters in 2001, 2002, and 2005.
a. Construct a time series plot. What type of pattern exists in the data? Discuss some of
the factors that may have resulted in the pattern exhibited in the time series plot for
this time series.
b. Given the pattern of the time series plot developed in part (a), do you think the forecasting
methods discussed in this section are appropriate to develop forecasts for this
time series? Explain.
c. Would you recommend using only the Nielsen ratings for 2002–2008 to forecast the
rating for 2009, or should the entire time series from 1997 to 2008 be used? Explain.
Year Rating
1997 11.2
1998 8.6
1999 7.9
2000 7.6
2001 10.7
2002 8.1
2003 6.9
2004 6.7
2005 8.0
2006 6.9
2007 7.6
2008 7.3
Month Sales Month Sales
1 105 7 145
2 135 8 140
3 120 9 100
4 105 10 80
5 90 11 100
6 120 12 110
226 Chapter 6 Time Series Analysis and Forecasting

WEB file
Masters
17. Consider the following time series.
t12345
Yt 6 11 9 14 15
a. Construct a time series plot. What type of pattern exists in the data?
b. Use simple linear regression analysis to find the parameters for the line that minimizes
MSE for this time series.
c. What is the forecast for t _ 6?
18. The following table reports the percentage of stocks in a portfolio for nine quarters from
2007 to 2009.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use trial and error to find a value of the exponential smoothing coefficient _ that results
in a relatively small MSE.
c. Using the exponential smoothing model you developed in part (b), what is the forecast
of the percentage of stocks in a typical portfolio for the second quarter of 2009?
19. Consider the following time series.
t1234567
Yt 120 110 100 96 94 92 88
a. Construct a time series plot. What type of pattern exists in the data?
b. Use simple linear regression analysis to find the parameters for the line that minimizes
MSE for this time series.
c. What is the forecast for t _ 8?
20. Because of high tuition costs at state and private universities, enrollments at community
colleges have increased dramatically in recent years. The following data show the enrollment
(in thousands) for Jefferson Community College from 2001 to 2009.
Year Period (t) Enrollment (1000s)
2001 1 6.5
2002 2 8.1
2003 3 8.4
2004 4 10.2
2005 5 12.5
2006 6 13.3
2007 7 13.7
2008 8 17.2
2009 9 18.1
Quarter Stock%
1st—2007 29.8
2nd—2007 31.0
3rd—2007 29.9
4th—2007 30.1
1st—2008 32.2
2nd—2008 31.5
3rd—2008 32.0
4th—2008 31.9
1st—2009 30.0
Problems 227

SELF test
SELF test
a. Construct a time series plot. What type of pattern exists in the data?
b. Use simple linear regression analysis to find the parameters for the line that minimizes
MSE for this time series.
c. What is the forecast for 2010?
21. The Seneca Children’s Fund (SCF) is a local charity that runs a summer camp for disadvantaged
children. The fund’s board of directors has been working very hard over recent
years to decrease the amount of overhead expenses, a major factor in how charities
are rated by independent agencies. The following data show the percentage of the money
SCF has raised that were spent on administrative and fund-raising expenses for
2003–2009.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use simple linear regression analysis to find the parameters for the line that minimizes
MSE for this time series.
c. Forecast the percentage of administrative expenses for 2010.
d. If SCF can maintain its current trend in reducing administrative expenses, how long
will it take SCF to achieve a level of 5% or less?
22. The president of a small manufacturing firm is concerned about the continual increase in
manufacturing costs over the past several years. The following figures provide a time series
of the cost per unit for the firm’s leading product over the past eight years.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use simple linear regression analysis to find the parameters for the line that minimizes
MSE for this time series.
c. What is the average cost increase that the firm has been realizing per year?
d. Compute an estimate of the cost/unit for next year.
23. FRED® (Federal Reserve Economic Data), a database of more than 3000 U.S. economic
time series, contains historical data on foreign exchange rates. The following data
show the foreign exchange rate for the United States and China (http://research
Year Cost/Unit ($) Year Cost/Unit ($)
1 20.00 5 26.60
2 24.50 6 30.00
3 28.20 7 31.00
4 27.50 8 36.00
Year Period (t) Expense (%)
2003 1 13.9
2004 2 12.2
2005 3 10.5
2006 4 10.4
2007 5 11.5
2008 6 10.0
2009 7 8.5
228 Chapter 6 Time Series Analysis and Forecasting

WEB file
ExchangeRate
.stlouisfed.org/fred2/). The units for Rate are the number of Chinese yuan to one U.S.
dollar.
a. Construct a time series plot. Does a linear trend appear to be present?
b. Use simple linear regression analysis to find the parameters for the line that minimizes
MSE for this time series.
c. Use the trend equation to forecast the exchange rate for August 2008.
d. Would you feel comfortable using the trend equation to forecast the exchange rate for
December 2008?
24. Consider the following time series.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use a multiple linear regression model with dummy variables as follows to develop
an equation to account for seasonal effects in the data. Qtr1 _ 1 if Quarter 1, 0 otherwise;
Qtr2 _ 1 if Quarter 2, 0 otherwise; Qtr3 _ 1 if Quarter 3, 0 otherwise.
c. Compute the quarterly forecasts for next year.
25. Consider the following time series data.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use a multiple regression model with dummy variables as follows to develop an
equation to account for seasonal effects in the data. Qtr1 _1 if Quarter 1, 0 otherwise;
Qtr2 _ 1 if Quarter 2, 0 otherwise; Qtr3 _ 1 if Quarter 3, 0 otherwise.
c. Compute the quarterly forecasts for next year.
Quarter Year 1 Year 2 Year 3
1467
2236
3356
4578
Quarter Year 1 Year 2 Year 3
1 71 68 62
2 49 41 51
3 58 60 53
4 78 81 72
Year Month Rate
2007 October 7.5019
2007 November 7.4210
2007 December 7.3682
2008 January 7.2405
2008 February 7.1644
2008 March 7.0722
2008 April 6.9997
2008 May 6.9725
2008 June 6.8993
2008 July 6.8355
Problems 229

SELF test
26. The quarterly sales data (number of copies sold) for a college textbook over the past three
years follow.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use a regression model with dummy variables as follows to develop an equation to
account for seasonal effects in the data. Qtr1 _ 1 if Quarter 1, 0 otherwise; Qtr2 _ 1
if Quarter 2, 0 otherwise; Qtr3 _ 1 if Quarter 3, 0 otherwise.
c. Compute the quarterly forecasts for next year.
d. Let t _ 1 to refer to the observation in quarter 1 of year 1; t _ 2 to refer to the observation
in quarter 2 of year 1; . . . ; and t _ 12 to refer to the observation in quarter 4
of year 3. Using the dummy variables defined in part (b) and also using t, develop an
equation to account for seasonal effects and any linear trend in the time series. Based
upon the seasonal effects in the data and linear trend, compute the quarterly forecasts
for next year.
27. Air pollution control specialists in southern California monitor the amount of ozone, carbon
dioxide, and nitrogen dioxide in the air on an hourly basis. The hourly time series data
exhibit seasonality, with the levels of pollutants showing patterns that vary over the hours
in the day. On July 15, 16, and 17, the following levels of nitrogen dioxide were observed
for the 12 hours from 6:00 A.M. to 6:00 P.M.
July 15: 25 28 35 50 60 60 40 35 30 25 25 20
July 16: 28 30 35 48 60 65 50 40 35 25 20 20
July 17: 35 42 45 70 72 75 60 45 40 25 25 25
a. Construct a time series plot. What type of pattern exists in the data?
b. Use a multiple linear regression model with dummy variables as follows to develop
an equation to account for seasonal effects in the data:
Hour1 _ 1 if the reading was made between 6:00 A.M. and 7:00 A.M.; 0 otherwise
Hour2 _ 1 if the reading was made between 7:00 A.M. and 8:00 A.M.; 0 otherwise
...
Hour11 _ 1 if the reading was made between 4:00 P.M. and 5:00 P.M.; 0 otherwise
Note that when the values of the 11 dummy variables are equal to 0, the observation
corresponds to the 5:00 P.M. to 6:00 P.M. hour.
c. Using the equation developed in part (b), compute estimates of the levels of nitrogen
dioxide for July 18.
d. Let t _ 1 to refer to the observation in hour 1 on July 15; t _ 2 to refer to the observation
in hour 2 of July 15; . . . ; and t _ 36 to refer to the observation in hour 12 of
July 17. Using the dummy variables defined in part (b) and t, develop an equation to
account for seasonal effects and any linear trend in the time series. Based upon the
seasonal effects in the data and linear trend, compute estimates of the levels of nitrogen
dioxide for July 18.
28. South Shore Construction builds permanent docks and seawalls along the southern shore
of Long Island, New York. Although the firm has been in business only five years, revenue
Quarter Year 1 Year 2 Year 3
1 1690 1800 1850
2 940 900 1100
3 2625 2900 2930
4 2500 2360 2615
230 Chapter 6 Time Series Analysis and Forecasting
WEB file
Pollution

WEB file
SouthShore
has increased from $308,000 in the first year of operation to $1,084,000 in the most recent
year. The following data show the quarterly sales revenue in thousands of dollars.
a. Construct a time series plot. What type of pattern exists in the data?
b. Use a multiple regression model with dummy variables as follows to develop an
equation to account for seasonal effects in the data. Qtr1 _1 if Quarter 1, 0 otherwise;
Qtr2 _ 1 if Quarter 2, 0 otherwise; Qtr3 _ 1 if Quarter 3, 0 otherwise.
c. Let Period _ 1 to refer to the observation in quarter 1 of year 1; Period _ 2 to refer to
the observation in quarter 2 of year 1; . . . and Period _ 20 to refer to the observation
in quarter 4 of year 5. Using the dummy variables defined in part (b) and Period,
develop an equation to account for seasonal effects and any linear trend in the time
series. Based upon the seasonal effects in the data and linear trend, compute estimates
of quarterly sales for year 6.