Economic 3406---Summer 2001 (Boldt)
Problem Set # 2---Forecasting Computer Assignment
Due July 3 (worth 9 points)

Part I.  Computer Lab Exercise ---Complete and turn in the computer lab exercise for forecasting.  This will be handed out in class on Friday, June 29 (we will meet for part of the class on June 29 in the computer lab).  In addition, you need to complete and turn in answers to Parts II and III below.

Use the data below for parts II and III of the problem set. A, B, C, and D are the last four non-zero digits of your social security number.  Parts II and III will be completed outside of class.
 

PART II  Forecasting with Management Scientist
 
Year Period Quarter Sales
1997 1 1 1A
2 2 1B
3 3 1C
4 4 2D
1998 5 1 2A
6 2 2B
7 3 2C
8 4 3D
1999 9 1 1A
10 2 2B
11 3 3C
12 4 3D
2000 13 1 2A
14 2 3B
15 3 3C
16 4 4D
.

where A,B,C and D are the last four non-zero digits of your social  security number.

a) Using Management Scientist, determine the best moving average forecast for the above sales data
(only consider two-period and four-period moving averages).  Provide the Management Scientist printout
for each of the above moving average periods.

b) Using the Management Scientist software, determine the "best" exponential smoothing model for the sales
data by reporting MSE's.  Provide a listing of alternative MSE's based on smoothing constants (0,.1,.2,...,.9,1).
Also, provide the Mgt. Scientist printout of the best exponential smoothing model.

c) Using Management Scientist, compute a linear trend forecast for the  above sales data.  What is the slope
and the intercept of this forecasting equation? What does the slope value suggest about these data?  Include
a copy of this forecast.

d) Using Management Scientist, compute a trend and seasonal forecast for the above sales data.  How many
seasons did you select? Why?  How much are sales above average in the strongest quarter?  Include a copy
of this forecast also.

e)  Based on the above (parts 1-4), which forecasting technique is best for the above data? Why?
 

PART III  Linear Regression Forecasting with Excel

A real estate sales conmany is attempting to determine what factors impact the sales price of homes.  Data
for 10 homes is reported below:

     Sales Price         Home Size        Condition Rating 
     Y, $ thou            X1, hundreds   X2, 1 (poor) to 10
                               of square ft.    10 (best)

        60.0                2A                       5
        32.7                1B                      
        57.7                20                        9
        45.5                17                        3
        47.0                1D                        8 
        55.3                2B                        4
        64.5                24                        7
        42.6                1A                       6
        54.5                1C                        7
        57.5                2D                       2

where A,B,C and D are the last four non-zero digits of your SS#.

a) Find the "best fit regression line" for your data.  Assume Sales Price is the dependent variable.  To do a
regression in EXCEL, select "Tools" then "Data Analysis."  Provide a  printout of your data and the regression
output?

b) Based on the t-test, are the individual independent variables significant? Conduct the t-tests for each of the
independent variables at the .05 level of significance.

c) Based on the value of the r-squared term, what can you say about the relationship between the independent
variables and the dependent variable? Can the regression equation be used to accurately forecast sales price?

d) Based on your regression, what would be the forecasted sales price  for a house with a home size
of 20 (2000 sq. ft.) and a condition of 7?
 

NOTE:     PLEASE STAPLE THE PAGES IN ORDER.