CALIFORNIA STATE POLYTECHNIC UNIVERSITY, POMONA
College of business administration
Department of Finance, Real Estate and Law

Business Forecasting
FRL 363

 I. COURSE OBJECTIVES
Two quite distinct general techniques are used in forecasting time series. Traditional models (regression technique) capture the behavior of economic variable through a structural model based on theory. Time series models, on the other hand, concentrate on the dynamic characteristics of economic and financial data, but largely ignore economic and finance theory. This course investigates the two alternative methodologies.

II. PREREQUISITE:

OM 301, FRL 301, and Math 125

III. Text
a. Using Econometrics, Studenmund A. H.

b. EViews Computer Package

c. Introductory Business Forecasting, by Newbold and Boss.
(The book is in Reserve Book Area in Main Library).

IV. GRADING
Homework and quiz 10%
Two midterms each 20%
Final Exam 40%
Term Project 10%

V. IMPORTANT THINGS TO REMEMBER

1. Students are required to use blue books in midterm and final exams. Quiz will include short assay questions and problems. Midterms and Final exams include essay questions and problems.

2. Makeup exams are not given under any conditions, so please check your schedule before taking this course. Late projects and cases will not be accepted.

3. Punctual attendance is expected from student at all class meetings. If you must be absent in one or more of the class meetings, please let me know as soon as possible.

VI. COURSE OUTLINE AND READINGS:

A. Simple and multiple regression Ch. 1&2

B. Assumptions and tests Ch. 4&5

C. Dummy variables Ch. 3

D. Time Trend Models

E. Specification Ch. 6&7

F. Multicollinearity Ch. 8

G. Event study methodology Hand-out

H. Serial correlation Ch. 9

I. Heteroskedasticity Ch. 10

J. Time -series Methods: Hand-out
Simple exponential smoothing
Holt's linear trend model
Multiplicative seasonality

VII. GROUP TERM PROJECT GUIDELINES:

Each group should have no more than three students. Each group selects a variable as the dependent variable and prepares two forecasts of the dependent variable: one forecast by the regression method, and one forecast by the exponential smoothing method. The accuracy of the forecasts and the advantages of one forecast over the other should also be discussed in the paper.

How to write the project:

1. Select Variables
Study the stock and determine all the variables that can affect you stock. Make sure that you choose some macro-level variables like interest rates, or for example GNP that affects stocks on general, and some micro-level variables that mostly affect the company that you are studying.

2. Data
Remember from your finance classes that it is very difficult to predict stocks on a day to day basis. You will get much better results if you choose semiannual or annual data. However, if you don't have any other choice, you can use monthly data; just remember monthly data may result in low R square and insignificant t-values.

3. The Regression Model
Write down the regression model that you plan to estimate, giving attention to functional form (i.e., linear, log linear, etc.), and specification of the set of explanatory variables (be careful about the direction of causation between dependent and explanatory variables). Try to relate the regression model as closely as possible to the theory that motivated it. Set up the tests of hypotheses you propose to conduct, again relating these tests to the implications of underlying theory. Correct all the problems you might have like multicollinearity, serial correlation, heteroskedasticity, etc. Forecast with your regression and show the accuracy of your forecast by plotting the actual data and your forecasts. Report you R square, and the ratio of RMSE to average Y.

4. Exponential Smoothing Model
Look at the plot of the past observations of the variable under study and determine the model that in your opinion can forecast the data in the best way. Consider trend and seasonality in the past observation. After you determine the model, estimate the forecast of future values.

How to present your project:

1. Present a short history of the company and explain the logic for choosing the independent variables.

2. Present the first estimation of the model; discuss R square, t-values, and the problems of your regression.

3. Correct the problems and present the final version of your regression. Discuss the factors that your dependent variable is sensitive to. Discuss R square, t-values, consistency of your model with theory and etc.

4. Compare the forecast estimated by the regression model with the forecast estimated by the exponential smoothing model, and discuss the accuracy of each forecast and the advantage of each forecast over the other.

5. Remember you have only 15 to 20 minutes for your presentation, so be organized and efficient, spending too much time on presentation will have a negative effect on your presentation.

6. All the students should use Power point and Excel to present their project.

7. Both your classmates and I will grade your project, so a well-prepared presentation will have a significant effect on your grade.

The written project is due on the last day of class. Projects will be presented on Tuesday and Thursday of the tenth week.