# Economics 694

Prerequisite:  Admission to the MBA program or graduate status, STAT 205 (or equivalent) and MATH 151 (or equivalent).

Credit Hours: (3)

Forecasting involves making the best possible judgment about some future event. Topics coverage includes introduction to forecasting, a review of basic statistical concepts, exploring data patterns and choosing a forecasting technique, moving averages and smoothing models, regression analysis, time series analysis, the Box-Jenkins (ARIMA) methodology, and judgmental elements in forecasting. Students will be trained in using computer-based models, databases, and programs.

Detailed Description of Content of Course

Get to know your data and the software:

• Overview of forecasting:  how do you predict the future? What kind of accuracy is possible?
• Where to obtain data; data sources on the Web?
• How to move data around:  useful things you can dowith your word processor and spreadsheet.
• What to look for in data:  seasonality, inflation, trends, cycles, etc.
• How to transform data to reveal its structure; deflation, logging, seasonal adjustments.
• Illustration of basic operations in Minitab.
• Famous forecasting quotes.
• Stationarity and differencing.
• The logarithm transformation.

Introducing to forecasting:

• Forecasting a stationary series:  the “mean” model.
• Forecasting a non-stationary series I:
• Geometric random walk:  the basic stock price model
• Three types of forecasts:  estimation period, validation period, and long-term extrapolation.
• How to evaluation forecast errors and compare models
• Mean
• Linear trend
• Random walk
• Random walk with growth
• Geometric random walk

Three types of forecasts:  estimation period, validation period, validation period, and the future

Modeling of seasonality

• General considerations in working with seasonal data: causes of seasonality, stability of seasonal patterns
• Seasonal random walk; and seasonal random trend models
• Seasonal adjustment by the ratio-to-moving-average method
• Trend/cycle decomposition of time series.
• Seasonal differencing
• Seasonal random walk
• Seasonal random trend

Averaging and smoothing models

• Simple moving average model
• Exponential smoothing model
• Combination of smoothing and seasonal adjustment
• Robust models for noisy data
• Massively parallel forecasting
• Moving averages and smoothing models
• Averaging and exponential smoothing models
•  Time series regression models:
• Fitting time series regression models
• What to look for in regression output
• What’s a good value for R-squared?
• Not-so-simple regression models

Introduction to ARIMA models

• Naïve + Autoregressive + Exponential Smoothing = ARIMA
• Using ACF and PACF plots to determine the “signature” of a time series.
• Fitting non-seasonal ARIMA models
• The spectrum of ARIMA models
• Introduction to ARIMA: non-seasonal models
• Identifying the order of differencing
• Identifying the orders of AR or MA terms
• Rules for identifying ARIMA models

Seasonal Models

• Identification of seasonal models
• Examples of seasonal model-fitting
• Estimation of ARIMA models
• Seasonal ARIMA models

Automation

• Combination of ARIMA and regression models
• Recap of steps in choosing a forecasting model
• Automatic forecasting software
• ARIMA models with repressors
• Steps in choosing a forecasting model
• Forecasting flow chart
• Automatic forecasting software

Politics and ethics

• Politics and ethics
• Review of models: what to use and when
• Political and ethical issues in forecasting
• How to avoid trouble
• Data transformations and forecasting models:  what to use and when

Detailed Description of Course

Lectures and class discussion will be the primary teaching methods.  The course will combine the use of lectures, videotapes, guest speakers, reading materials, project reports, and case analysis.  Students will receive hands-on computer experience.  In addition, forecasting techniques based on subjective and judgmental methods and their applications in long-range forecasts will be discussed at length.

Goals and Objectives of the Course

After successfully completing this course, students will:

Assessment Measures

1. Know how and where to obtain data needed for making forecasts
2. Understand the hidden patterns in the data
3. Recognize the need for adjusting data for inflation, trend, seasonality, and cycles
4. Understand naïve forecasting models
5. Know and be able to apply statistical concepts
6. Understand the topics of autocorrelation analysis and the use of Box-Jenkins techniques
7. Understand judgmental forecasting techniques
8. Recognize the importance of selecting the correct forecasting technique and the means by which the data derived should best be analyzed
9. Solve forecasting cases using the appropriate forecasting techniques
10. Become familiar with structural forecasting models
11. Be familiar with the application of ARIMA models
12. Compare forecasts derived from different techniques
13. Appreciate the ethical and political issues involved in forecasting

Student progress will be evaluated through several means that may include written examinations, oral presentations, and extended project reports.  The course will be offered as a hands-on course, which would place particular emphasis on application of theoretical concepts.  The specific evaluative measures will be defined by the instructor and would be consistent with the stated objectives of the course.

Other Course Information

The instructor will be expected to integrate substantial readings from the current and/or original professional literature into the course.

Approval and Revision Dates

4/17/00 New course approved