Eco 309 economic forecasting

ECO 309

Class Project Guideline

In this project, you are using company revenue data (Company Revenues V16) along with GDP, price index, and employment (in Macroeconomic Data) to analyze and forecast the revenue for the company that is assigned to you.

For analysis, make sure you provide the corresponding Minitab output and provide word explanations.

I. Introduction: use two to three sentences describing the company: including the year the company was established, the products/services the company provides, and the location of the headquarter.

II. Analysis:

1. Create a time series plot and an autocorrelation graph for the revenue data. Do you observe any trend or seasonality? Explain.

2. Based on what you observe in part 1, choose one model from the following three smoothing methods to estimate the revenue data: simple exponential smoothing, Holt’s smoothing (double exponential), and Winter’s smoothing. Note that you do not need to estimate all three models. Pick only one model that you believe to be the best choice. methods to estimate the revenue data. Choose your own parameters for level, trend and seasonality. Explain why you choose a multiplicative or an additive model. Identify the accuracy measures.

3. Use a multiplicative model to estimate the revenue data and compare the decomposition model with model in part b. Which model is better?

4. Use the better model in c to predict for the next four quarters. Report the prediction intervals.

5. Multiple regression: use the three macro economics variables, GDP, price index and employment as the independent variables and company revenue as the dependent variable and run a multiple linear regression. Performance the following analysis

a. Interpret regression coefficients; do they make sense to you?

b. Which variables are significant? Which are not?

c. Do you have a good overall fit of the model?

d. Check the residuals. Are all assumptions satisfied?

e. Is there multicollinearity?

f. Check for autocorrelation using the DW statistic. Do you have autocorrelation?

g. If you observe either multicollinearity or autocorrelation in the residuals, implement a solution and re-estimate the model.

The part will be on the final. You can work on it ahead of time to save time during the final exam.

6. ARIMA

a. Obtain the partial autocorrelation graph. Together with the time series plot and autocorrelation graph, determine the appropriate ARIMA model(s)

b. Run the ARIMA model(s) of your choice. Report the results. If you run several ARIMA models, mention the model specifications and report the results from the best one.

c. Compare the ARIMA model with the results in 2. What model is the best?