Business Statistics and Forecasting

Learning outcome 1: Demonstrate a comprehensive understanding of quantitative and qualitative forecasting approaches and underpinning principles in business statistics

Learning outcome 2: Select and critically apply appropriate statistical tools and techniques to identify, diagnose and resolve strategic issues

Learning outcome 3: Develop, manipulate, and interpret complex spreadsheet models for data-centric statistical approaches used to respond to strategic business issues.

Learning outcome 4: Critically analyse complex business scenarios, make sound judgements in the absence of complete data, and communicate analysis clearly to specialist and non-specialist audiences.

The Task

Question 1. Select a concept, theory, or framework from economics, business, or finance. Briefly describe this concept, theory, or framework. You may either use either a diagram or equations. Then, state the objective of your report. An example would be ..."this report would like to investigate whether exports and inflation affect the money supply in the UK". This should be the introduction section.

Question 2. In your methodology and data section, you are to accurately specify a multiple regression model to capture the relationship or concept described in task 1. Be sure that your model is well specified and all variables well defined. Then download a 40 (or more) time series sample for a minimum of three variables (i.e., one dependent variable and at least two independent variables). Note that any model specified that is not in agreement with your concept/theory/framework in your introduction section would receive a lower mark. You can download the dataset from multiple data sources although for consistency and ease of computation it is advisable to seek to stick to one data source.

a. Your data frequency for each series should be the same, i.e., monthly, quarterly, yearly, weekly, or daily. (Remember that the more data points you use, the more accurate your estimators will be).
b. Split the dataset into in-sample and out-sample.
c. Present the data professionally in the appendix of your report.

Question 3. Now move to the Results and Discussion section. In this section, you must show evidence of three aspects: pre-estimation analysis, regression estimations and post-estimation diagnostics.

All analyses and estimations carried out should be done using the EViews software only. You are strongly encouraged to present evidence from EViews in the appendix of your report. For example, you could cut/copy and paste the EViews results in the appendix (and accurately title them) and present the results on a formatted table in your report.

At this stage, all regressions should be carried out using only the in-sample data.

Remember also that you could estimate different variants of your original model. For example, you could add dummy variables or consider other functional forms (e.g., using natural logarithms) or add interaction and squared terms, where applicable.

Interpret and discuss your results.

Question 4. The final section should be the forecasting section. Again, this should only be done using the EViews software. Create one-step-ahead point forecasts for using your out-sample data. Provide a commentary of your results, including the forecasting evaluation.

Question 5. Provide a very brief conclusion to your report. Avoid making general conclusions which has nothing to do with your report.

Also note the following:

a. Include a title page that states your student ID (no name). An abstract is not mandatory. However, if you prefer to include that, it will not count towards the overall word limit.
b. The total report word limit is 1,500 words (-/+ 10%). Thus, the maximum word count is 1650. This will strictly be enforced. The word limit does not include tables, graphs, references, appendices, abstracts, and content pages.
c. You are strongly encouraged to adequately proofread your work prior to submissions. Significant errors could affect your overall score.
d. Your report should be properly referenced (Harvard Referencing System).