Students taking this module must carry out an empirical project. This project accounts for 60 % of the final mark. The word count must not exceed 3000 words. The project must be your own work. The main aim of this project is:
- To gain skill in the use of available specialised software to carry out an empirical project. This include Python, SPSS, R and Eviews.
- To acquire (improve) research skills including finding relevant literature and data sources.
- To be able to apply in practice different econometric models acquired using real world data.
You can choose any topic as long as it is consistent with the content of the MSc course you are undertaking. You might want to choose a topic based on an interesting data set or a published article. Do not agonise too long over choosing a topic. Once you have chosen a topic and collected the required data, do not be tempted to switch.
A replication study can be a good option. Get a published paper that use models similar to those taught in Predictive Analysis for Decision Making (or can be replicated using taught methods). A replication study involves repeating similar empirical exercise using (i) extended data sets, (ii) the same model on different data, (iii) testing the robustness and sensitivity of the published results and/ or (iv) any other extension – minor or otherwise – that one may add or make on the existing work.
Alternatively, you may start thinking of a topic that can be developed into a dissertation. In any case, the coursework is an empirical exercise, which requires the following:
i. Applying econometric models taught in this module.
ii. Interpreting econometric findings in line with existing/ known theory or conceptual framework.
iii. Finding data suitable for the study. The sample size need to be relatively large. Remember this module and the methods taught require a large data set. Here is a rough guide for you to follow:
Finding the appropriate data can be the most difficult part. Make it your first priority and check that the data is available before deciding on a topic. There are various databases from the university (DataStream1, Bloomberg) and other credible websites.
Make sure you have enough observations and variables. The sample size plays an important role in the precision of your results and what you can do. Make sure that you know the exact definition of the data. Terms like income and prices are not acceptable. Are the variable in constant or current prices? What is their base year? What is their coverage (Net or Gross, National or Domestic)? Are they seasonally adjusted?
The final project must be typed, structured and well organised. Do not just transcribe the results of performing dozens of regressions. Try to structure the interpretation of the results, pose questions and explain how the regressions provide answers to them. As you write up the results you are almost certain to think of something else you need to do. Therefore, start writing up early.
You should inform the reader about the things they are not expected to know and will need to know in order to understand what you have done. Do not copy out large chunks from econometric textbooks. It is likely I know most of that, just give a reference.
I strongly recommend that the project takes the following form. The length of each section will differ from project to project.
(I) First page (5 Marks): This should contain your name, title of the project and a short abstract of no more than 100 words. Give any acknowledgements for any help that you may have received whilst working on your project.
(II) Introduction (10 Marks): Introduce the subject and give some background information and refer to any relevant literature. This should follow the questions that you are going to attempt to answer and the significance of the results from this study.
(III) Theory/ Literature Review (20 Marks): Set out very briefly the economic theory or motivation for the topic. Use it to specify a model. Discuss the interpretation of the parameters (e.g. elasticities, marginal propensities, etc) and set out any a priori expectations of the signs and magnitude of the parameters if necessary. Also comment on the hypotheses to be tested (e.g. efficient market, stability of parameters, etc).
(IV) Data (10 Marks): Discuss the sources of the data, the sample size and frequency of the data, definitions of the variables. Describe the main features of the data with graphs. Submit the data and codes you used directly to the module leader.
(V) Econometric/ Statistical Model (15 Marks): Use the economic model and the structure of the data to choose an econometric model (e.g. linear regression, ARIMA, GARCH, Probit/ Logit etc.). Explain why you choose a particular econometric model. Report the results briefly.
(VI) Interpretation (15 Marks): Evaluate your chosen empirical econometric model in the light of the theory/ framework that was postulated and compare your results with those of past studies.
(VII) Conclusion (10 Marks): Explain the significance of your results and how they relate to the questions posed in the Introduction. Discuss future avenues for research.
(VIII) References2 (10 Marks): There should be a list of works cited at the end. Statements, assertions and ideas made in the project should be supported by citing relevant sources. Sources cited in the text should be listed at the end of the assignment in a reference list. Any material that you read but do not cite in the report should go into a separate bibliography. Bibliography though is not needed. Unless explicitly stated otherwise by the module teaching team, all referencing should be in Westminster Harvard format. If you are not sure about this, the library provides guidance (available via the library website pages).
(IX) Appendices: Report further and additional results when appropriate. The appendices are not part of the word count. The main results should not be reported here. The remaining 5 marks are for the overall presentation.
Before you hand your project in check that your project has your name, a title, an abstract, page numbers, references, the pages are numbered in the right order, the tables are numbered properly, and the data are submitted.