ECO-7026A

Summative Coursework 2

Timings

This assessment paper will be available from the 9^th of December 2022 until the 19^th January 2023.

Instructions

The TOTAL number of marks available for the paper is 100 MARKS:

• 50 marks for the code
• 50 marks for the findings

Word count

A recommended maximum word count is 1000 words.

Any work submitted beyond the recommended word count will be at the discretion of the marker to consider.

Academic Integrity

To maintain academic integrity students are required to:

• Not keep an electronic copy of this paper or share the content with others

• Ensure that all work produced is solely their own. Students are advised that their work will be checked using text matching software to determine any similarity with the work of other students or with other published materials.

The University policy regarding plagiarism and collusion will apply to this examination. A copy of the policy can be viewed here: Plagiarism and Collusion

Submission

The submission point for this assessment will close at 15:00 UK time, 19^th January 2023. It is recommended that you submit your work as soon as completed.

All coursework must be submitted as a self-contained zip file containing all codes, Stata datasets, and the written report in Word format.  Python codes may be submitted as Jupyter notebooks or standalone scripts, at your option.

Guidance on submitting un-typed answers, such as diagrams and graphs, can be found on Blackboard, beneath this paper.

Students are advised to keep a copy of their submitted work in case of unforeseeable technical difficulties.

If you encounter difficulties submitting your work and are unable to submit in the time allowed for your exam, please complete this form as soon as possible to contact the Assessment and Quality Office for help.

Assessment Task

In this project, you will undertake economic research using Python and Stata.

Below there are two Task Options. You will need to choose ONLY ONE TASK OPTION and follow the instructions to collect and organise data in a dataset and use that dataset to conduct the analysis, and report your findings.

Submissions should take the form of a zip file including:

• A report on your findings (max 1000 words excluding tables and figures). This format should match the empirical section of a research article, with a brief introduction and conclusion.
• Python and/or Stata code to reproduce your findings. Python codes should run in the environment used on the module; if you need to use additional Python packages, these should be clearly specified (including version number).  Stata codes should run on Stata/SE; any additional packages used should be available from ssc, and their versions stated.
• It should be clear how the code maps to the results you report, i.e. how to generate each result in your report. You may optionally include a brief README describing how to reproduce your results.
• Any data files apart from the main project dataset should be included in the zip file, along with information on the source (e.g. a URL or web page). You need not include the main project dataset, but for large datasets you should make clear which files are needed, how they should be named and the URL source.

Marking criteria

The project will be marked on based on your findings and code.

Findings (50%):

1) The data should be accurately and informatively described.

2) You should define a clear research question and hypothesis, which the empirical analysis should test. The suggested topics include possible research questions.

3) Results should be accurately and informatively reported.

4) Results should be reproducible. It should be clear how to reproduce your results by running your code, along with the relevant project data. Results should match the report.

5) Finally, the broader significance of your results should be briefly discussed.

Note that originality of the research question is not a marking criterion. It’s fine to replicate some existing published research, so long as you make clear that is what you are doing. Important! Your code, analysis and report should be done independently. Replicating other students’ work is plagiarism and it is punishable according to UEA guidelines.

Code (50%). Code should be:

1) Correct. Doing what it is intended to do without errors, and giving output which matches the description in your findings.

2) Complete. Starting with the raw project data and finishing with the reported results.

3) Well-structured. Functions and/or other programming constructs should be used where appropriate to aid reproducibility and maintainability.

4) Readable. Code and algorithms should be as clear and straightforward as possible. Variable and function names should convey intent. Comments should be used appropriately. Formatting should aid readability.

5) Secure. Code shouldn’t connect to the internet, read or write files except the contents of the zip file, or do other unexpected or dangerous things.

Criteria for a distinction mark

1. Code and findings should meet all the criteria above.
2. The report should be well structured (using section headers) and should contain: a brief introduction (complete with the research question), methods, results, and conclusions.
3. The figures and tables should be correct, readable and easy to understand without reading the report; use appropriate formatting, captions and notes.

Task Option 1:  The effects of the Russia-Ukraine war on stock prices

Create a dataset of historical daily stock prices for 20 companies between 1^st January 2021- 30^th November 2020 listed on the London stock exchange. Choose 10 technology companies and 10 food-producing companies.

Using this dataset, your task is to investigate:

1. Whether there was an effect of the start of the Russian-Ukraine war on the stock returns.
2. Has the war (e.g., major attacks and Russian advances) continued to affect the companies’ stock returns?
3. Have positive news such as the retake of Kherson by Ukrainian forces on November 11 affected these companies’ stock returns?
4. Are there any differences in the effects of war news on stock returns between technology companies and food-producing companies?

In order to create a dataset, you will need to download the data from an appropriate source (e.g. Yahoo Finance), append the data for different companies in a single file, and reshape the data to obtain a panel data structure. You will need to download price data and any characteristics of company you feel are relevant for your analysis (e.g., market capitalisation). You can also download other company characteristics, or financial performance indicators from your preferred if it helps your analysis. You will also need to independently research important dates or main events during the war in Ukraine and include them as appropriate variables in your dataset.

Using this dataset, you are required to use data visualisation and regression analysis to display and estimate the effects of the war in your overall sample and in different sub-samples. For question 1 you will need to include in your analysis a variable for the onset of the war and the following period. For questions 2 and 3 you will need to research the dates of 2-3 major events or announcements, some negative, some positive, and use that information to code variables for the timing of those events in your data and test their effect on stock returns. You can investigate question 4 by using econometric models which include interaction terms.

In your report, briefly describe the data, sample and variables used, econometric model, and present your results. Discuss your findings and why you find/do not find any differences between the stocks in terms of the effect of the war? What do your findings suggest about investor behaviour? Consulting published research and relating your findings to published research on the topic is recommended, but not mandatory. Specifically, if you refer to other research to discuss the broader significance of your results, it may help you to earn the marks for marking criterion 5) in Findings. But if you have a very good discussion of the implications of your findings without citing other research, marks will not be deducted.

Suggested reading to inspire you for this task:

Sun, M. and Zhang, C. (2022). “Comprehensive analysis of global stock market reactions to the Russia-Ukraine war”, Applied economics letters. https://www.tandfonline.com/doi/full/10.1080/13504851.2022.2103077

Task Option 2: Demographic mix, seasonality effects and reference dependence in performance at Parkrun events

Create a dataset of all individual performances at all the Norfolk-based Parkrun events between 1st January 2014 and 9^th December 2017, using the data available in the .csv files at https://www.dropbox.com/sh/7s9j3z8o45c9wtz/AAAvD19TAJhATfrG92WMu9b-a?dl=0

Using the appended dataset, your task is to investigate:

1. Compare the overall demographic mix of the participants in terms of gender and age across different event locations. Are there important differences in the composition of participants at different events across Norfolk?    
2. Compare average performance across months. Are there any seasonality effects in performance across the year? 
3. Do the performance data suggest reference-dependent behaviour?  For example, are participants more likely to finish just under a full-minute time (e.g. 23:50-23:59 compared to 24:00-24:09)? Do males show more refence-dependent behaviour than females?

In order to create the dataset, you will need to append all the separate event files into one document. The files contain information for individual runners regarding: event location, date of the event, individual performance and rank, individual characteristics such as age category and gender and other data. The variables contained in these files are described in the “Task Option 2 Readme” file. The final dataset should have all individual participants on rows and the variables on columns, similar to the separate raw data files. Note that parkrun is a casual event, and there are no checkpoints that confirm a finisher completed the course; look out carefully for unexpectedly slow times (from, e.g., running alongside a less-fast friend) or fast times (from, e.g., record-keeping error).

Using this dataset, for question 1 you will need to group data by event location and display appropriate graphs comparing the gender and age mix across locations. Think about displaying confidence intervals in your graphs to help you test the hypothesis that there is a 50-50% gender mix amongst participants.

For questions 2 and 3 you will need to further restrict the sample to work only with participants aged 16 to 45 years old, when we can expect performance to reach a peak. Display graphically the trends in average monthly performance in the restricted sample. Specify appropriate regression models and estimate regressions to test formally whether performance differs systematically across months. You will need to think carefully about how to account for the fact that the same individuals appear in the data more than once.    

For question 3 you can think about displaying the distribution of running times (think carefully about how to construct appropriate bins that will showcase any bunching around exact times). You can estimate regressions to test formally whether the males are more likely than females to display reference-dependent behaviour, for example whether they are more likely to finish the race in a tight window before exact times (e.g., 23:50-24:00). To estimate this kind of regression you will need to think carefully about how to define the outcome variable (e.g., “bunching before exact time”).     

In your report, briefly describe the data, sample and variables used, econometric model (including appropriate controls, correcting the standard errors where necessary), and present your results. Discuss your findings and what they suggest in terms of performance and strategic behaviour of runners. Consulting published research and relating your findings to published research on the topic is recommended, but not mandatory. Specifically, if you refer to other research to discuss the broader significance of your results, it may help you to earn the marks for marking criterion 5) in Findings. But if you have a very good discussion of the implications of your findings without citing other research, marks will not be deducted.

Suggested reading to inspire you for this task:

Allen et al, “Reference-Dependent Preferences: Evidence from Marathon Runners” (Management Science 2017), https://pubsonline.informs.org/doi/pdf/10.1287/mnsc.2015.2417

Dawson, P., Nasamu, E. and Turocy, T. L. (2018) UEA ECO Blog.  https://ueaeconomics.wordpress.com/2018/01/03/parkrun2/

 

Get expert help for Programming and Analytics and many more. 24X7 help, plag free solution. Order online now!