Assignment This assignment is worth 10% of the total mark. It should be handed in to the drop box in LMS (in a pdf form) by Monday 5pm Week 12 (12th October). Your report should contain your answers to the questions from Sections 1 to 3 below, and with supporting Eviews outputs for Section 3. Section 1 (3 marks: You can start this section after Week 4 lecture)
Read the following article: Harford, T. (2014), ‘Big data: A big mistake?’, Significance 11(5), 14–19. Question: Critically evaluate the main points of the article using three bullet points, in less than 150 words in total. Critical evaluation means
• To give your opinion on something • To support your opinion (with evidence where possible). • Note: Critiquing is NOT simply stating that something is “bad”. • Weigh up strengths and weaknesses. • Appraise the worth of something – test assumptions – judge the worth of an argument
Your points of evaluation may include the following (but not limited to):
• Correlation vs. causation • Importance of theories or insights in statistical analysis • Multiple testing problem • Sampling error • Sampling bias • Big data hubris
In providing your answers, you can also refer to the contents of Lecture in Week 4 (Statistical Significance in Empirical Research) Section 2 (3 marks: You can start this section after Week 7 lecture) The following paper is published in Kaplanski, G., Levy. H. 2010, Exploitable Predictable Irrationality: The FIFA World Cup Effect on the U.S. Stock Market, Journal of Financial and Quantitative Analysis, 45(2), 535- 553. However, you do not need to read the article to complete this part of the assignment.
Question: Critically evaluate the above statistical results in relation to the claim that the FIFA World Cup has a significant effect on stock market return. Use three bullet points, in less than 150 words in total. Your points of evaluation may include the following (but not limited to):
• Economic plausibility of the model • Economic significance of effect size • Validity of statistical significance based on the p-value criterion • Possible sampling bias
Section 3 (4 marks: You can start this section after Week 8 lecture) One of the problems in the analysis in Section 2 is that the sample covers a long period with a massive sample size. This will make the p-value very small regardless of economic significance of effect size, which can bias the outcome of statistical inference. In addition, there is a danger that the results may be distorted due to structural changes. The latter include the changes in trading behaviour of investors’, trading technologies, investment patterns, the development and regulatory framework of stock market, and market crash, among others. In this section, we estimate the regression model over a different subsample of 20 years, which includes 5 FIFA World Cup periods. You should choose a period based on the last digit of your Student ID as below:
Period Last digit of your Student ID 1988 – 2007 1, 6 1990 – 2009 2, 7 1992 – 2011 3, 8 1994 – 2013 4, 9 1996 – 2015 5, 0
The data is stored on the file “fifadat.wf1”, which is available from LMS. Note that this file is too large for Student version of Eviews, so you will need to use the full version available from Virtual Desktop. Note that the sample period can be adjusted in Eviews by changing the sample period as below:
Question: Critically evaluate the adequacy of the proposed model using a range of model diagnostic tests listed below. Choose three from the list below. Discuss your results, combining the evidence from these three tests, in less than 150 words in total.
• Bayes factor for H0: β5 = 0 against H1: β5 0 • Signal-to-noise ratio or Cohen’s f2 • RESET • A test for non-normality, heteroskedasticity or autocorrelation (either using data
visualization or statistical tests)
Your discussion should be based on the relevant Eviews outputs, which should be included in your report or attached as an appendix. A zero mark will be given to Section 3 if no Eviews output is provided, or if an incorrect data set is used. End of the Assignment