RESIT Assessment – Statistical Exercise.
Instructions
Students should refer to the original assignment instruction outlined in the course guide.
Consider any feedback you may have received and submit your revised assignment. In addition, include a new section at the end of the assignment (up to a maximum of 200 words) in which you comment on what you changed in your assignment in response to any feedback received or your reflection on the task.
The FEEDBACK IS :
Youhaven’tsubmittedtheexcelwiththedataandtheregressions. Thesubmittedexcelfileistheexamplethatispostedonmoodle.
Theintroductionandthediscussionon CAPM areveryconfusing.
Itisnotclearwhatsortofdataisconsideredfortheanalysis.
Verygeneraldiscussionisgivenonregressionanalysis.
Whydoweneedhistograms?
Needtocollectdataanddotheregressionanalysisinordertogetsomeconclusionsontheperformanceof CAPM.
We test the CAPM for this coursework; you will use one market index and a list of 10 companies’ weekly data and these companies should be from at least two sectors. Consider three sample periods: (1) Jan 2006 – Dec 2008, (2) Jan 2009 – Dec 2011 and (3) Jan 2012 – Dec 2014. The required data can be downloaded from Yahoo finance: http://uk.finance.yahoo.com/
Tasks:
(1) Make sure that your choice of companies and sectors would capture high market capitalisation.
(2) Using data for the entire sample period, run time-series regression on each of the selected companies onto a constant and market excess return and verify whether there exists a significant beta.
(3) Report the t-static for alpha and the R^2 for each company.
(4) Do a cross-sectional regression: for all i = 10.
(5) Discuss your results and merits and demerits of CAPM analysis.
(6) Discuss whether your results are sensitive to sector characteristics.
(7) Word limit: 1500
Brief notes:
1. Why use simple regression to estimate ß? Here dependent variable is stock return of individual firms (yi) and independent variable is market return (xi).
Because ßi=sim/(sm^2)=Covariance (market return(xi), individual stock return(yi))/variance of market return (xi)
In simple regression,
So the coefficient is actually ßi=sim/(sm^2)
