DATA AND METHODOLOGY ANALYSIS VIA STATISTICS

Im looking for a responsible person who can do this great job in 10-11 days.
It includes calculations and description; the first part of the Research I will send (Comovement of russian and other equity markets).
It should be about 40 pages.
My Supervisor asked to do the next things.
Everything you have to to acoordoing to it and change something if firstly it was incorrected.


Chapter 4: Data and methodology
o 4.1 Data and properties
- Introduce the data, their time range and frequency (put here your former chapter 4 that you need to better structure to give any relevant information as in the papers that you read)
- Descriptive statistics and other properties: Incorporate your former section 5.1. You also need to propose stationarity (i.e. unit root) tests, heteroskedasticity tests, and distributional tests (e.g. normality tests).
o 4.2 Methodology
- 4.2.1 Correlation analysis: You need to explain why it is useful here as a statistical tool to answer your research question. You need also to explain how you analyse them (i.e. introduce the related significance tests, the tested assumption, how it works and how you validate the assumption, if you validate it what is the meaning on a statistical and financial point of view).
- 4.2.2 Linear regression: You need to explain what you mean here. Provide related equations and explain how it works on a statistical viewpoint.
- Major comments:
o Descriptive statistics: You have 7 monthly time series, thus be careful to perform individual statistics and not sample based statistics unless you intend to do some cross-section (i.e. multivariate) study.
o Drop the GARCH part and DID
o Regression: You have 2 choices.
- Data are heteroskedastic and/or autocorrelated: You can perform a regression based on returns but choose a robust method (e.g. HAC method of Newey West so as to get the true significance statistics in your regression). You can run HAC regressions with Eviews software for example. If you know how to code, you can use Cran R software as well.
- Try to make returns conform to the linear regressions assumptions: You can try to check if the change in returns is not correlated (no autocorrelation) and not heteroskedastic.
o About the crisis impact: You can incorporate a dummy indicator in your regression and test for its significance.
- Choice 1: Dummy = 1 before crisis, 2 during crisis, 3 after crisis.
- Choice 2: Dummy = 0 before crisis, 1 during crisis, 0 after crisis.
o Using linear regressions to test for linkages between markets returns:
- Co-movements: Do the regression of Y_t on X_t plus Dummy_t
- Causality of X towards Y: Do the regression of Y_t on X_(t-1) and Dummy_t


About the recuiter
Member since Aug 27, 2017
Ethan
from Queensland, Australia

---

---

Looking for help? Checkout our video tutorial
How to search and apply for jobs

How to apply? Do you have more questions about the Job?
See frequently asked questions