Master of Science (MS)
Though Pearson's correlation coefficient provides a convenient approach to measuring the dependency between two variables, in the last few years, there has been a significant amount of literature cautioning against the use of Pearson's correlation coefficient, as it does not remain invariant under monotone transformations of the underlying distribution functions. Since we are interested in examining the dependency pattern observed by the return on the Sterling Pound with that of the Japanese Yen, we will use the notion of a copula to approximate the joint density function between the daily returns on the Sterling Pound and the Japanese Yen. In particular, we use a result that is fundamental to the development of copula theory, namely Sklar's Theorem, to examine the observed joint density function between the daily returns on the Sterling Pound and the Japanese Yen. We will attempt to capture the approximated joint density function using a theoretical Gaussian Copula Model. This comparison is performed in the case where the underlying marginal distributions are both uniform, as well as the case where the underlying marginal distributions are both gaussian.
Document Availability at the Time of Submission
Release the entire work immediately for access worldwide.
Coelho, Ryan, "Characterization of the dependency across foreign exchange markets using copulas" (2006). LSU Master's Theses. 2964.
Ambar N. Sengupta