• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

need help with risk modelling

hey guys,

I have a favor to ask of you. I am thinking to write a thesis on modelling currency risk, and come up with some optimal strategies. However, I am quite weak in the risk modelling. I am hoping to learn more abt while writing this thesis.
I have write a draft for my proposal. But I am not too sure that it is applicable, or is there other things I should consider.

hope to hear from you.


the research q.

Is the time-varying hedge ratios estimated by the dynamic conditional correlation (DCC) more efficient than time-varying hedge ratio estimated by constant conditional correlation model (Garch)?

Data

Daily closing price of Spot (S) exchange rate for Euro/SEK, Euro/Japanese Yen, and Euro/US$ for the period in 2005-2015, the value of Euro in one SEK, one Japanese Yen, and one US dollar. By using the daily data can estimate the risk better than weekly data.

Assume 250 trading days per year; we expected to have 2500 observations.

One month Euro currency interest rate

One month Japanese Yen interest rate

One month interest in Sweden, and USA

Methodology

First examine the features and characteristics of exchange rate return for each of the currency, mean, variance, and distribution.

1) Does the exchange rates exhibit a prolonged depreciation/appreciate

*How long does the behavior (depreciation/ appreciation) persist to consider as significant?

*How can I test this? Use OLS-method, regression method to these results are significant?

2) Is the variance time-varying as suggested by literatures for the currencies in the sample?

Use graphs to show this behavior?

3) Has exchange rate return a fat-tailed distribution?

*Display in the graph

According to the articles, the exchange rate returns are time-varying mean, and has time-varying variance, with a fat-tailed distribution, use the data to show that the exchange rate return, it does follow have these features. After confirm the features of the exchange rate data, and then we can model the exchange rate for all the mentioned currencies. The model used are GLL-GARCH-student -t, and GLL-DCC-GARCH_student-t,the reasons are

1. If the exchange rate returns do indeed has the following features, Engel 2000, and Change et al (xxx) articles mentioned that performed the best. Therefore, it is interesting to test whether DCC-GARCH models the time-varying variance better for exchange rates than normal multivariate GARCH model.

Simulate data using the estimates generated from the above mentioned models. Bos et al 2000, used Gibb sampler by using markov chain monte carlo to simulate the GLL- GARCH-student-t models.

· Can Gibb sampler be used to simulate the GLL- DCC-GARCH-student-t models, if not, how can this are achieved?

Hedging instrument is the forward contract and uses the covered interest rate parity to estimate the forward rate.

When exchange rates are successfully simulated, apply the hedging strategies mentioned in Bos et al 2000, except, the alternative strategies, VaR and Sharpe Ratio, to check which one strategies has the best return, and minimal variance. The main focus of the thesis is to test whether DCC-GARCH improve the hedging ratio for the hedging strategies.

Hedging strategies:

1) Random walk

2) No hedge

3) 100% hedge

4) Optimizing the manager’s utility function (assuming the managers are risk neutral)
 
Top