recommended topic for thesis

[ray jack]

Does anyone have some topics in time series analysis which i can work on for my msc thesis in statistics using any statistical software package

 

[moretodo]

You can try to develop an n factor stochastic model for evolution of interest rates - current models are not as accurate as they perhaps ought to be since they are typically 1 factor models. Alternatively, you may also at the effect of the evolution of interest rates along with any shocks like interest rate hike etc. on other asset prices like NYMEX, gold etc. You can dwell upon the effects of geopolitical factors solely on asset prices such as gold and petroleum as well.

 

[Ken Abbott]

GARCH is often used in a risk management context to proxy implied vol when there is no market-observable data. To what extent does GARCH(1,1) on any underlying asset price approximate the observed implied volatility? Do any of the GARCH variants work better? When do these measures converge and diverge? Did the financial crisis have any affect on usefulness of GARCH in this context?

 

[IntoDarkness]

modeling implied vol is a very good idea.

 

[moretodo]

modeling implied vol is a very good idea.

I personally disagree since the whole concept of implied volatility is due to the fact that the BS model is idealistic and has shortcomings in its mathematical ability to account for realistic dynamics of supply and demand in option pricing. Furthermore, it is generally really hard to determine whether the implied volatility values are diverging due to the the inefficacy of the underlying modelling method at hand or BS model's inability to price the option accurately thereby resulting in preposterous implied volatility values. Therefore, the whole exercise can be pointless from an application point of view. Ofcourse, the mathematical aesthetics of such a project would be appealing to some.

 

[IntoDarkness]

I personally disagree since the whole concept of implied volatility is due to the fact that the BS model is idealistic and has shortcomings in its mathematical ability to account for realistic dynamics of supply and demand in option pricing. Furthermore, it is generally really hard to determine whether the implied volatility values are diverging due to the the inefficacy of the underlying modelling method at hand or BS model's inability to price the option accurately thereby resulting in preposterous implied volatility values. Therefore, the whole exercise can be pointless from an application point of view. Ofcourse, the mathematical aesthetics of such a project would be appealing to some.

ya implied vol is not real vol but rather market price. modeling implied vol wont get any real vol but can give insight of how market evolves. once thats figured out, just use BS to easily convert the modeled vol cube to prices and make profit. easy said than done, but i believe many shops r doing it