Thursday, January 28, 2010

The Gausian Copula:



Nice demo on the copula. I'm not in full agreement with the conclusion.

"That does not mean it was useless. The Gaussian copula provided a
convienent way to describe a relationship that held under particular
conditions. But it was fed data that reflected a period when housing
prices were not correlated to the extent that they turned out to be when the housing bubble popped."

From:In defense of the Gaussian copula, The Economist

You have to ask, if you were basing risk correlation on the copula, how can you incorporate the uncertainty and instability created by the fed maintaining artificially low interest rates for so long. Bubbles/Booms usually result from an accumulation of mistakes. Markets typically by their very nature consist decentralized decisions. What would lead so many decentralized decision makers to all make the same mistakes when it came to the financial crisis? This kind of coordination and the accumulation of mistakes is very likely explained by the fed's interventions. Decisions about risk, leverage, and asset prices would very likely become more correlated in an environment of centrally planned interest rates than under 'normal' conditions.

3 comments:

ranfuchs said...

Any assumption that the correlation is fixed, will lead to the wrong conclusion in the long term. Especially in the case of default, where we clearly know that any additional default increases the probability of subsequent defaults.

Then of course, we can also question the sensibility of using the normal distribution, when we clearly know that defaults are not normally distributed

burosys said...

http://burosysfurniture.blogspot.com :: Burosys Furniture is the best place to buy modern and contemporary furniture. We offer high-quality office chairs, office furniture, workstations, conference table, desking systems, canteen furniture, frezza italy, italian furniture and more in our innovative furniture collections.

Matt Bogard said...

ranfuchs- thanks for the additional analysis - that makes sense. I really am just getting introduced to copulas and financial economics and financial statistics.