Comments on: Summary of a Workshop on Recent Advances in Modelling Systemic Risk Using Network Analysis http://www.aleablog.com/summary-of-a-workshop-on-recent-advances-in-modelling-systemic-risk-using-network-analysis/ Alea Jacta Est Mon, 12 Jul 2010 20:04:31 +0000 hourly 1 http://wordpress.org/?v=3.0 By: David Merkel http://www.aleablog.com/summary-of-a-workshop-on-recent-advances-in-modelling-systemic-risk-using-network-analysis/#comment-3699 David Merkel Sat, 09 Jan 2010 05:35:00 +0000 http://www.aleablog.com/?p=4484#comment-3699 Okay, I will keep watching, but I have my doubts... Okay, I will keep watching, but I have my doubts…

]]> By: jck http://www.aleablog.com/summary-of-a-workshop-on-recent-advances-in-modelling-systemic-risk-using-network-analysis/#comment-3698 jck Fri, 08 Jan 2010 11:39:58 +0000 http://www.aleablog.com/?p=4484#comment-3698 I think total system leverage is not a sufficient metric, I agree with you on the prediction issue, but my view is that a visual representation of the network, while not predicting anything, makes it easier to spot the potential trouble spots. Look at the last link (slides: Note the clustering between key banks and ‘outside’ non bank (triangle) around page 29), no prediction there but "outside entity" surely should be closely monitored by regulators. This is early days, but some interesting potential... I think total system leverage is not a sufficient metric, I agree with you on the prediction issue, but my view is that a visual representation of the network, while not predicting anything, makes it easier to spot the potential trouble spots. Look at the last link (slides: Note the clustering between key banks and ‘outside’ non bank (triangle) around page 29), no prediction there but “outside entity” surely should be closely monitored by regulators.
This is early days, but some interesting potential…

]]> By: David Merkel http://www.aleablog.com/summary-of-a-workshop-on-recent-advances-in-modelling-systemic-risk-using-network-analysis/#comment-3697 David Merkel Fri, 08 Jan 2010 04:55:58 +0000 http://www.aleablog.com/?p=4484#comment-3697 jck, I find their models to be complex and hard to estimate. I was reading a journal on nonlinear dynamics, and there was an article on (something to the effect of) "If nonlinear models can't predict, what good are they?" Complex models have a difficult time estimating parameters with any degree of certainty. I would imagine that is true for network models. Personally, I think it would be simpler to just limit total system leverage. But maybe I dream. jck, I find their models to be complex and hard to estimate. I was reading a journal on nonlinear dynamics, and there was an article on (something to the effect of) “If nonlinear models can’t predict, what good are they?” Complex models have a difficult time estimating parameters with any degree of certainty. I would imagine that is true for network models.

Personally, I think it would be simpler to just limit total system leverage. But maybe I dream.

]]>