Comments on: Credit models get even more complicated http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/ A slice of lime in the soda Thu, 23 May 2013 23:51:59 +0000 hourly 1 http://wordpress.org/?v=3.4.2 By: Harry Seldon http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-6728 Harry Seldon Sat, 12 Sep 2009 12:05:14 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-6728 "seems to accept as an article of faith that the problem with credit risk models was that they weren’t sophisticated and complicated enough." This freakes me out. Looks like people do not understand that complexity generates chaos. Conversely, to fight chaotic behaviour one needs regulation and simplicity... “seems to accept as an article of faith that the problem with credit risk models was that they weren’t sophisticated and complicated enough.”

This freakes me out. Looks like people do not understand that complexity generates chaos. Conversely, to fight chaotic behaviour one needs regulation and simplicity…

]]> By: David Harper http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5364 David Harper Tue, 11 Aug 2009 20:28:01 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5364 Read it last night. Turns out to be a nicely-written & succinct primer on the state of credit risk models; including highlighting drawbacks (e.g., Gaussian copula). highlights @ http://www.bionicturtle.com/learn/article/imf_on_recent_advances_in_credit_risk_modeling_credit/ Read it last night. Turns out to be a nicely-written & succinct primer on the state of credit risk models; including highlighting drawbacks (e.g., Gaussian copula).

highlights @ http://www.bionicturtle.com/learn/articl e/imf_on_recent_advances_in_credit_risk_ modeling_credit/

]]> By: Don the libertarian Democrat http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5344 Don the libertarian Democrat Tue, 11 Aug 2009 16:17:47 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5344 By the way, when I say that these models can be interesting, I'm referring to this paper by Li especially: http://www.defaultrisk.com/_pdf6j4/On%20Default%20Correlation-%20A%20Copula%20Function%20Approach.pdf Here's the point: "There exist three methods to obtain the term structure of default rates: (i) Obtaining historical default information from rating agencies; (ii) Taking the Merton option theoretical approach; (iii) Taking the implied approach using market prices of defaultable bonds or asset swap spreads." It strikes me that looking at these products and risks through these various methods can be quite interesting, even if dubious for actual investing. By the way, when I say that these models can be interesting, I’m referring to this paper by Li especially:

http://www.defaultrisk.com/_pdf6j4/On%20 Default%20Correlation-%20A%20Copula%20Fu nction%20Approach.pdf

Here’s the point:

“There exist three methods to obtain the term structure of default rates:
(i) Obtaining historical default information from rating agencies;
(ii) Taking the Merton option theoretical approach;
(iii) Taking the implied approach using market prices of defaultable bonds or asset swap spreads.”

It strikes me that looking at these products and risks through these various methods can be quite interesting, even if dubious for actual investing.

]]> By: Ken http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5336 Ken Tue, 11 Aug 2009 13:54:25 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5336 I keep reading that quote in the update, and failing to understand. It seems to be saying "We can do better at causing widespread disaster," but that makes no sense at all. Are they perhaps pointing out that the downside effects of the algorithms also have to be taken into account? I keep reading that quote in the update, and failing to understand. It seems to be saying “We can do better at causing widespread disaster,” but that makes no sense at all. Are they perhaps pointing out that the downside effects of the algorithms also have to be taken into account?

]]> By: Ken http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5335 Ken Tue, 11 Aug 2009 13:49:28 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5335 Felix, an interesting perspective on this dilemma may come from the hard sciences. Regulators were designed and commissioned to regulate the activities of relatively benign financial institutions, with tools more akin to Newtonian physics than the Quantum Calculus that now seems to be required. Management must seek to understand the implications of their decisions. When a portfolio consists of 6,000-10,000 loans, that is one thing - mundane, predictable, and "modelable." However, when the portfolio and securitizations and hedges value at 1,000 or 10,000 times the size of that "Newtonian" portfolio, something changes in the game. A model can predict behavior, but it will, under no circumstances, provide any more clarity than Newtonian calculus in an Einsteinian context. Just a thought, but perhaps we had best prepare ourselves for the contingency that behemoth banking is simply beyond prudent capability - I am not saying we can't do it, we clearly are doing it - rather I am saying that it isn't a good idea, systemically. Limitations. We have them. Felix, an interesting perspective on this dilemma may come from the hard sciences. Regulators were designed and commissioned to regulate the activities of relatively benign financial institutions, with tools more akin to Newtonian physics than the Quantum Calculus that now seems to be required.

Management must seek to understand the implications of their decisions. When a portfolio consists of 6,000-10,000 loans, that is one thing – mundane, predictable, and “modelable.” However, when the portfolio and securitizations and hedges value at 1,000 or 10,000 times the size of that “Newtonian” portfolio, something changes in the game. A model can predict behavior, but it will, under no circumstances, provide any more clarity than Newtonian calculus in an Einsteinian context.

Just a thought, but perhaps we had best prepare ourselves for the contingency that behemoth banking is simply beyond prudent capability – I am not saying we can’t do it, we clearly are doing it – rather I am saying that it isn’t a good idea, systemically.

Limitations. We have them.

]]> By: Dan http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5324 Dan Tue, 11 Aug 2009 10:18:29 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5324 The math allows us to deceive ourselves and pretend to know much more than we do. The math allows us to deceive ourselves and pretend to know much more than we do.

]]> By: Luis Enrique http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5323 Luis Enrique Tue, 11 Aug 2009 08:38:28 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5323 right. on the other hand, modeling risk isn't a complete waste of time, and I'd rather have models that incorporate "contagion in the network" than not. What you make of these models, is another matter. right. on the other hand, modeling risk isn’t a complete waste of time, and I’d rather have models that incorporate “contagion in the network” than not. What you make of these models, is another matter.

]]> By: dsquared http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5319 dsquared Tue, 11 Aug 2009 06:49:03 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5319 William: tell me more about this "understanding the company you are lending to" idea of yours; I believe that it may have applications to equity investment. William: tell me more about this “understanding the company you are lending to” idea of yours; I believe that it may have applications to equity investment.

]]> By: The Scientician http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5315 The Scientician Tue, 11 Aug 2009 04:22:34 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5315 This is another alarming sign that the Panglossian paradigm of efficient markets and the magic equations that describe them have survived the Great Recession. Indeed, the hydra has grown additional heads, or at least terms... This is another alarming sign that the Panglossian paradigm of efficient markets and the magic equations that describe them have survived the Great Recession. Indeed, the hydra has grown additional heads, or at least terms…

]]> By: Doug http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5312 Doug Tue, 11 Aug 2009 03:04:05 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5312 Who needs these fancy credit models when internet aggregators can provide you with a free credit score Oops - sorry wrong thread. Who needs these fancy credit models when internet aggregators can provide you with a free credit score

Oops – sorry wrong thread.

]]> By: Anal_yst http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5308 Anal_yst Tue, 11 Aug 2009 01:36:09 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5308 Just remember ladies & gentlemen: GiGo. You could have David Shaw and James Simons working at the SEC or any other regulator but the complexity of the models/math is irrelevant when the assumptions are rooted-in nonsense. I remember doing refis and acquisitions where debt/EBITDA (sometimes just senior debt!) was modeled to ramp-up to 12x post-deal, and these weren't LBO deals, they were corporate! A few tweaks in the underlying assumptions though and the deal got down to 6x with 4 strokes of the keyboard. I don't have the answer (yet?) but I'm quite certain there is, and will continue to be an inverse correlation between complexity of teh maths/models and regulators' ability to actually do their job. Just remember ladies & gentlemen: GiGo. You could have David Shaw and James Simons working at the SEC or any other regulator but the complexity of the models/math is irrelevant when the assumptions are rooted-in nonsense. I remember doing refis and acquisitions where debt/EBITDA (sometimes just senior debt!) was modeled to ramp-up to 12x post-deal, and these weren’t LBO deals, they were corporate! A few tweaks in the underlying assumptions though and the deal got down to 6x with 4 strokes of the keyboard.

I don’t have the answer (yet?) but I’m quite certain there is, and will continue to be an inverse correlation between complexity of teh maths/models and regulators’ ability to actually do their job.

]]> By: Eric Dewey http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5307 Eric Dewey Tue, 11 Aug 2009 00:51:59 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5307 Felix, the FT had a good editorial on the prevalence of deferring to numbers in executive decision-making today. Makes some points that are germane here... Felix, the FT had a good editorial on the prevalence of deferring to numbers in executive decision-making today. Makes some points that are germane here…

]]> By: MrM http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5306 MrM Tue, 11 Aug 2009 00:39:31 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5306 Some of these equations are the foundations of Basel II, so the regulators would know them. Over-reliance on model in socials sciences is bad, but turning regulators and bankers into Luddites is not much better either Some of these equations are the foundations of Basel II, so the regulators would know them.

Over-reliance on model in socials sciences is bad, but turning regulators and bankers into Luddites is not much better either

]]> By: Don the libertarian Democrat http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5304 Don the libertarian Democrat Mon, 10 Aug 2009 22:48:03 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5304 When I look at these papers, they seem very clever and interesting. As ways of looking at these investments and risks, they could be useful in a heuristic sense. But I'm amazed that people invest using these models. Although using math, they don't seem to generate precision. On the other hand, as I said, they are interesting ways of viewing these investments. But I'm no expert. I do think that everyone would get something out of reading the paper. However, I'm sure that I could find people who could help me more clearly understand the points being made and how dicey they are for real world use. I would certainly do this before allowing someone to spend my money, especially if I had a lot. When I look at these papers, they seem very clever and interesting. As ways of looking at these investments and risks, they could be useful in a heuristic sense. But I’m amazed that people invest using these models. Although using math, they don’t seem to generate precision. On the other hand, as I said, they are interesting ways of viewing these investments. But I’m no expert. I do think that everyone would get something out of reading the paper.

However, I’m sure that I could find people who could help me more clearly understand the points being made and how dicey they are for real world use. I would certainly do this before allowing someone to spend my money, especially if I had a lot.

]]> By: William Wild http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/comment-page-1/#comment-5298 William Wild Mon, 10 Aug 2009 21:36:31 +0000 http://blogs.reuters.com/felix-salmon/2009/08/10/credit-models-get-even-more-complicated/#comment-5298 David, I think anyone working in credit needs to actually try to understand the company whose risk they are taking. David, I think anyone working in credit needs to actually try to understand the company whose risk they are taking.

]]>