PDA

View Full Version : Credit risk modeling market to grow 1000% by 2008


DW Simpson
09-29-2004, 09:43 AM
From Business 2.0 this month:

Credit Risk Modeling

2004 Market Size: $100 million
2008 Market Size : $1 billion
Companies to watch: Kamakura, Moody's KMV, MSCI Barra

What's the probability that a firm will default on its debt? That's the hottest question in finance, thanks to the wild popularity of credit default swaps - expected to approach a whopping $5 trillion in nominal value this year. To value a credit swap, you need a guess about the chance of default.

That's why financial institutions spent an estimated $100 million on credit risk modeling this year. In the last 12 months, Standard & Poor's launched a credit-risk modeling tool, as did risk-management consultancy MSCI Barra, and top banks have developed in-house models. Industry leader Moody's KMV draws on the work of Nobel Laureate Robert Merton, who attacked the problem by predicting when a firm's assets will fall below liabilities. New models mine correlations between say, financial ratios and historical defaults. The introduction of Dow Jones iTraxx, a default swap index, is providing data to make models more accurate and also stimulating demand for modelers. Says Stanford University finance professor Darrell Duffie, "Suddenly, every week I get a call from a recruiter looking for a Ph.D. who can do this stuff."

JGatsby
09-30-2004, 12:21 AM
Why isn't the SOA in there like a dirty shirt singing all of our praises? I'd love to do that type of work, although it doesn't really seem like brain surgery to me.
You look at the cash burn, and the balance sheet, and a few accounting ratios over a period of time, and you can usually get a pretty good idea of where a company is going to be in a year or two. That's why subdebt lenders use covenants, so they can get out when things start to go south.

WinnieThePooh
10-02-2004, 12:29 AM
Why isn't the SOA in there like a dirty shirt singing all of our praises? I'd love to do that type of work, although it doesn't really seem like brain surgery to me.
You look at the cash burn, and the balance sheet, and a few accounting ratios over a period of time, and you can usually get a pretty good idea of where a company is going to be in a year or two. That's why subdebt lenders use covenants, so they can get out when things start to go south.

It is more sophisticated than you portray. You need a high level of mathematics. Actuaries don't have that level in their studies. Writing investment or finance track graces the surface. You don't ask an actuary to do high level derivative pricing unless they have advanced degrees in applied mathematics. Credit derivatives falls in the same category.
This highlights the fallacy of the actuarial profession.

JGatsby
10-02-2004, 03:42 AM
I understand that Credit Risk modeling is sophisticated (and was in no way trying to argue otherwise) but I question it's applications.

The success of any given business is contingent on so many unquantifiable (and quantifiable) variables. The role of management, marketing, product demand, input prices and focus all drive the success of a business. Until you really dig and to the leg work by speaking with management, reading the footnotes in annual reports, analyzing the market and competitors, I don't see you how you can get a sense of the business.

Of course, models would be a great tool and probably far more efficient than doing a month worth of research, I don't know how you can put a probability on many of the aspects listed above. Management is not a quantifiable risk. How does it fit in a model?

Obviously, I don't have the slightest idea of how to model a business' success or failure, but I have a hard time understanding how it would work at all. I guess a $100 million industry can't be wrong , and I am probably taking the wrong approach to this whole topic. I do agree that currently actuaries don't come close to having the required skills to do this type of work, but I think that of any existing profession it would be a good place to start, and there are many within the industry that would like to have shot at it.

campbell
10-02-2004, 12:50 PM
Why isn't the SOA in there like a dirty shirt singing all of our praises? I'd love to do that type of work, although it doesn't really seem like brain surgery to me.
You look at the cash burn, and the balance sheet, and a few accounting ratios over a period of time, and you can usually get a pretty good idea of where a company is going to be in a year or two. That's why subdebt lenders use covenants, so they can get out when things start to go south.

Actually, I just came back from an SOA seminar on Asset Liability Management, and Credit Derivatives was a pretty hot topic discussed while I was there. It was the modeling aspect of default over a wide bucket of bonds - not necessarily about doing the groundwork to figure out the stat for an individual company. Because the real money and growth right now is in bundling credit default options and securitizing them just like mortgage-backed securities.

It is somewhat complicated in the modeling, and it helps if one has a =really= strong math background (more than the exam system tests) -- but one can pick that up in a masters in mathematical finance program (NYU has one). There's a lot of people out there looking into securitizing credit risk, and more interest in using this to hedge various risks. Those who are really into investing topics like derivatives and hedging strategies would do well to look into this. Lots of opportunities for quants and actuaries. John Hull (who has some texts on the current SOA syllabus) has written much about this lately, and said he was about to put up his latest preprint on modeling these instruments.

Will Durant
10-02-2004, 02:36 PM
It is more sophisticated than you portray. You need a high level of mathematics. Actuaries don't have that level in their studies. Writing investment or finance track graces the surface. You don't ask an actuary to do high level derivative pricing unless they have advanced degrees in applied mathematics. Credit derivatives falls in the same category. This highlights the fallacy of the actuarial profession.
I agree with this. It has been discussed before in the context of expensing stock options, with Michael Davlin expressing pretty much this same opinion.

http://www.actuary.ca/phpBB/viewtopic.php?t=33070

campbell
10-02-2004, 06:21 PM
I guess I should mention that I have a masters degree in math, with a heavy-duty background in probabilistic modeling. Yeah, it's tough to just jump in Ito Calculus and know what you're doing. It goes way deeper than knowing Black-Scholes.

Still, there are actuaries who do have the technical expertise for modeling. And the actuarial biz could be attracting people with heavy-duty math backgrounds.

Incredible Hulctuary
10-03-2004, 12:43 AM
The success of any given business is contingent on so many unquantifiable (and quantifiable) variables. The role of management, marketing, product demand, input prices and focus all drive the success of a business. Until you really dig and to the leg work by speaking with management, reading the footnotes in annual reports, analyzing the market and competitors, I don't see you how you can get a sense of the business.

Of course, models would be a great tool and probably far more efficient than doing a month worth of research, I don't know how you can put a probability on many of the aspects listed above. Management is not a quantifiable risk. How does it fit in a model?

Couldn't the same thing be said about life span or the probability of having a car accident? There are numerous unquantifiable risk factors in those. But those unquantifiable factors often manifest themselves in quantifiable data, just like the nebulous problem of "aggressive driving" results in a increased number of traffic tickets or accidents.

You can't quantify the quality of management, neither can you quantify a driver's level of aggression. But you can quantify events that result from the quality of management, such as employee turnover or same-store sales growth. And those quantifiable factors may have strong correlations with credit risk. Of course, if you're going to make securities out of it, you'd have to be able to pinpoint it much more precisely and involve all sorts of financial equations, which brings us into the type of PhD mathematics that they use for derivatives.

Will Durant
10-03-2004, 01:30 PM
And the actuarial biz could be attracting people with heavy-duty math backgrounds.
Aye, it could be, but it isn't. That's where the SOA should be coming in, but instead they seem to be moving in the opposite direction (from looking at the 2005 education redesign).

DW Simpson
05-13-2005, 03:04 PM
http://news.ft.com/cms/s/6eb78c92-c0f1-11d9-a3da-00000e2511c8.html

New tool helps pension fund trustees assess risk of sponsor's collapse
>By Nicholas Timmins, Public Policy Editor
>Published: May 10 2005 03:00 | Last updated: May 10 2005 03:00
>>
Standard and Poor's, the credit rating agency, has launched a tool to allow pension fund trustees to assess one of the biggest risks they face: how likely it is that the pension fund's sponsoring company will go bust.

It is a risk that most pension scheme trustees to date have paid little or no formal attention to, according to pension specialists.

But from next April, under draft guidance issued by the pensions regulator, trustees will have to make such an assessment in deciding whether they have sufficient funds in defined benefit pension schemes, most of which still have sizeable deficits.

John Ralfe, an independent pensions consultant, said yesterday: "The risk that the sponsoring company will default is the biggest single risk that most company schemes are facing."

During the 1990s, when pension schemes were apparently in surplus, he said, trustees paid no attention to that because if the sponsoring company went under there looked to be enough money to pay out benefits. "But in the last few years it has become an ever more important issue, and more important than just the size of the deficit, which can go up and down quite quickly."

Trustees for pension schemes run by the biggest companies can currently make an initial assessment based on their credit rating.

But most of the 5,000-plus smaller companies with pension schemes do not have a formal credit rating. Trustees may have a sense of how strong the sponsoring company is. But Tim Keogh of the consultancy, Mercer, said: "There is a shortage of tools that can help trustees make a formal assessment."

S&P is offering a service that it says will cost between 6,000 and 10,000 a time, which provides a credit estimate and considers the prospects for the sponsor and the industry in which it operates. It will also set out the company's exposure to the scheme, helping trustees assess how quickly a sponsor should be pressed to plug any pension deficit. The ratings agency said it would "better arm trustees in their negotiations with sponsors about funding levels".

S&P claim to be the first to offer such a service, although it is one that some of the big actuarial firms are thought to be considering.