Mark, this is a good point.  In Moody's June 1999 study, Debt Recoveries for 
Corporate Bankruptcies, there is a graph of Firm-Level Recovery Rates by Year 
of Bankruptcy Resolution (as opposed to default year).   Reading from the 
graph finds:





As the data above is based on resolution year, the same study finds the 
average length of time spent in bankruptcy is 1.30 years (median 1.15 years 
and standard deviation of 1.20 years).  The interesting point is the slide in 
recovery value 1996 to 1998.   One would wonder if the mix of high-yield 
issuance slanted towards the tech companies would yield lower recoveries 
(quick technological obsolescence--little tangible recovery value).  

Interesting to note is the December 1999 to December 2000 change in spreads.  
The table below divides the December 2000 spreads by the December 1999 
spreads.   One would have expected a pronounced flight to safety over this 
time-frame, but the figures are more mixed:









To: Michael Tribolet/Corp/Enron@ENRON, William S Bradford/HOU/ECT@ECT, David 
Gorte/HOU/ECT@ECT, Vince J Kaminski/HOU/ECT@ECT
cc:  

Subject: Re: Default rates  

Per our discussion, see attached for impact of assumed recovery rates:  



Michael Tribolet@ENRON
12/11/2000 08:09 AM
To: William S Bradford/HOU/ECT@ECT, David Gorte/HOU/ECT@ECT, Vince J 
Kaminski/HOU/ECT@ECT, Mark Ruane/HOU/ECT@ECT
cc:  
Subject: Default rates


Please see below for my note to Jeremy at the bottom and his reponse.   I 
have placed Mark Ruane's yields against a mid November default frequency 
table.   Note there may be a slight shearing in dates, but the concept is 
more important:


Market implied cumulative default rates (%):

                 1 year     5 year     10 year
AAA         0.51           5.74         14.54           
AA            0.67           6.39          16.61
A               0.98           8.98          21.03
BBB         1.17           9.88          22.39
BB             3.27        18.62          37.51
B                4.65         24.21         46.27



S&P Historical default rates (%):

                 1 year     5 year     10 year
AAA         0.00           0.13          0.67           
AA            0.01           0.33          0.90
A               0.04           0.47          1.48
BBB         0.21           1.81           3.63
BB             0.91          8.82          14.42
B                5.16         20.95         27.13


In looking at the One-Year transition rates as a very rough proxy for how 
many more defaults occur in a recession (1991) versus average (1981-1999)  
historical default rates (%):


                             Investment grade                Non-Investment 
grade
Avg. 1981-99            0.07                                             4.21
1991                            
0.12                                           10.40
Multiple                      1.7x                                            
2.5x


Looking at where the market implied default rates divided by the historicals 
default rates to obtain a "multiple" (how much more severe than historical):


                 1 year     5 year     10 year
AAA         infinite      44.2x       21.7x
AA            67.0x        19.4x       18.5x
A               24.5x         19.1x      14.2x
BBB           5.6x           5.5x        6.2x
BB             3.6x           2.1x         2.6x 
B                 1.1x           1.2x        1.7x


On the 10 year historical figures,  we need to be careful as the S&P static 
pool figures show a definite seasoning (lower defaults in late years probably 
due to prepayment) versus our contracts.  Secondly, the S&P figures have 
Withdrawn ratings, which usually mean they are stale, but loosing some 
information content.   

I will ask Emy to set up a meeting to discuss further.







---------------------- Forwarded by Michael Tribolet/Corp/Enron on 12/11/2000 
07:06 AM ---------------------------
From: Jeremy Blachman@EES on 12/10/2000 07:21 AM
To: Michael Tribolet/Corp/Enron@Enron
cc:  

Subject: Default rates

Thanks. I would STRONGLY  suggest an offsite sooner than later with a handful 
of the right people so that we can step back and design the right 
architecture for looking at credit in our deals. It is broken, not clear, 
killing our velocity and true capabilities. We also need to look at staffing, 
skills sets, the credit reserve model etc. Perhaps you should take a crack at 
an agenda.
---------------------- Forwarded by Jeremy Blachman/HOU/EES on 12/10/2000 
07:08 AM ---------------------------


Michael Tribolet@ENRON
12/09/2000 03:51 PM
To: Jeremy Blachman/HOU/EES@EES
cc:  
Subject: Default rates

I visited with Vince Kaminski for about 20 minutes today regarding the market 
implied defaults rates  and the disconnect in investment grade land.  He is 
seeing the same anomaly and agreed that we as a company need to revisit the 
methodology employed in calculating the implied figures.     I will follow 
through and report back.