he said, "i'll be gone, you'll be gone." they failed to adjust their ratings to take into account credit risk from the fraud and lax underwriting standards that increasingly characterized the mortgages securitized and sold on wall street. in august of 2006, an employee wrote, "there is been rampant fraud in the industry for quite some time as pressures mounted to feed the -- nation machine." in september of 2006, another employee wrote, "i think it is telling us that underwriting fraud, appraisal fraud, and the

general appetite for new product among some originators is resulting analysts being made that should not be made." the surveillance group told me that she is seeing losses as bad as 50%. i would like to be able to publish a commentary with this data. not taking into account mortgage fraud, it did indeed turn into a powder keg. one that helped blow up the cdo market and triggered the 2008 financial crisis. in the fall of 2007, mukasey c -- moody ceo call the town hall meeting to talk to his employees.

"what happened in 2004 and 2005 is that our competition went nuts. everything was investment grade. it really did not matter. no one cared because the machine just kept going." the managing director later responded, "our errors make us look either incompetent or like we sold our souls to the devil for revenue or a little bit of both." he said, "i would like more candor from senior management about our errors and how we will address them in the future." that is what we're calling for today as well. kantor, not only about what went

wrong, but what can be done to prevent another crate -- credit rating disaster in the future. the final -- the house and the senate financial reform bills before congress offer a number of measures to increase credit rating oversight. both bills would eliminate the statutory prohibition, the clearer language to set standards for grading brady models, methodologies, is still needed. the bills would also beef up the sec's enforcement authority. the bills should be further strengthened, in my judgment, by

directing regulators to tackle the inherent conflicts of interest that arises when rating agencies are paid by the people that the rate. our investigation provide strong support for better controls and credit -- -- credit rating agencies. one more matter comment yesterday, the subcommittee was made aware of a longer version of an e-mail that was included in the exhibit book as exhibit 23. we were not aware of the longer version earlier when that book was put together, so we have added its to the book as exhibit 23 addendum. the e-mail shows that merrill lynch was trying to make a direct link between the fees that it paid and the ratings it would receive on a deal.

the longer e-mails shows that moody's told merrill lynch that its deal and analysis was independent of any fee discussion. that was surely a welcome response and a very inappropriate request on the part of merrill lynch. senator coburn had planned on being here today and had looked forward to it. but he was called away to matters and oklahoma, so we will have to proceed without him. i do want to thank him again and his staff for their tremendous ongoing support with this investigation. i want to call on senator kaufman, who has also been a major supporter of what we're trying to do on the subcommittee. we're very grateful for him being on the subcommittee.

>> thank you, mr. chairman. since this meltdown occurred, after concern about jobs and what happens on wall street and what we're going to do about it and i have to tell you that the most scorn is the idea that you have this incredible growth in revenues at the same time their ratings thousands mortgage- backed securities aaa and they now turn out to be junk. aaa to junk is a hard thing for people to deal with. how can you this so badly? the problem we are dealing with is what happens. it really more importantly, what

are we going to do about it? both are important. aaa will never mean what it used to mean. i do not care what we do here, but it will allow me what it used to me. people are just beginning to determine the conflict of interest that the chairman pointed out, the incredible conflict of interest, and also the way that credit rating agencies have been protected and the way they do their job when you hear them talk about the fact that they really have no responsibility to the average investor out there. that is not their customer. that just does not go with me. here is the problem. i believe that the two greatest things that make this country great is democracy and our capital market.

that is to make such a great. the credibility of our capital market is in tatters. fortunately, in the rest of the world, their capital market, where you take a greater chance. that is not the perception. the idea that the average person cannot use rating agencies to determine the quality of the product means that people will stop using the product. our markets will long be credible. they have lost a lot of credibility already. if we do not deal with this, we will not go on as business as usual. i spent a lot of time in new york. it is not going to be that way. this is not going to go on forever. our market -- we lose our credibility, the united states

of america has lost one of the keys to its success. what we're doing here is incredibly important if we're going to maintain the credibility of our u.s. market. thank you, mr. chairman. >> to it, senator kauffman. let welcome the first panel of witnesses. frank raiter, forming -- former managing director at standard & poor, richard michalek, former vice president and senior credit officer for the structured derivatives products group at moody's, a former team at managing director of the structured derivatives products group, and dr. arturo, a former

moody's senior vice president and current director of the finance center at the university of santiago chile. frank raiter, i understand that you are here under subpoena. i appreciate you being here, whether a subpoena was issued or not. we look forward to your testimony. all witnesses who testified before this subcommittee are required to be sworn. at this time, i would ask all of you to stand and raise your right hand. do you swear that the testimony you're about to give will be the truth, the whole truth, and nothing but the truth so help you god?

we have a timing system here about a minute before the red light comes on, you will see the light changed from green to yellow. that will give you an opportunity to conclude your remarks your written testimony will be printed in the record in its entirety. frank raiter, i think we will have you go first. >> good morning. from 1995 until my retirement in 2005, i was the managing director at standard & poor. i think i had an inside view of the role of the rating agencies in the recent economic crisis. the failure of the major rating agencies to adequately address and assess the risks associated

with new mortgage products is the result of several factors. the first one is the lack of oversight by the sec. the second was the impact that these decisions had audit management of the rating agencies. and the disconnect between senior managers and the analytical managers responsible for assigning ratings. the final factor was the separation of the initial ratings process. the first factor, a lack of regulatory oversight, resulted from the failure of regulators to appreciate the unique position the agencies assume the financial markets. they were granted preferred status by the sec. other regulators followed suit. there was no regulatory oversight, no standards established to measure the performance or quality of the

ratings. the preferred position of the rating agencies led directly to the second factor. management of the rating agencies came to believe that the increasing revenues were the results of superior management skill and insight rather than the accommodative fed interest rates. this success bred complacency and an aversion to change. this resistance to change was the primary cause of the failure of the ratings and the ultimate -- in the ultimate financial crisis. senior management, on the other hand, was focused on revenue, profit and share price. management wanted increased revenue and profit will analyst wanted more staff, data, and i t support, which increased expenses and reduced products -- profit.

the analysts were responsible for both producing ratings and developing and maintaining ratings criteria. balancing these two missions was a significant mission in the group were revenues grew tenfold between 1995 and 2005. trading volume grew five or six fold. ratings production was increased -- achieved. adequate staffing was not the only challenge faced in trying to maintain quality of the ratings process. the accuracy of the model was also critical to the quality of the ratings. it was based on a status set of approximately two under 50,000 loans -- 250,000 loans. analysts continued to collect larger data for the next

versions of the model. in late 2002, another version of the model was introduced based on 650,000 loans. by early 2004, preliminary analysis of this more inclusive data set was completed. that analysis suggested that the model in use at the time was underestimating the risk. in spite of this research and development of the model, it was postponed due to lack of it resources. to my knowledge, this version of the model based on 2.8 million los was never implemented. the final condition that

contributed to the failure of the -- -- ratings agency was the separation of the initial ratings process. while the ratings process utilized ever improving models, surveillance operated under their own criteria. the manager of surveillance refused to use the ratings model in reviewing the performance of outstanding bonds. the concern was that we rating outstanding deals with new information would significantly increase rating of volatility and result in lost revenue. by 2005, -- it is my opinion that if they had pushed to implement the version of the model based on 2.8 million keolis in late 2004 or early 2005, the economics of deals

inc. the lowest quality of subprime would have disappeared. in addition, the riskiest transactions submitted for ratings in 2005, 2006, and 2007 would have been assigned much higher which might have made it unprofitable for lenders to make additional loans. if the surveillance department had raided these existing deals each time ratings criteria were adjusted, transactions would have been put on credit watch or downgraded in 2005. that i would have us -- that would have sent an early warning to investors. this concludes my opening comments. >> thank you very much. >> is your microphone on? >> yes. >> my name is richard michalek. i join the structured derivatives products in june

1999. my position was eliminated in december of 2007. i held the title of vice president credit officer. i performed legal analysis, and to assist in the development of refining trading practices. my regular responsibilities included participating in rating committees -- committees within the group. consulting on legal matters brought me to new york, london. i published the group's quarterly and annual review, assisted with the legal portion of the semi annual training sessions for new hires. during my last year, my primary responsibilities were split between serving as a senior analyst and being the project

leader responsible for developing a methodology for rating collateral managers. i was a securitization consultant for a law firm. it before that, i was an associate lawyer. i was admitted to the bar as a solicitor in new zealand. my testimony today is based on my experience working in the structured derivatives product group. well i have the opportunity to interact with several other groups within moody's, i do not profess any particular expertise or advanced knowledge. the testimony today is also not been delivered with the intention to bring harm to any individual or stand in judgment of individual behavior. imperfections are neither

surprising or unexpected in light of the framework presented to the rational people comprising the credit rating agencies. credit rating agencies serve an important function. in theory, the agents are independent. because of a repeat experience, they should be able to provide this measure of risk at a lower cost. my experience in the derivatives group provides that what i hope will be a prospective with respect to a couple of questions. one question that is being asked, how independent are these agencies? another question is, what consequences do rating agencies suffered under the proposed and current framework?

as for that first question, the independence of the derivatives group changed dramatically during my tenure. the willingness to declined to rate or to just say no to proposed transactions steadily diminished over time. the unwillingness to say no group in parallel with the company's share price. the apparent loss of bargaining power by the ratings agency in general and as a group in particular was coincidence with the steady drive towards the modernization -- as it produced more products, as the quality between the different rating agencies lost some of its important, the threat of losing businesses absolutely tilted the balance away from the

independent arbiter of risk toward captive facilitator of risk transfer. the second question, it is asking a handful of more fundamental questions. who should bear the risk of getting it wrong? if we except that the ratings of the rating agency products should all the ratings be issued by ratings agency, i am of the opinion that much more could have and should have been done to improve processes and procedures. i am not so naive to fail to appreciate that in the extremely competitive environment, or the message of management was not just say no, but must say yes, in the available resources had to be spent on remedial corrections. i believe that ratings provided

an important public good. i believe that some ratings deserve some measure of protection from liability and claims of negligence. to the extent that agencies remain wholly private entities, a distinction based on the extent of the public good provided might be made with respect to the products been raided. -- read it. -- rated. the liability for getting it wrong must be a sign to the entities involved. the agencies would redirect some of the extraordinary profit margins in the research and once again have incentive to just say no. that concludes my remarks. thank you.

>> thank you very much. >> thank you very much. good morning. i would like to thank the chairman and the subcommittee for holding this hearing on the role of the rating agencies in the financial crisis. during the majority of 2007, i was the managing director in charge of the business line that rated subprime cdo. i was suspended by moody's. in my opinion, because of the financial crisis lies primarily with the misaligned incentives in the financial system. individuals across the financial food chain are compensated based on quantity rather than quality. the situation was no different at the ratings agencies. the vast majority of analyst at

moody's are honest individuals who try hard to do their jobs. however, the incentives in the market for rating agency services favored and still favor short-term profit over credit quality and quantity over quality. at moody's, the source of this conflict was the quest for market share. managers of the rating groups were expected by their supervisors to maintain market share. it was an unspoken understanding that loss of market share would cause a manager to lose his or her job. management would periodically distribute e-mails indicating the department's market share. these e-mails were limited to managing directors only. even if market share dropped by a few percentage points, managers would be expected to justify the deals that were not rated. colleagues were under extreme pressure. senior management never explicitly forced the lowering

of credit standards. there was one easy way for a managing director to reduce market share. i do not believe that this was done a deliberate manner. instead, during the bubble years, this need to rationalize changes and methodology with performance was quite exceptional. easier still was avoiding why collateral standards had declined or whether some of the parties had ulterior motives. i began to receive these e-mails when i was promoted to managing director. it would list all the deals in the market for the relevant. we did for the relevant. -- for the relevnant period. i was responsible for rating cdo's. the group generated over $2

million of revenue. this amount accounted for one- fifth the total revenue of the entire rating agency. however, trouble for securitization was already growing. during the course of the year, the end of this initial phase of crisis was heralded by the fall of hedge funds. to the resulting -- it sent bankers to clean up their balance sheets bridge in september of 2007, i was told that the ratings of the subprime bonds were about to be downgraded. while the understaffed needed time to under three determine the new ratings, i left the meeting with the knowledge that the then current ratings were wrong and the lager reflected the best opinions of the rating agencies. this information was crucial to the few cdo's in my pipeline.

it the ratings were wrong, the ratings of the cdo's would be wrong also. it would constitute securities fraud. i immediately notified my manager and proposed a solution to the problem. my manager declined to do anything about the potential fraud. as a result of my intervention, a new process was announced. i believe this action saved moody's from securities fraud. just about a month later, in mid october, another e-mail was sent to the managing directors in my group. along with the e-mail, our business manager noted that our market share dropped from 90% to 94%. -- 98% to 94%.

this was the most disturbing e- mail i have ever received in my professional career. a few days before, at the time, this was the largest single down grate at moody's. as a direct result, over 570 billion would be downgraded through the end of 2008. despite the net -- the amount of errors, it appeared to me that my manager was more concerned about losing a few points in market share then about violating the law. in late october, my manager asked me to leave the group. i was given a small position with less responsibility and less pay in a different group. while moody's has the knowledge that the rating situation constituted a problem, it failed to activate a nearly identical situation in 2009.

instead of following some common-sense steps to prevent a violation of the law, management chose to suspend me after i pointed out the breach. recent ratings activities indicate that market participants still prefer the most aggressive rating. rating firms that have taken conservative positions have seen their market shares tumble. we will no doubt see the results of this when the regulatory spot light is turned off. credit standards will once again plunged as rating agencies race to build a market share. the only way to prevent this is to recognize that the function they perform is a quality regulatory one. much like accountants. -- quasi-regulatory one. thank you very much. >> thank you very much.

>> thank you, mr. chairman. thank you, senator kauffman for the invitation to be here. i am a professor at the university of chile. i spent the last 15 years working in finance. i worked at moody's. i worked there from 1996 until 1999. when aaa meant something. i testified two years ago and i made some observations and some suggestions that i thought could be implemented. i am not going to read anything. i want to make three points.

the first observation that i would like to make is the following. they give ratings and that is basically nothing but an opinion. moody's gives ratings based on expected loss. s&p gives ratings based on default. they both profess to give ratings based on different benchmarks, different standards. if you take a look at the ratings, there is a high degree of differences in ratings. that is something that raises

some issues and should be investigated in more detail. you become very suspicious when you see such a degree of difference between the ratings agencies. are these two companies dancing independently? is this carefully calibrated to make sure the ratings are aligned with market share? that is something that is concerning. the other issue is something that' is linked to the subprime market. the ratings agencies have said that the ratings they gave to transactions involving subprime loans suffered so badly because

they relied on data provided by the bankers and it was no good. allegedly, they were not at fault because the use information and it turned out to be wrong. i do not believe that is a reasonable explanation. that was not the case when i worked at moody's. we always checked everything that the bankers told us. we look for a certain amount of detail. if that were the case, it seems to me as a market participant, if the ratings were given information that you did not verify, and they should come up with a warning. something along the lines where ratings are based on information that they believe this is true, but they have not checked. something like that. finally, this is a very serious

problem, the situation where we are right now, what we -- i believe the situation is really bad because what we have right now is a marriage made in hell. you have a situation where nobody believes in and the ratings. -- nobody believes in the ratings. i cannot tell you how critical that is. the securitization market is paralyzed. that has significant effects on the market. it is the perception that this market is not regulated, this is

probably the most regulated market in the world. i think i am going to stop here. i would be happy to answer any questions that you might have. thank you very much. >> i think we will have a 20 minute round. if you would take a look at the exhibit 94b, this is a cdo note has a vertical abs 2007-1. an analyst complained about how the verticals issue were was not cooperating with them. the deal was unlikely to

perform. in a 2007 e-mail, the one that you are looking at there, exhibit 94-d, one analyst said that the vertical is closely tied to the na -- b &a. i do not see why we have to tolerate lack of cooperation. the deal is likely not to perform. despite bad judgment, if it is unlikely to perform, s&p rated it. several months after that deal was raided, the lungs began to show delinquency. a little later on, it was downgraded. the securities are now below

investment grade. there in junk status. one of the purchasers of the vertical securities is a hedge fund cold pursuit. -- called pursuit. the suied because of the lack of ability to sue the rating agencies. the court ordered ubs to set aside some funds to pay a possible all ward to that investor. the investor had also uncovered an internal e-mail at ubs what baker wrote. "sold some more crap to pursuit." that is exhibit 94-n, by the way. it is it possible for an analyst

to reach a deal even though the analyst thought that the deal was "not likely to perform." >> i have to say that i was not -- i really do not know how they did things down there on the residential side during the time i was involved. it was not unusual to turn down deals if we did not think they would perform or if they did not meet our criteria. things got dramatically more hay wire after 2004 and 2005. i honestly could not tell you what they did in cdo's. we would not -- there were occasions where we had some real

concerns about collateral and we would put such high levels of the transactions, it was not economic for them to do it. we would not write it. in the late 1990's, prudential mortgage changed their waterfall structure, which did not meet our criteria. for a number of years, we would not reach any of their deals because they did not meet the standards. >> you have years of experience eighth. should a cdo -- take a look at 94-e.

94-e it shows some of the assets that were included in the cdo had already been downgraded. d.c. that? >> yes, i do. -- do you see that? >> yes, i do. >> it had already been downgraded at the time they were recruited. they were not performing as expected. if it downgraded assets is included in th a cdo, is that something that an analyst with notes? >> we have rules that if an asset was on what, it would be taken down a few notches. once an asset was downgraded, it would be used at that ratings

and no more, no less. >> would that be warning signs of the quality of the assets? >> it would have been a qualitative warning. quantitatively, it would not leave an analyst much room. at that point, we would have said, this is our best view. >> should it affect the credit analysis? >> absolutely. this is something that i tried to do later. once you take the path to downgrade, that is not likely to stop. we tried to implement something where the securities that had been downgraded already, but not on what. -- on watch. >> should it affect the analysis? it did not affect the analysis here. the s&p analyst said in his e-

mail that bank of america was using it to take risk off of its books. vertical is a company that was partly owned by bank of america and was brought by several former bank of america employees. if an analyst thinks that what is going on wednesdawith a cdo,d affect the ratings process? >> i think so. >> how long is expected that a triple a rating should hold? >> the waterfall typically pays the higher rated bonds offer. as they pay down, the bonds that are underneath should last the life of the transaction.

>> would you all agree with that? >> in this case, these ratings were downgraded within a year. they're now below investment- grade. these are junk. >> if i could respond, i think that we are touching on something that you are likely to hear later on in the day. common perception of what a triple a rating is and means is not necessarily what the definition of a triple a is. it is something of a legal distinction. it is very important that this is not lost on the committee. there is the migration rates in the history of the different ratings that are assigned. those migration rates represent

an average migration for a particular rating. it is simply a population of which there are tales on the there is. some aaa's never get downgraded. some are down greeted quite quickly. there is an expectation in the market and there is a proper expectation that aaa i's are not going to be issued on monday and on friday, downgraded to anything else. there should be a necessary element of stability in the rating, at least at moody's. it was one that was ongoing. >> thank you. let's take a another look at another failed rating. this time involving mortgages issued by fremont. take a look at exhibit 93-b. in january of the 2007, s&p was asked to rate subprime loans issued by fremont investment.

it -- they were known for poor quality loans. at that time, an analyst sent an e-mail to a supervisor saying the following: "i have a goldman deal. since they have been performing not so good, is there anything special that i should be aware of it? " one of the supervisor responded, "we do not treat their collateral any differently." which got s&p employees that no there is a problem, but treated those lungs like any other. in 93-d, there is an e-mail which s&p analyst for circulating an article about

fremont and how they stopped using 8000 brokers because of lowland's with high delinquency -- with some of the highest delinquency rates in the industry. in march, a couple months later, fremont announced an 8k filing that the court of appeals found significant -- sufficient evidence that the company was " marketing and extending an adjustable rate mortgage products to subprime are worse in an unsafe and unsafe -- and sound manner." -- and unsound manner." the suit could proceed against the company. just a few days later, they

entered into a publicly available cease and desist order regarding fraud and lax underwriting standards. despite that information, they were rated by both s&p and moody's february and march -- before the ratings were done, they knew of those facts which i just described. by the end of the year, in 2007, both companies began downgrading those securities. my question -- does either s&p or moody's take into account and issuers reputation for issuing either good loans or bad ones and incorporate that into their credit analysis? frank raiter? >> yes, they do. at the time this occurred, the policy had been that when

information is provided to the analytical staff by investors or other originators, they would look into the matter. if it was justified, it would have resulted in a visit to them and a review of their practices and procedures proved it could also have resulted in their deals being put on creditwatch are being held up for ratings until the information could be worked out. at this time, in fairness to what was going on, there were rumors rampant about the quality of appraisals and the quality of all underwriting standards that could have been quite overwhelming to the staff to try and track them down. it routinely should have been investigated. >> when the analyst said no special measures, that was not what was supposed to happen? >> it was the manager's responsibility to say, this is a problem. can you get to the bottom of it? they took their marching orders

from the managers that were in control of the criteria. >> again, when they said, no special measures, there should of been special measures? >> yes. >> i do not know if i feel competent to comment too much on how the team would have responded to this particular allegations. i know that with respect to collateral managers, if there was an issue that was raised regarding reliability of origination that a collateral manager was involved with, we might explore it. we would definitely send a team to the collateral manager to do review period to about a week those procedures and the information that was contained in those reviews would be used analysis. as to the underlying collateral, i am not competent to comment. >> should there have been measures taken in the circumstances? >> i think so.

i am not confident in what occurred in the group, but have this information come into the cdo side, we would have looked into it. i was not involved in 2007 as this information went through with the folks who raided the prime directly. but there was almost a feeling when dealing with them that there was a see-no-evil and hear-no-evil kind of attitude. they did not want to believe what was going on. they were part of this market and it just should not be happening. >> part of the culture. >> in many occasions, when we

were presented with transactions that did not make sense or made us nervous, the response was either no, you cannot do it for the rating is going to be way below. we do take into consideration, when analyzing the transaction, it may be applicable to that transaction. that was the way it was done in the cdo group. >> when you were there, that was how it was handled? >>. i remember several -- yes. i remember several situations where we had to make conservative estimates. >> that is the way it should have been done? >> i believe so. at moody's, are rated the first emerging-market cdo in 1996.

we re did about 10 to 15 deals during that time. one of the problems was that we did not have enough information. the way we handle that, we made conservative assumptions regarding certain pieces of information that we did not have. at the risk of sounding too technical, you better get a technical thing right. we did something called stability analysis. all of those emerging market deals went very, very well. >> take a look at 93-8. this says volumes about fremont. this is a moody's e-mail.

december 2006, they were talking about delinquency rates. if you look at page to comment here is the chart of the top-10 issuers that have high delinquency rates. on the bottom of of the next e- mail, it says, "is this data correct? " more evidence about fremont, by the way, as bad as it gets. they had those articles about the 8000 brokers and the other information that i mentioned. in terms of the deals, i want to talk about an e-mail -- 95-a. this was an e-mail in which an investment banker described a deal. this was rated aaa by both s&p

and moody's. it describes it as having a lot of cushion. they needed a lot of cushion to protect against losses. despite that cushion, it was downgraded six months after being rated rate that may not be monday to friday, but it is not the way it is supposed to be. look at 95-b. this is an s&p e-mail chain from august of 2007. regarding -- "it appears that the closing date porfolio they gave us were not the same. it appears that 25 assets that included in our closing date porfolio were dummies.

they were placed in less than 24 hours with assets that would have been notched and made the porfolio worse." the cdo had given s&p down the assets that they used to come up with its ratings. at the last minute, they replaced those dummies with new assets, less than 24 hours before the closing. the new assets, they made the porfolio worse. if that had been acted on, the rating would have been worse. is that the correct reading of that e-mail? >i need you to put your microphone on. >> it suggests that they pulled

-- they put dummy deals in the original run and then they substituted them with loans or credits that did not meet the standards of view the original loan. -- standards of the original loan. it looks like a bait and switch. it did not meet the criteria. i do not know who addressed it. >> i will tell you, these e- mails are just devastating to the kind of culture that was going on here. it is incredible that you could have this type of dummy assets being put into a cdo and then substituted at the last minute. >> i would like to point out that there were a lot of good analyst that writing is in memos and asking for guidance but the guidance was not forthcoming from the top. >> i cannot agree with you more. that is the culture.

i am going to turn this over to senator kauffman. >> thank you, mr. chairman. you attributed this to bad incentives. quality was gone. competition. the sec oversight. how much of this was outright fraud it? i understand this about the business. the things that are going on here, because of the incentives, because of the lack of oversight, people were doing stuff that they knew was absolutely wrong. >> it is clear from a lot of these e-mails that people were making a very poor calls in terms of the analytics. whether that is fraud or not, it was wrong, it did not look good.

but there was no one over there watching over them to tell them, this is not the right thing to do. >> i understand that and i am very sympathetic to that. i spent a lot of time in business. i worked for corporations. there were always incentives of some kind to make the quarterly earnings, push this stuff out the door. when you have the incredible amount of charge that he put out there on the number of r &b that are rated aaa and turned to john, many after you were gone, it sounds like -- maybe fraud is too strong a word. but these are not common business practices. no matter what the incentives are. >> they worked with business practices there was a huge disconnect between management. senior management thought they

were responsible for all these wonderful things. the analyst or just the soldiers in the trench is doing what they were told. . .

>> and they come back and say but revenues will go down, you are faced with that choice. either you continue to work there and fight, or you quit. >> i am very sensitive to this, and i understand this. i have been there. but how much do you think the management really knew what was going on down in the trenches because of the precious they were putting on people down in trenches? do you think they were oblivious to this? >> you have kathleen corbutt on

the third panel. she could possibly shed some light. the process was a mystery. we never sat down and were explained why we didn't get things, why they wouldn't make the changes recommended. they just didn't respond. they didn't communicate from the top down. >> i would have to say i can tell you at moody's while i was there it was frequently the case the information did travel upward. i was involved in one transaction where i was asked to consult with somebody in public finance. they had a ratings problem that was waiting to litigation. general counsel for moody's was at the table. we were on conference calls together, talking to the client. we were trying to resolve this. brian clarkson at the time was on the call. and this was for a relatively

small matter. it was probably $50,000. it was not as long as one of our c.d.o.s. from that alone i would infer that the information regarding some of the much larger conflicts and potential problems was definitely reaching upwards. >> thank you. first, before i answer, chris, i want to quickly come to the defense of the analysts mentioned in exhibit 95-a. that analyst mentioned was extremely bright, and the reason she was not on that deal was she was asked a lot of good questions. i wanted to come to her defense. >> no problem. >> second, i don't believe, in the cause of the crisis, there were a lot of ins tanses of

outright fraud, legal fraud. there may have been some on the front ent with the mortgage brokers and filing applications that were clearly fraudulent. but the way the system worked, you had a chain. it was almost like a game of telephone where you pass some information down the line, and everybody changes it just a little bit. not enough to jump over the fraud, but because of the length of the chain from the originator to the aggregator to the c.d.o., everybody changes it a little bit, and by the time you get to the end of the line, it was different. that is why you haven't seen cases of outright fraud. everybody was pushing the envelope. because everybody was pushing envelope, the end product was garbage. >> i can say -- see that one day, two days, five days, two

months, three months. these are very smart people. when you look at the middle of a chain and see what is happening, at some point you say what is going on here? maybe it is non-fees annuals, maybe not malfeasance. >> maybe since of omission. there was no incentive to look under the rocks. if the rocks looked polished and nice, ok, we will pass it on. >> this is the third hearing we have had, and this is kind of a common theme. no one knew what was happening. it has rings to me of other cases that nonknew what was happening. these are not dumb people. they are coming to work, coming to their cubicle, not reading

the "wall street journal," or baron's and seeing what is happening. and as mr. raider said, they are good people, and they are stuck. i don't think it was the good people, like the analyst you mentioned, that are the problem. i think this is a systemic problem here. we talk about it in surgical terms, like incentives, regulation and all that stuff, which are all absolutely totally on the mark. it was the incentives, and it was the regulatory environment, and it was a flawed business strategy, and it was competition, and it was quality over quantity. when you put all these things down, somebody in the middle of this thing would have to be totally and completely -- they had to put together the big picture. somebody had to see look, we are putting incentives on here, we have this quantity problem, the competition problem, we are

not having much regulation oversight, there is the potential here for something bad to happen. remember think you bring up a good point, senator, and you refer to your experience in the private sector. the thing to keep in mind here is this is a very peculiar sector in the private market. >> yes. >> to be perfectly blunt, if you are in private business, and you do a lot of things wrong, eventually you go out of business. in the rating business, because of some really vlade regulatory framework, there is no penalty for giving bad ratings. >> having some conversation two years after my experience, and they keep issuing ratings, and they keep collecting fees for that. one of the problems to me is the regulatory framework. there was a law passed by

congress. if you want to start a new rating agency, you have to show you have been in operation for three years giving ratings and collecting fees. now who is going to pay you to give you ratings that don't have any validity? >> yes. >> in my opinion, that is something that needs to be changed. one of the problems right now is there are no penalties for giving out ratings that are unreliable. they were a free gift, and you cannot blame them. they were taking advantage of an opportunity to make money. >> i think you have the other key point in this whole discussion, especially on the hearing today, and that is you want to make sure that the committee understands what triple-a really means. >> correct. >> and i think there is such an

incredible disparity between what of read triple-a to be, and what triple-a means -- >> to the rating agencies, and the sclour documents, senator levin used the term safe, and i don't think safe -- it is definitely the commonly understood adjective that you would use, but to be perfectly honest -- when this debate first came up, there was a transaction i was working on with an analyst that because of the pick assets there were, and it was not yet placed on negative credit watch, so quantity tatively -- quantitatively we couldn't take the hair cut that was coming. there was going to be a likelihood that this would not

hold its rating for a very long period of time. however, on the closing date, it did meet the published criteria and reached that quantitative number, and we couldn't say it is a different triple-a. >> i am very sensitive to that. it's hard to give ratings to agencies. it's hard to figure out what is going to happen. in cases like that, if what transpired after that hadn't happened, i would be sitting here as the most sympathetic man in the room to what you are saying. the question is, at some point it went from that, which was a tough decision -- and if you did not to what the rating agencies say, they think it is still like that. we have these very difficult questions we have to deal with. triple-a doesn't mean it is going to be here forever, and it doesn't mean this or that.

triple-a doesn't mean thousands of securities that are rated triple-a that two years later are junk. we are not having the first discussion, the discussion i had in business school about rating agencies and what they are. but when you are faced with a situation where because of regulation, that is that. so what is it that rating agencies are? triple-a doesn't mean triple-a. it doesn't mean triple-a based on your very good definition of what it used to be. so how do you deal with the fact that you had this systemic -- is it fair to say systemic problem? >> i think so. >> and how do you get at that so that that specific problem -- and we are going to have all kind of new problems, we do it around here all the time. but how do you deal with that

systemic problem? >> in my opinion, i think that we would really have to begin with what we are disclosing. i personally believe that there are products that deserve a commonly under stood rating, that the public can say this is safe because the rating is saying that it is safe. i think for a large number of the highly complex structured products, it's a different ball game. i think that to the extent that you are able to distinguish between those products that are clearly in the different ball game, then the caveat of buyer beware is more appropriately applied. but for that portion of the products that i think -- the enormous public good that comes from having an independent ash

tar of risk -- arbiter is something that we should seek to preserve and eliminate that bleed, if you will, from the extreme debate or debatable conversation that goes on with respect to the highly complex products, and what really should be beyond debate with respect to what is safe and what is definitely contributing to the public good. >> and do you think you could do that? i mean realizing there would be some securities in the middle, but having the idea that for what is commonly known as triple-a corporate bond, you have one thing for credit default swalms, you have something different? >> i think it would require an exe tuesday -- exe tuesday --

impetus from outside to install that kind of change, or to be the first one to install that kind of change. i had the somewhat naive idea when i joined moody's that it was a particular quality that moody's was offering, and that was something that the company was going to seek to defend over time. and that effectively our brand meant something, and that i expected people to step in much earlier to say we are diluting our brand. if we had this 96% or 98% market share, once our brand becomes equivalent to the other three, mathematically we are not going to stay there. >> you have just defined this whole thing. in my opinion, everybody just decided to go for the fastest possible money they could ever possibly make, not worry about

the brand or the whole thing. the big thing that was missing was the referees on the field. we just pulled the referees off the field not just in the rating agency but with everybody else and said we don't need referees anymore. i think just like in football, here or anywhere else, like we need police on the beat. not because people are crooks. but we need police on the beat that people will become crooks because the temptations are there. who do you think we deal with this situation down the road in terms of rating? >> well, i think you have to have some rules. actually, you have to have some penalties for not doing the right thing. and there are none, and there are no measures of whether you are doing the right thing. >> what we are here-to-is define what is not doing the right thing, offense? >> if you have developed a model in-house that is much

better than anything you are running, and it shows that you have been too optimistic with the ratings you have assigned, and you do not immediately start to use it and go back and rerate the old deals so you can warn the investors that we have been wrong, then that is not doing the right thing. and i will point out from a cultural perspective, there were two mantras we heard at standard and poor's all the time before i joined, and i am sure they went on before that. one is a triple-a is a triple-a, and it doesn't matter if it was a corporate or a structured product. they used the transition studies to prove that. they said look, what is the probability it might be downgraded or going from triple -a to default as in central? there had never been a triple-a

mortgage back that has failed until this debacle. so triple-a's were all the same. the other thing that was heard constantly, and it was in one of these e-mails, if we change, everybody will think that we have been wrong. and that just put a real anchor on any new ideas quickly going through the process because they were afraid somebody would suggest that they hadn't been right before, and they would have liability, or they would lose supermarket share. that was not doing the right thing, and they don't have a referee or anyone to tell them when they have crossed that line. >> well, the other thing is, look -- there isn't a single thing raised here in terms of what the behavior was. as an elected official, you had to change your position on anything because that is admitting you were wrong. this is a common thing. the fact that there are no

penalties is key, and i would like to key at least a few questions in the second round about grandfathering and exactly what you said, why there wasn't more grandfathering with more of the witnesses. with that, i yield to the chairman. >> there is always going to be a debate on how to cure a system, but we know there are a lot of things that should not have happened but that did happen. we are going to debate the cures and remedies legislatively in the week coming up. but i want to go back. some of the things it is obvious to me were wrong at the time. as complex as some of the remedies are, some of these issues are not complex at all. market share should not be driving ratings. would you all agree with that? let's take a look at what drove

the ratings here. let's look at more evidence. exhibit 5, an e-mail dated march 23, 2005 between s&p employees. here is exint 5. when we first reviewed results a year ago, we saw the subprime and other numbers going up, and that was a major point of contention which led to all the model tweaking we have done since. version 6.0 could have been released months ago, and resources assigned elsewhere if we didn't have to massage the subprime and other numbers to preserve market share. should those numbers be preserved to -- i mean be massaged to preserve market share? does anyone believe that? mr. raiter? >> no, sir, they should not have been massaged. i think i stated earlier that as the models were developed by our consult apt, and they were tested in-house to verify that

they were accurate and their predictiveness was improved, the models were immediately put in force. we ran out of funding in 2005 to put this version 6.0 model in place, but the preliminary analysis, as was suggested, was that we were not adequately rating the transactions. that model was delivered, i believe, in seven of 2006. they did an accuracy evaluation. it was determined to be accurate, better than what they were running, and the consultant was paid. but they also performed what was called an impact analysis on the ratings. we had never done that before, so i do not know where the order came to start doing impact analysis on the effective new models on market share, but it is apparent that that is what happened. >> what was your reaction when

you read that e-mail? what was your reaction when you read that e-mail? >> i was pretty amazed. frank was one of the p.h.d.'s who worked on these models. he is one of the best annualests we have and very outspoken. >> were you bothered by it? >> certainly. >> should market share be preserved, massaging numbers to preserve market share? >> no. >> troubling when you see that kind of an e-mail? >> it's troubling in the sense that it is one more piece of evidence of what i was observing while i was at moody's. one of the comments that might be somewhat illustrative of this is when i had some discussions with brian clarkson about the process. his perspective was yes, we could effectively produce perfect ratings, but we wouldn't be on the deals, and

if we are not on the deals, we wouldn't be able to add any value whatsoever. so yes, we take a little bit this poison, but we are going to save the patient because you have the opportunity to get in there and fight the good fight. >> and make profit? >> he didn't mention that. >> he may not have mentioned it -- >> it was definitely the part of it. it was in the one text of a discussion where my job was on the line. that had already been said earlier in the conversation. if you are difficult in the transactions, there is no choice but to replace you. >> boy, that is pretty devastating, i will tell you. what is your reaction to this kind of an e-mail here? >> it is very disturbing, and as the folks on this panel said, is something we witnessed. market share did drive the credit analysis, and i think

that is why i was also let go from the rating agency. >> because you objected to it being the driver? >> i objected to it being the driver. i went ahead and tried to prevent us from what i believed was committing securities fraud. >> doctor, what surreaction when you see an e-mail like this? >> this isn't a lot to add after what my colleagues have said, but it is a bit troubling. >> take a look, if you would, at exhibit 24-a. in october of 2007, moody's had downgraded hundreds of hmvs securities and was in the process of downgrading billions of dollars of c.d.o.s.

it was written, quote, take a look at the deals you didn't rate for the spread sheet that we sent out last night to double check the information. let me know of any stories. and an earlier e-mail said market share by keel count dropped to 94%. it was lower than the 98% in prior quarters. is this something you got frequently, this kind of reference to market share up or down as being a driver? >> yes, sir. these e-mails were sent out. the market share e-mails were sent out quarterly and sometimes on a monthly basis. they were sent out to the managing directors in a given group. >> for a firm that is supposed to have a reputation of high quality right in the middle of a financial crisis, to be looking at the market share issue instead of whether the ratings are decent and what

happened, how could our ratings have been so wrong and how can we improve it? what is on their mind, market share, market share, market share. let me ask you, add moody's, did employees understand that the amount of market share that was maintained by the groups influence the size of the employee year-end bonus? >> it certainly impacted revenues and stocks, and employees owned options or restricted stock in the company. and profitablity influenced the size of the bonus, yes. >> more pointed than that. the revision in the compensation structure -- i am not going to remember the date, i think it was in 2006 -- where a larger percentage of our compensation was going to be delivered in terms of deferred compensation. so it become more important to see that what we were looking

at was whether or not we were reaching our revenue numbers on a quarterly and annual basis which would allow us to max out our bonuses. >> and that meant that the ratings that you would give or not give to the banks would affect your bonuses, or could affect your bonus? >> could affect your bonus. if for any reason you were stopping a day, or delaying a deal, or creating an issue with the relationship between the banker and moody's, that was a problem. >> and this is the fundamental conflict of interest that we need to do something about in the legislature. would you agree? >> yes, sir. >> would you agree? >> absolutely. >> mr. raider, would you a-- agree with that sflr yes. >> and dr. cifuentes? >> yes. >> this is something debateded. we will find that out monday

night. this is an issue not in the bill yet, and it has to be somehow or other put in that bill. there has got to be a way that the regulators are going to find to eliminate to eliminate this conflict of interest. it is shocking, that whether or not something is rated triple-o or whatever, and that means something to people. it may mean too much to some people, by the way. it may mean more than technically it should mean. but it means a great deal, and legally means probably too much in terms of where some entities are allowed to invest or not invest. but that fact of life, it should not be didn't upon -- dependent on -- the credit rating or bonus should not be determined on the rating. there is such a conflict of interest to me. your testimony is going to be

helpful in hopefully getting rid of that conflict, and in whatever the new regime is, to end that conflict of interest. it goes right to the heart of the rating, which is supposed to be an honest, objective, independent assessment of the likelihood of an investment paying off, and it is not performing that function when you have this kind of pressure on people to rate a certain way in terms of their own pay. now we had a situation -- mr. raider, when you were head of rmbs, whud a model to rate the rmbs securities, and there was a period when the s&p was doing well in terms of revenue, and you asked senior management to buy mortgage data on different types of mortgages so you could improve that model. is that correct? >> yes, sir. >> and did you get the money? >> no, sir, not while i was there.

>> so this was supposed to keep models curent and to do surveillance, stharkts? >> predominantly to build models, but to maintain surveillance at the low level detail. >> do you know why the s&p did not spend the money or better analytics? >> no, sir, i do not. >> now, there are also some things that shouldn't happen regardless of the complexity of how you design a better system. there are some things, it seems to me, that are clearly wrong that happened and should not happen. we had in the subprime loan deals -- you had a number of loans in which borrowers paid a low initial rate, sometimes interest-only payments. and then after a specified

number of months or years, switched to a higher floating rate that was often linked to an index. did you have any data at the time as to how those subprime loans would perform? mr. raider, did you have data? >> the moderate theal is researched in a number of these exhibits we have looked at was a data set that had a significant amount of information on those hybrid pay option type of loans, and the analysis of those loans suggested that we were underenhancing or being overly optimistic, and that was the primary reason for prying to push that model into production in 2004 or 2005. >> because you were trying to test as to whether or not that kind of a product would increase the risk of non-payment, is that correct? >> well, the product started to appear in 2003, but in very small numbers.

by 2004 when we welt this database, we had more significant information on those types of products that indicated that how we thought they were going to behave in the initial versions, it was behaving worse than that, and we needed to get the new model in place because it had more data and gave us a better look at how these things might perform. >> was there a delay in putting that new model in place? >> well, we had some preliminary results in early 2004. i left in april of 2005, and i believe the model was delivered in september of 2006, and i don't know if it was ever implemented. >> some of the subprime loans used stated income in which the borrower accepted the stated income level. does that make loans riskier to have unfair vide income in

these applications? >> yes, sir. >> would you all agree with that, that it would be riskier with no verification of income? >> yes. >> doctor? >> if you know it is riskier, you do your annuals with much more conservative assumptions. that to me is really the bottom line. >> and mr. raider, going back to you, you had insufficient data to predict how a new loan would perform. was there a sign called a magic number? >> when we couldn't get the data in order to do a full analysis on a model, they would come up with a multiplayer that -- a multiplier used to beef those numbers up. it was assumed they would be replaced with full-blown

analytics when the data came in. it was my understanding there were some magic numbers installed in 2005 in the existing model, and if they were massaging information in the 6.0 model when it came out, it would typically be in the form of these modifiers they used. >> was there a fee for surveillance? >> yes, sir. surveillance was a profit center. >> and there was programs a large fee that was supposed to last for the life of the security? but surveillance fees smaller? >> yes, very much smaller. >> but they were supposed to pay for on going ratings and re-ratings, were they not? >> right, the ongoing review of the rating. >> let me ask you. if it becomes materially more conservative, how important is it that the credit rating agency use the new assumptions

or criteria to retest or re-evaluate securities that are under surveillance? >> it is very important for two reasons. if you don't do that, you are basically creating two classes of securities, a low class and an upper class, and that creates a discrepancy in the market, and at the same time you are not being fair because you are giving an inflated rating or you are not communicating to the market that the ratings before were of a different class. so i think the right thing to do is to analyze or actually reanalyze all of the trance as with the new parameters. and there is no revenue in doing that, as you probably suspect. >> was there any discussion about using ratings models to conduct survalue ands? >> yes, there were. >> were those discussions at times? >> i'm sorry? >> were there heated

discussions? were there disagreements over there? >> yes, there were disagreements. >> what was the argument about? >> there was a certain number of analysts on our side that said it would make a lot of sense to protect the investors, looking at the new criteria and the marks in the properties, whether it went down or up. and there was the other side of the argument that it would increase ratings volatility, which may make us look bad in the eyes of the investor and cost market share. >> so market share was a factor in whether or not you would use the new information to rerate the securities? >> yes, sir. >> was there a restricted list for rating andists at moody's that were prohibited from

working with certain banks? >> i don't know if there was such a list. >> were you on a restricted list where you couldn't work for goldman or credit swees? >> there were quite a number of banks that requested i not be assigned to their transactions. >> they were complaining? >> there was a variety of complaints, that i would be two aggressive, abrasive, or that i was asking for things that weren't being asked for by other analystings in the transaction. >> was that in your mind because, at least in some of the cases, you were asking too many questions, which would negatively affect the rating? >> absolutely. it was a case -- i attempted to provide the same anls to every transaction in my view. and we were not facing the same

set of bankers on the other side of the phone. >> a couple of questions quickly about the synthetic c.d.o.s, and there was a huge increase in those. i think everyone knows the large numbers that were being rated. did those synthetics cause any problems for credit rating agencies, anybody there? >> yes, sir. the synthetics, the key element of synthetics was their complexity as well as flexibility. >> were they being used to short the market a lot? >> from news reports, i understand, yes. >> and if they were being used to short the market, would that be saying something about the quality of the reference to the assets? logically would it be saying something about the assets being referenced? >> i think if you hear all the stories of folks shorting, not just a few securities, but in

size and massively, i think they obviously had a view that these securities were not that good and the whole market was going to clams. >> is it correct that you supervised the staff at moody's that rated a transaction known as epicus? >> that is correct. that was under my business live. i think you are referring to the ac-1. >> that is correct. >> that was under me. i staffed the transaction. i was not involved in the day-to-day. >> were you aware that the bank that presented the deal to moody's was goldman-sachs? >> i was. >> and have you seen reports that the paulson firm shorted the transaction using goldman-sachs as its agent? >> i have seen the reports in the s.e.c. kment. >> have you also seen reports that paulson played a role in selecting assets, and for the c.d.o., he expected it to perform poorly? >> i have. >> and were you or your staff

aware at the time that moody's, that was working on the abicsa rating, that paulson was shorting the assets and playing a role in selecting assets expected to perform poorly? >> i did not know, and i suspect and fairly sure that my staff did not know either. >> are these facts you or your staff would have wanted to know before rating abacus? >> from my perspective, it is something i would have wanted to know because it is more of a qualitative, not a quantitative assessment, someone who tends the deal is picking the portfolio. >> are people that are usually putting deals together want the deals to succeed. isn't that the assumption? >> that is the basic assumption, yes.

>> if a people wanting the deal to blow up is picking the assets, that would run counter to the assumption? >> it changes the dynamics of the structure. if the person putting it together wants it to blow up. >> well, i couldn't agree with you more. senator kaufman? >> thank you. i would like to touch on a few things here. grandfathering. what are the factors in deciding whether you grandfather or not? mr. raider? >> could you define the term grandfather? >> going back and looking at securities and things you have already rated and deciding whether you are going to apply grandfather to that? >> well, if you don't grandfather, they do get surveiled. there are different levels of criteria you can apply in looking at how a deal is performing. one way is to look at the pool level and decide we are just going to track the delinquencies, the foreclosures and the losses.

another way is to use the loan level on the trance as. if you can get updated loan level files, then you can rerate the transaction on today's economics, today's house prices and changes in the credit quality of the borrowers and get a better look at how it would perform today if you were going through the process. those two different ways to surveil can produce dramatically different results. >> after you changed the way -- the model you were using, zg to go back to use the newer model for older loans -- >> again, this is outside of my expertise. >> all right. >> in the c.d.o. world, a lot of times deals that were out there enclosed ended up being grandfathered. some of the models that the deals used, for example, a

divertities model, or models provided by moody's' were baked into the deal, and they had an affect in terms of the deal's compliance, meaning that there is a direct line that you have to comply with this. there was no way for us to say you have to stop using that old model and use the new model. we could have applied the new model on the portfolio. we had the portfolio, but that stuff was not taken. usually we let deals had who used the old models continue to use those models. >> i left moody's in 1989. we didn't really have that issue at that time. the market was really very small compared to what it is today. >> mr. raider, did you have any indication that stated income loans were being used in any of

the instruments that you were dealing with? >> yes. stated income loans were there. they were known as liar loans, ninas. when they started using the stated loan concept in the late 1990's, it was applied to the highest credit borrowers. doctors, lawyers, and self-employed people. as they started developing in the subprime arena, again you started out with the top of the subprime market with the initial loans that were coming into the bonds. by 2004 and 2005, with the new hybrids and the starting numbers, you were stepping down to much lower fico scores and scores of the borrower, and there was evidence starting to bubble up that brokers were impacting the way stated income was put on the various applications, that there were questions about appraisals, whether they were accurate or not. so when they first started out with the no income low-doc kind

of loan, we did have modeled in the ratings process a higher enhancement for those loans. and as we tried to collect data on the new products they were developing and how they performed or were expected to perform, we were factoring that into the models. again, i hate to beat the dead horse, we had a 2.8 loan set that was used to build the 6.0 model, and at that time we had the most information. the next data set we had had almost 10 million loans in it and was more powerful. so we were looking forward to get additional information to make a better judgment as to how things were performing without waiting for the portfolios to start going bad. >> let me ask this question. you kind of imply -- do you think the decision not to move with the more advanced models

was a financial decision, or do you think it was a decision made with the fact that it was going to make things more difficult to give higher ratings, and therefore be not as competitive. >> i think the initial decisions not to fund it were because of resource constraints and the desire to maintain higher profits. i think the decisions that were made when it was finally developed and available for implementation would indicate whether they were starting to take a more serious look on what the impact of market and reliability was on analytics. i was gone by there. >> i wanted to bring to the attention of the committee is what you are in part drawing out is an extreme reliance on modeling and on the quantitative analysis going on. right. >> to the extend you had a stated loan, it was another data input for which you could

make some asudges little. so there was necessarily a stepping back on some level in my mind of the qualitative analysis of what was going on down below, and instead you were saying can we model it? yes. do we need to adjust our assumptions? programs. the debate would then be around is this the correct adjustment. >> at some point, as mr. raider said, it starts out being a very small problem at the very top for people with a high income, and now it is getting wider and wider use. i can understand, modeling works great because those stated income loans, when they first started, had very few defaults. but it was clear to everyone that, early on -- mr. raider, when you were still there, stated income loans were becoming a larger and larger part of the portfolio? >> well, they were growing. but in 2004 and early 2005 they

hadn't reached the numbers that were on your slide. again, we had a data set that hold us we needed to increase the numbers in our model, and the fact that it didn't get implemented in a timely matter. those increases were postponed. but the other side of it is, in all candor -- and i don't want to get into an analytical debate, but when you have a pool coming in with 10,000 loans, each one with 85 data points, it is hard for an individual to look at it and come up with a decision on would is a five basis point enhancement on the triple-around. you can't do these without models. there is an area you might a dress. the fair credit reporting act prevents the agencies from getting the kind of in-depth information on borrowers that would help them gauge credit performance expectations. it is post the loan being made,

so it is not going to be a disparity impact on people trying to get a loan, but it gives the credit agencies additional information in helping them track. we did not have the income. we had to back into the income numbers with the ratios they gave us from the front and back end because they were not allowed to collect it. so there were issues there with the information available that would take us out of the box. >> if i could add something. for example, in that situation, if you receive 10,000 loans, there is no way you could examine each one of those. i wouldn't expect the ratings agency or anyone to do that. having said that, if a banker comes to you and says look, i have 10,000 loans, one reasonable thing you can do is take a random sample, see if what you find agrees with what the banker tells you. we did that many titles when i was there.

not in the context of mortgages, but in the context of different things. so that would be a way to handle that situation. >> all right. >> on this issue, i believe models are important, and you can't do these without models. you need a human being to have a qualitative judgment of what the results are and what the input is into a model. a financial model is like a weapon. it could be useful -- if you are holding it, not so useful if it is being pointed at you. >> my experience has been that once the model is out there, once it has been published, bankers and originators understand how to gain that model. >> exactly. >> and they will -- anecdote ally, there is information on these loans. a lot of times, the loan officer new the borrower and

had that information. but to extend that information to the whole universe of borrowers may be statistically workable but makes no sense. this is where you need the qualitative judgment. i know why this was happening, but the model doesn't make sense any more. >> one thing was barbelling. are you familiar with barbelling? >> yes, it is layering of poor assets with goo assets in a portfolio. >> and isn't it almost a perfect of are example of what you were saying? people now begin to gain the model? they are not trying to figure out what the best product is, they are trying to put together a product that meets with the model? >> one of the lessons learned for me personally is that averages lie. if you have a model run bean aaverage, it is not telling you all the information. >> do you have any commenting on barbelling? >> it was an early example of how we had to respond in what

became an aggressively -- or an increasingly aggressive game of cat and mouse. effectively once we had published criteria, and this was the established requirlte, we would quickly see that here was a portfolio that presented some compliant averages as eric was referring to. and then there was our response to that, in which case that generated yet another response to that, and so it goes. >> mr. raider, was bar belling -- barbelling a problem? >> i think it may have been raising its head, but you can fight the concept of barbelling only by maintaining the models as accurately and gaining the most amount of data you can get. that requires, frankly as the market exploded and the new products arrived, we should have been looking at coming out with new models every six months. a year was the worst.

anybody is going to try and game it, and the only way you can avoid gaming is to keep improving it to capture the nuances that allow the gaming. >> did you have any indication that management was concerned about this? this has come out at another problem. this massive flood of new business. business just exploded based on the chart we saw earlier. it is a little like having a restaurant and tripling the number of people eating there without keeping the same kitchen. did you see any indication by management that they were concerned about the fact that we weren't doing these new models, we were getting a massive influx of new business, and this could be a real problem? >> no, we didn't get any indication that it really bothered them because they were turning us down for staff. they were turning us down for the resources we needed. what they were looking at was you must not have been working very hard because your volume is doubled, and nobody is

quitting. so i guess you had slack down there, and they were just enjoying the revenue. by 2005 when i left, we were getting calls from corporate monthly on how much money are you going to make this month? structure was driving the whole ratings business, and rmbs was the fastest growing unit. >> mr. 2005 when mr. raider left, by 2007 or 2008, there must have been some understanding by management that this was not all going well? >> absolutely. there had to be. >> did you ever see any indication of that? >> certainly one good example would be in the effort to try to catch up, it was clear that from 2004 on, that we were playing a game of catch-up in terms of staffing and in terms of systems. we were trying to transfer a lot of the input, data input, for our monitoring process to

sources offshore to try to speed up the kwanity that was being driven. we were working on a model. but effectively all of the resources are being directed backward, trying to fix what was acknowledged to being broken, as aposed to forward. it was how do we survive this onslaught of work? >> did you ever think about what was being done in management? clearly the brand was not doing well. you could see an erosion of the quality. >> sure. i think it was common to commiserate, particularly among the rear guard, those that had been there longer and had seen the culture change most dramatically, that the place was not what it won't was. personally i was anticipated a

steady erosion until there was just a pure equivocation between the rating agencies. so it was a matter of just throwing a three-sided dice, and you would pick that rating agency, it didn't matter any more. it was disappointing, but those that did want to try to continue to pursue the good fight, if you will, i think we did our best under the circumstances. >> i think the drive for market share was up front and center, and any other resources didn't really matter as much, and the focus was how much revenue we made. >> i got that. did you have any indication that management was aware of the innocentivities, all the things we talked about, market share profit. did they realize there was an

erosion of the product and that this was not a good long-term strategy? >> i don't know. they must have. they must have understood that this was not something that could continue, but i am not sure if they did or not. >> from your vantage point far away from this, how did it look to you? >> well, in my view is a little bit from own outsider because i left in 199. but being in the market, and to some extent i am in the market. there were a few funny things going on regarding the ratings assumptions. i remember i gave a talk at the c.d.o. conference at the end of 2006, and i made the point that there were many, many changes to many, many things. the probabilities, correlation,

things that might sound technical here. but the changes were a little suspicious. there were some changes at that time to something called correlation, which we probably don't have the time to discuss what it means here, but basically it amounted to a relaxation of the standards. so anyone looking at the rules of the game from the outside would have noticed that certain assumptions that were made in the past, now they were a little bit relaxed, and certain trance transactions that were receiving triple-a ratings, with a few calculations, you could come to the conclusion that they probably were not. >> we have been here for a while talking about this. mr. raider is there anything you can say to sum up what you think went on during this period? >> thank you, senator. i would like to add a brief point here, if i may, because i think it is relevant. it may sound like a subtle

point, but i think it is very profound. the problem we have here i believe is a little more serious than what we think it is. we have been talking here about triple-a and triple-b, and the rating agencies were given the right to determine whether something is triple-a or triple-b or whatever. fine. but there is another level of complexity here. congress has given the ratings agencies also the right to define what triple-a and triple-b means. congress has given the agencies the right to legislate. this is crazy, but i think i am going to finish my statement here. i think sometimes an example is more clear than an explanation. suppose you pass a law saying in washington, d.c. you cannot build a tall building, but you

forget to define what tall means. now there is a private company that will decide what tall building means. so that company might decide it is a five-story building. next year they may change, and it is a 10-story building. nobody knows what triple-b means. the only thing that is known is that if you are an insurance company, you cannot buy anything below a triple-b. if you are a pension fund, you might be forced to sell. the rating agencies not only determine whether something is triple-b, but they can change the definition of what triple-b means, and i think that is a very extraordinary state of affairs. it is very screwed up oned fundamental level. >> because the ratings are used by the agency to make decisions. you are right. i never thought of it that way,

but essentially the rating agencies are determining on their own what these ratings are. >> exactly. legislating all the time. >> thank you. thank you very much, mr. chairman? .. >> it was part of the daily work

load. >> it was not particularly prominent in residential because our model was destroyed it. everybody got the center. we could not have bear stearns from a deal and print in any destructor and give something else to someone else. were somewhat insulated and this particular area, yes sir. >>3? in terms of that kind of pressure that you felt, let me ask mr. kolchinsky, let me ask you because you already made reference -- have you ever heard the phrase ibgybg? >> that was a quote to made. i was working on a transaction

with a large bracket bank. >> what does that mean? >> that is what the largest banks that has a large balance sheet. >> which bank was that? >> i think it was due to back. >> --deuschebank. >> it was explained as "i'll be gone, you'll be gone." what are you making a big deal over this comment? you said laughingly as if you are losing perspective here. did it mean to you that you should think short term because everybody would be gone before the chickens came home to roost? >> what was told to me, i did not realize how that thinking was driving much of what was going on, actually they're not

short-term thinking? >> short term and get this deal and get this quarter closed and get this bonus. i don't know whether or not my group will be here at the end of nextlra corporate i have to thik of this next deal. >> basically, which was your client? the investment baker -- banker? >> the banker. the bankers will looked on as clients. if an investor called, that would be fine but they never did. it was simply the bankers and

they were the clients. >> anyone? >> thank you, you have been helpful, all of you. you've come a long distance and we appreciate it. [no audio] [no audio]

>> we will now move to our second panel. the witness will be the former north american practice leader of residential mortgage-backed securities at standard and poor's. yuri yoshizawa, a group managing director for structured finance at moody's investors service and finally, peter d'erchia, currently managing director of u.s. public finance and global practice leader of surveillance at standard and poor's. we thank you for being with us today.

we have a rule here which you are familiar with. it is oral sex. it requires -- it is rule six. it requires our witnesses to be sworn in. do you swear the testimony you are about to give will be the truth, the whole truth, the -- and nothing but the truth, so help you god? >> yes. >> the timing system we will use today will be a red light that comes on about one minute -- about five minutes after you begin. one minute before that light comes on, it will be changing from green to yellow which will give you a chance to conclude your remarks. we will print your entire testimony in the record, of course. we would ask you to limit your oral testimony to no more than five minutes. miss burns, we will have you go

first. mr. yoshizawa will be next and then mr. derr shepherded miss burns, please proceed. >> mr. chairman and members of the subcommittee, good morning. from 2005-2008, with the north american practice leader for residential mortgage-backed securities. i have been asked to appear today to discuss ratings for our and the product. i want to begin by saying dead s and p, we have learned hard lessons in the difficulties in the subprime area the subprime mortgage area improved credit for minds of americans, apparent abuses have been -- have had a reverberating effect of the economy. we began downgrading some of its rigs in this area in 2006 and had warned of a deterioration in subprime sector before that. we were watching this market for the end of the day, the assumptions and criteria underlying our ratings did not

anticipate the extent of the collapse of the housing market which has been more severe and we, along with some others, had anticipated. although my focus is on the s&p process for rating securities, it is helpful to discuss the nature of our credit rating. the s&p credit ratings represent our opinions of the likelihood that a particular financial obligations will repay principal and interest. ratings do not speak to whether an investor should buy, sell, or hold rated securities or whether the price of a security is commensurate with its credit risk. this is part our ratings process, as simply does not raise the underlying mortgage loans made to borrowers or evaluate or regulate whether making those loans was a good idea in the first place. our resonators make loans and are responsible for verifying information provided by borrowers. the also make underwriting decisions.

in turn, insurers and the rangers of mortgage-backed securities under those loans perform due diligence on those loans constructor transactions, identify potential buyers, and underwrite the securities. our role in the process to reach an opinion as to the ability of the underlying loans to generate sufficient proceeds, to pay the purchases of securities issued under stress scenarios that correspond to reading levels. in doing so, we rely on the data coming from issuers, rangers, and servicers that other market participants also rely on. the market must be able to rely on these participants to fulfill their roles and obligations to verify and validate information before they passed on to others. the s and p analysis of the transaction it valued the credit worthiness and cash flow of a pool of mortgage loans. we use models that reflect our analytic assumptions and criteria. the models supply those criteria to a particular loan

pools using 70 different data points. the assumption and analysis are under regular review and are updated as appropriate. after reviewing relevant information, an analyst present this to a committee agreed to qualitative judgments of committee members are an integral part of the ratings process as they provide for consideration of asset to transaction factors, taking into account the tet and experience of the committee members. the chief component of our analysis is assessing the credit rating. for subprime loans, which are expected to perform worse than prime loans, it would have significantly more credit enhancement than similarly rated transaction. through securitization this does not become solid investments request is what amount aaa

ratings candace support text was the rating is determined by the rating committee, s&p noted by the issuers and disseminate the rating to the public. along with the redding, we republic published a narrative to release further information. the subcommittee has asked me to speak to the awareness of brought in the subprime mortgage market. the s&p was aware of these reports and from 2005-2007, we consistently but informed the market about their concerns. we revised our models and took into action when action was appropriate. i think it for the opportunity to participate in the hearing today and i would be happy to enter any questions you may have them not at the very much. mr. yoshizawa -- >> i am a senior managing director of the derivatives group at moody's my group rate

the various derivatives including collateralized debt obligations. i would like to thank you for the opportunity to be here today. moody's placed an important but narrow role in the investment opportunity industry. we publish credit ratings on entities including corp. governance. our credit ratings or opinions about the future likelihood of a full and timely repayment of debt obligations like notes and commercial paper. we guard was whether the debt is issued by a cdo or a corporation , we follow methodologies and it here to established procedures i will discuss these as they pertain to cdo's but let me give you a brief overview as to what they are. they have been around since the early 1990's. the cover a wide range of instruments and can have various types of assets

including securities issued by financial restitutions, corporations, and other structured finance should courteous. they may be static or manage transactions. ecstatic transaction is typically not subject to change in managed cdo's the financial institution can vary. our methodology for rating cdo's incorporates quantitative and qualitative factors this includes a credit risk associate with the collateral backing the cdo its structure. the qualitative factors that we use include governing documents of the cdo, a lateral manager and a trustee. these factors vary. moody's run the trading process through committee system. rating committees decide the reddings rather than any one individual. after the analyst attains

relevant information from the issuer, he or she and corporate information from public sources as well as movies on macro economic perspective. they formally debut and presented to a committee. the committee members are selected based on relevant expertise and diversity of opinion. it never is encouraged to express views and discusses differences in an open and frank manner. once a full discussion takes place, the members of both currevote. each committee members boat carries equal weight and a majority vote decide the outcome. once the rating is published, we monitor the rating on an ongoing basis. we will modify it as needed. one common misperception is that our credit ratings are derived from the application of mathematical process or model. this is not the case.

models are tools and sometimes we use them in declining ratings but the credit rating process always involves much more. most importantly, we exercise independent judgment by the members of the rating committee. our committee system is at the core of everything we do and it is designed to protect the quality, integrity, and independence of our ratings. having said that, we recognize we must continually reevaluating our methodologies and processes to determine how a -- how they might be enhanced further. let me make clear that we at moody's are not satisfied with the performance of our ratings instructed finance cdo's over the past several years. thank you and i would be happy to take your questions. >> that you very much ms. yoshizawa missed. dear ship.

-- mr. tursha. >> i am peter door shut, the managing director of standard and poor's rating services. i have been through -- from 1997 through 2008. as the head of the group, a supervised surveillance for five major categories, residential mortgage-backed securities, commercial mortgage-backed securities, collateralized debt obligations, asset backed securities, and servicer of valuations. each of those groups was headed by a separate manager who reported to me as a member of my management team. the s&p's structured finance surveillance portfolio grew substantially during my time as the group's head. our resources expanded to meet this increase in workload and our growth respect to rmbs

outpaced the increase in standard and poor's monitored increase. the rmbs increased by 75% from the beginning of 2003 through 2006. in 2007, responding to the unprecedented deterioration rmbs deal performance, the group increased an additional 57%. the standard and poor's surveillance work in 2006 and 2007, consider the basic process behind standard and poor's surveillance review. after a rating is assigned on and rmbs transaction, a new one, it is transferred to monitoring. the surveillance analysts take information related to the actual performance of the rated pool overtime and uses that

performance data to assess whether standard and poor's rating is appropriate in light of the evil and you of the current credit support of the deal. standard and poor's surveillance process comes from a new rating review. this reflects a practical reality of difference pre and post-deal metrics. it also serves an important function in providing a form of analytical check and balance therefore, each deal rated by standard and poor's is subjected to two and local processes provided for a more robust analysis than a simple reapplication of the same initial method over time. as i noted, the bottom of rmbs rated transactions increased throughout the years 2003-2007. in the past, standard and poor's look for at least 12 months of

performance data for seasoning. we used a variety of internal monitoring tools to identify and track individual deals for closer review. in late 2006, as the performance of recent vintage u.s. rmbs transactions experienced broad deterioration, standard and poor's surveillance group began a process of individual reviews to prioritize rmbs transactions and to monitor those entire annual but it is on a monthly basis as each month's tool data came in. starting in 2006 and for the first of 2007, delinquency data coming in each month did not resemble anything seen before. at that point, no significant realized losses had been reported to the deals under review. moreover, historically,

delinquencies did not always lead to losses. accordingly, there was a lot of analyses. we have to determine what that demand for it as receive more data in early 2007, standard and poor's new issue surveillance and criteria personnel all worked together to understand what was happening and how to respond. ;during our ongoing analyses in early 2007, standard and poor's took numerous significant steps to react to the deteriorating rmbs performance and inform the market of our analysis. we recognize the unprecedented nature of the early to like what is occurring in the 2006 vintage. it fundamentally changed in february of 2007 to reflect issuances -- issues on creditwatch without warning for losses to develop. -- without waiting for losses to

develop. standard and poor's continued to downgrade ratings as appropriate on an ongoing basis by july, 2007, standard and poor's had adopted its methodology sufficiently to issue such a substantial number of downgrades. that the pollution continued through july, 2007 resulting in further downgrades and a subsequent performance data and criteria warranted. finally, i would like to thank the members and staff of the subcommittee for giving me the opportunity to participate in this hearing. my experience at standard and poor's has always been committed to doing the best we can to develop and maintain inappropriate ratings and i am proud of the hard work of our team in trying to understand and respond to a historic market disruptions to the best of our abilities. i have set for additional information about my own work about the rmbs market in a

written statement. i am happy to answer any questions you may have. >> thank you very much. let me start with some of the failed deals we went into the first panel. first is the vertical deal. we are looking at 94b. this is a cdo known as vertical cds. in 2007, it was complained how the particle is sure, ubs, was not cooperating and how the deal was unlikely to perform. this is what one s&p analyst wrote in that 2007 e-mail. " particles political closely tied to bank of america is mostly a marketing shop helping to take risk of books of bankamerica. i don't see why we have to

tolerate lack of cooperation. the deal is not likely to perform." despite that judgment, the s&p rated it and so did moody's. both rated the top three charges as triple-a and defaulted -- it defaulted two months after. mr barnes, s and p was uncomfortable but it was raided. should that have been taken into consideration? >> i cannot speak to this example. >> but you can speak to the question if there is a lack of cooperation and a deal is not likely to perform, should be rated? >> the analyst would take that into account and it would be part of their approach. all the members in committee

would look at this. >> the analyst concludes that the deal is not likely to perform. should be given triple-a ratings? >> the committee process at standard and imports, no one person considers a redding. they can have different opinions that will come to committee. >> how about of the committee decides that a deal is not likely to perform? >> then it would either take action or change the ratings. >> if an analyst says that the bank of america is using the cdo take the risk of the books, should that be a factor? it the analyst banks that was going on with the cdo is taking bad assets of the company's books, does that affect the ratings process? >> i cannot speak to the aspect of the cdo credit area.

if they think it was material and could impact the performance, i would expect that they would, yes. >> then it should affected. >> if they believe that, it should be an aspect they should consider, yes. >> i am asking you have -- not if they think something, i am asking for your opinion. >> i would agree that it would be an aspect of the credit perspective. >> let me ask you about -- let me ask you ms yoshizawa, this is exit 94e ---- this is exhibit 44e, securities were included debt were downgraded the year before. they had not been performing as expected for it is that common? is the common for a cdo to

include downgraded assets? >> it can include assets that have been downgraded. >> that is common? >> it could be common, yes. >>qbñ ticket look if you would t 93d, in january 2007, the snb was bask -- the s&p was asked to rate cdo's from fremont. an analyst sent e-mail that said they have a deal with fremont collateral. fremont collateral has not been performing so good. is there anything special i should be aware of? one supervisor said no.

they don't treat their collateral any differently. in 93c, as long as they had current fight goes course for the bourse, the analyst was good to go. two days later an article was circulated noting that fremont is not now using 8000 brokers because their loans of some of the highest delinquency rates in the industry. that is 93d. by the way, fremont had also announced a filing in the california court of appeals that marketing and extending adjustable rate mortgage products to subprime bahr was in an unsafe and unsound manner greatly increased the risk that borrowers would default on loans that would otherwise cause losses the lawsuit was allowed to continue.

there was a cease and desist order that remind entered into with the fdic regarding fraud and lax underwriting standards. despite all this information, we got fremont's rmbs securities are rated both by s&p and moody's in spring of 2007. should they have taken those factors into account? >> managers share their views that the criteria at the time should not have been modified. >> i'm asking which should be the case. fremont was not performing so good. that does not make any difference? according to this e-mail, it doesn't make any difference.

should make a difference? >> when we look at those model, the data used in those models is used from and an array of originators and performance and used to establish the expected performance. >> i am asking you a simple question. it is a common sense question. you have an analyst who is saying that the collateral has been performing not so good. the supervisor says that does not make a difference. should it make a difference? >should the collateral mckay difference? >> yes, the collateral should make a difference. >> why doesn't he get an answer back that it does not make a difference. >> because it goes in the context of what the criteria is including. if you're looking of the overall performance of the industry and how far it is deviating, what they are reporting back to is that they do not think an

outside adjustment or a change to the assumptions is a program at that time. >> this is not assumptions. this has to do with the deal. he said he had a goldman sachs deal with remote collateral. they say pre what is not performing well. there are all kinds of problems with fremont. if fremont does not get a higher credit risk, i am trying to figure out who does. how can you just simply say it does not make a difference? something which says you folks are supposed to be assessing credit risks here and it does not make any difference that their collateral is not performing? that is what this e-mail says. they say they do not treat their collateral any different. should their collateral be treated differently if it is not performing? but as a simple question. -- that is a simple question. >> if it is performing differently than what our models projected, yes it should be

changed. >> is the fact they're not performing well relevant? >> it would be relevant to the analysis. >> was it relevant to the supervisor? >> you have to look at it in the context of how the assumptions are built. they are built on an array of data. and how far off and where is the performance is expected to perform and just because the performance is 4 it does not mean a deal was under enhanced. with a poor quality loans put into a transactions that have an hour of credit enhancements that could reflect that poor credit quality arabic because delinquencies are high at that time -- for that particular vintage. i would not say it was ignored. >> he was told to ignore it. take a look at 93a. this is what the -- what it seems to me will shake up folks that are listening to this

testimony, to the credit ratings can be relied on? here you have a chart. this is a movie deal. -- this is a moody's deal. they have issuers of high delinquency. it comes back. holy cow, is this data correct tha? is that relevant? >> it is death, irrelevant. >> why doesn't the supervisors say damn right, it is relevant. you're trying not to enter the question. should the supervisor had said it is relevant. you better dig into this. >> it should be relevant investment in the performance, yes. >> thank you. should that have been relevant? is that a factor that fremont as such an out liar that they're

one of the worst when come to delinquencies? is that relevant or should be relevant to your credit rating? >> my understanding -- i am not responsible for the rmbs area however, my understanding is that the originator and servicer and their collateral quality is a factor in the analysis, yes. >> and should be? >> yes. >> i am glad to hear that. we took an in-depth look at the early hearings by the home loans issued by washington mutual. they were riddled with lax lending standards, income that had not been verified, appraisal problems, load errors. they have one of the worst delinquent stearates in the country along with fremont.

it seems to me it is obvious you should distinguish between lenders when you do these kind of the analyses. i'm glad that movie did this. it seems to be that it is obvious that it should have been done. i am stunned that the reluctance of the s and p two say they obviously should be a factor. at the end of the eighth time i asked a question, i got the answer that it should be. a fact . let's look at exit 95. i'm not sure what the pronunciation is delfinnes. this is a moody's deal. i will address this to you, ms. cubists -- your shoes out. -- yoshizawa. this was raided and downgraded

over six months by both moody's and s&p. first of july, 2007, it was raided. january of 2008, it was downgraded. it became junk status. in 95a, an investment banker or to a moody's analyst. it was a mezzanine to hold a lot of cushion so they did not care that much that was july 2007. that question obviously wasn't big enough. and then you look it 95b which is an e-mail of july, 2007. this is an s&p analyst at wrote this -- regarding delifinnes, it appears to the closing date portfolio for analysis and the effective date portfolio were not the same. it appears that the 25 assets they include in the pros --

closing date portfolio were dummies. they were replaced and less than 24 hours with assets and would have been notched and made the portfolio worse. what are dummies being used in this way? this 95b, this is a standard and poor's document. i will ask you again, mr. barnes. the dummies. >> i cannot speak to that because i am not in that area. >> ok, mr. dershowitz? -- mr. derr shock d'erchia? >> this would be discussed and the new transaction side. >> do you know anything about the use of these dummies generally? have you heard about that practice? that is substituting assets shortly before the ratings come out. >> i am not familiar with that.

>> are you familiar with this ms. barnes? >> not with respect to cdo's. >> do you use dummies at rmbs's? >> sometimes but not in this manner. people would be -- would use something that is representative of collateral but would not be a substitution, so to speak. it would be was the collateral was originated. >> it would be a substitution of one asset for another. do youí0s know the use of dummi, ms. yoshizawa? >> as i mentioned in my opening testimony, cdo's need to be static or managed. i do not know about this transaction but it appears it was a static transaction and you would expect that the portfolio that was provided to us at closing would be the same portfolio as the origination date.

>> let's take a look at exhibit 1i, this is what happened to the triple a's. black rock made an assessment and looked at the entire universe of aaa ratings that have been given to rmbs securities from 2004-2007. subprime rmbs security , half of the rigs are junk and 93% in 2006 are jump, option arm securities or securities that

they looked at and we looked at this in the first hearing. they are fully -- full of negatively amateur rise to -- amateur rightist who amortized loans. nobody knew how many the bowers would default and would have to pay the higher loan payments. we examined the enough traffic show that washington mutual wanted to shout -- sell its up and farm loans starting in early 2007. it had already decided they were likely to fail so they wanted to get rid of them to get off their inventories. later in 2007, they securitized billions of dollars of their option on loans at 7 mortgage loans that resulted in polls that one aaa ratings. this shows that 97% of the aaa ratings given to option arm

securities in 2007 and 2006 had now been downgraded to junk. the numbers for other high-risk loans are equally shopping. this included fixed and variable loans with little or no documentation. the aaa ratings have fallen to junk status. prime, fixed, and very boulogne's -- and parabolan loans are not included a charge. does this shock -- does this shock you? does that chart shock you? >> yes, mr. chairman, it is shocking. 90%? my group would have been responsible for lowering these ratings.

in doing so, there is a tremendous amount of work. >> we will get to that. it takes awhile to downgradings but we will get to that later. does this shock you, mr. yoshizawa? >> that is not what we would expect. 'w>> ms yoshizawa, the first pal of witnesses raised concerns about moody's handling of the cdo's here are some things they said. pressure by the investment bankers, pressure applied by managers at moody's to maintain market share over and over again. that is what we heard. pressure to maintain market sharermbs and cdo's, next, managers felt like they would lose their jobs if they lost market share. next, the drive for market share led to the deterioration in credit standards applied by

moody's. next, emphasis on keeping the investment banker customer happy. keeping the peace, next, lack of adequate resources to rate cdo's effect of lake. changing culture was placed on meeting the investment banker demands. next, the quality assets assumed by moody's model was not the same as the collateral provided by the investment banks. next, banks were progressively pushing cdo's to get into the summer. all but that you heard, does that trouble you? >> yes, all those statements are troubling theme were you surprised to hear them? have you heard them before? >> i have heard that before, yes. >> senator kaufman.

>> for the record, thousands of rmbs's rated aaa into a dozen sex and 2007 are rated now his job. can you give me a couple of reasons why that happens? >> well, senator, the assumptions that we use in our criteria have obviously not panned out the way we had expected. the market had deteriorated more precipitously and dramatically then we had expected. >> the assumptions did not work out and it was the overall market went down? >> yes.

>> that's it? that is why an organization like yours with a long and honorable tradition of breaking triple-a all the sudden have a massive failure, i think that is fair to say, and your assumptions were wrong and it was just a marketing tthing? the housing market went bad? no one could perceive that would happen. if you look at the housing market, it would seem it was going into areas it had never gone in before. there was no discussion at s&p about the fact that you have a

chart like this to describe the housing prices and that might be a factor? >> there were many assumptions that we use in the ratings process. the market decline was one of number of the ultimate default rate and things that impacted doubleton us to parity -- the defaults and the severity. as market depreciations impacted people acquisitions, those were things to consider. >> ms yoshizawa, why do you think there were thousands of rmbs's that were created -- that were rated aaa in 2006 and 2007 and are now rated as john? >> once again, i am not in that area however, >> you got the flavor. >> we strive for and we believe

we have achieved historical accurate and reliable ratings. even outside of those structure transactions and others that had been affected by the housing crisis, we have ratings that have informed the structure. we certainly did not predict when we were rating these instruments and looking at the long-term credit expectations on these instruments the magnitude of the housing crisis. we did not predict the velocity or the length of time over which this crisis would extend. that is the main reason, i think. >> from a surveillance perspective, we had the value of seeing the actual delinquencies.

we had never seen that precipitous a rise on a comparative basis with previous in vintages in that period time. >> this is 2006-2007. you're doing surveillance but not rating. you're not involved in the national rating, you're just looking back and trying to re- rate it? >> yes, we have once a regular column that is to make sure the rating is appropriate. >> let me start with you. in 2005, the s&p was already talking about the deteriorating market.

in your testimony on page 6, you talk about an article with subprime lenders and clouds forming on the horizon. following the internal housing market simulation conducted in 2005, the s&p felt this study concerned the potential impact of the housing turned count -- downturn. it wasn't like a shot. into about -- into doubt -- it wasn't a shock in 2005 but then in 2006, maybe you're the wrong person. i guess i should talk to you, ms barnes. in your testimony, on page 11, you recount the same thing. it was not -- in 2005, you knew

there was a problem the housing market. there is the chart i was holding up. this shows the fact that by 2006, we were now operating june, 2006, the housing bubble was -- it just went through the roof. you knew in 2005 that this was a big problem yet in 2006 and 2007, you're still getting aaa ratings to a whole series of rmbs' wound up being a jock. >> yes, senator. through that time. we are looking at the market and trying to understand the

development and what was happening the study in 2005 about the housing bubble, people were discussing that in the marketplace. we want to see what impact the ratings of that scenario a corporate and 2005 on, we were looking at the different developments and collateral we were seeing. >> i got all that that during that period, the main thing i am interested in about these instruments -- you did not give a lower rating. they still got a solid rating which understand -- which i understand. you are not indicating any kind of consent about the subprime market. w-so many were triple-a which turned out to be jmnk.

>> we did released an update on our criteria during that time period. that increased their credit enhancements so we expected losses and defaults to occur and that increased their credit enhancement which we believed at that time would be reflective of the ultimate the polish -- exposure of those deals. >> your testimony on page 18, you and mr. mcdaniel, you go back to 2003 and talk about the loosening of underwriting standards. that was four years before 2007, if my arithmetic is right. there had to be some knowledge and the management of movies that there was trouble here. it did not all the sudden here that in 2007, there was a housing problem of massive

proportions and you're still reading everything triple-a. -- rating everything triple-a. you're still putting up aaa bonds which in retrospect >> from 2003 on, as we mentioned, we wrote about the various concerns they had a market. at the same time, my understanding is that they were increasing enhanced levels required to get certain ratings. there were changes made to the methodology for creating a triple-a rating. i do not know about the specific practices as to how much and has been a change. that was not my area. however, i am aware that they have continually been identified and justifying their methodology. >> did all three of you here the

previous panel? >> yes. >> mr barnes, the previous panel said that was not it at all. they said was lack of regulation. they said there were incentives given to folks within the organization. they said it was a search for profit pe. said was market share. do you have any firsthand experience of any of those things having an impact on what happened? >> i have no personal knowledge of people's incentives being directly tied to the number of deals they rate or raided. -- raided. --rated.

but overall financial performance of the organization was good. >> no one in the organization you know felt that the incentives distorted what decisions they're making? >> not on the deal level, no. >> in terms of personal incentives, the analysts were not compensated based upon the performance of an individual group. >> i am talking about everybody up the chain. there is a whole process with analysts. we're not talking about the analyst, the whole chain you had. you went into great detail to explain how this works. do you think folks in that chain fell there were incredible incentive to get as many deals out the door as they could with the highest possible rating? >> i did not perceive an undue

pressure from market share. i had been responsible for the cdo area. we have very low market share, in the 20%, 30% range. i was expected to explain why our market share may have been lower in terms of our methodologies. i was expected to know why. that was the case. at no time was i told my mandate had been to increase that market share. to retain market share we had. >> so what the other panelists talking about is something that happened with them but not with you. the more deals the better happened to the folks in the other panel but you did not feel that? when you were out to lunch, people did not talk about more pressure to do business and make

deals work? you did not feel any of that? >> i did not feel the pressure to do deals at the cost of credit. >> how about pressure just to do deals? >> we were supposed to trade deals that we could rate. we were expected to rate them up the levels they should be rated. it is not black or white as whether you can write something. sometimes you do not have enough information. sometimes the complexity of the action will not allow you to come up with a rating. >> you were in a different environment than the first panel. but the environment they talked about was something that did not impact you. >> i heard their testimony. >> they were working for the

same company but they were saying things differently from you. what do you think? >> i had wanted% market share and workload came from the new deal size. de. i would be in a different place. >> i can appreciate that. i think, mr. chairman, i am going to turn it back to you. [no audio]

>> ms yoshizawa, take a look at exhibit 24a. this is an internal moody's e- mail exchange from early october, 2007 for the second e- mail is to you. who is sunil surana? >> she was a business analyst at that time reporting to my boss. >> here is what she wrote. this is in 24a. she is writing to you.

"could share by discounting dropped to 94% by volume is 97%. that is lower than the 97 +% in lower courts. any reason for concern? our issue is being more selective to control costs? is it cheaper or is it an aberration? what was your answer to that? >> we were expected to know why we were not on deals that we were not on. it could have been about credit enhancement or other relationship issues. >> market share mattered. isn't that what that email says? >> we were expected to know what the story was part of a past that onto my directors to find out. i am sure there was a list of transactions and a one to let them let me know what was happening. there's no pressure to rate there's no pressure to rate those transaction