The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Statement of Richard Michalek Former VP/Senior Credit Officer, Moody's Investors Service Submitted to Financial Crisis Inquiry Commission June 2, 2010 and originally submitted to the Permanent Subcommittee on Investigations Committee on Governmental Affairs United States Senate April 23, 2010 My name is Richard Michalek. I am a former employee of Moody's Investors Service, a subsidiary of Moody's Corporation. I joined the Structured Derivative Products Group (the “Group”) at Moody's in June of 1999, and my position was eliminated in December of 2007. At the end of my tenure at Moody's, I held the title of Vice President/Senior Credit Officer. My responsibilities during my tenure included performing legal analysis on the structure and documentation of complex structured finance transactions in order to assign a Moody's rating to that transaction; reviewing and analyzing forms of standard and customized documentation used in interest rate hedging transactions and complex interest rate and credit risk swap transactions; participating in rating committees within the Group, and on request for other groups at Moody's; consulting on legal matters for other groups in the New York, London and Asian offices of Moody's when requested by applicable managing directors for those groups; speaking at industry conferences as and when requested by my managers; preparing and publishing the Group's quarterly and annual review and survey of activity; and participating in and, when requested, organizing and leading, the legal portion of semi-annual training sessions for all new Moody's analysts working in structured finance. During my last year at Moody's, my responsibilities were split between serving as the senior legal analyst on the team responsible for developing, refining and implementing the methodology for rating Credit Derivative Product Companies, and being the project leader responsible for developing a methodology for rating collateral managers and rationalizing the Group's use of resources in maintaining accurate data on collateral managers. Immediately prior to joining Moody's, I was a securitization consultant advising the pre-eminent New Zealand law firm of Chapman Tripp, and prior to that I was an associate lawyer in the structured finance group in the New York office of the law firm Skadden Arps. I am admitted to practice law in New York State, and was admitted to the bar as a solicitor in Wellington, New Zealand. I have a JD/MBA from Columbia University Law School and Columbia University Graduate School of Business. Statement of Richard Michalek page 1 of 21 My statement today is not delivered with the intention to bring harm to any individual or to stand in judgement of individual behavior. On the contrary, as I hope my remarks will illustrate, I believe that any imperfections, flaws and failures observed or identified in the credit rating process are neither surprising nor unexpected in light of the context and framework of incentives presented to the competent and otherwise capable and rational people comprising the employees and management of the credit rating agencies. I. Practical Overview of the Rating Process Credit rating agencies serve the important function of providing buyers and sellers of credit – that is, investors in and issuers of a promise to pay – an independent measure of the risk presented. In theory, these agents are independent, and because of repeat experience and a rationalization of costs, are able to provide this measure of risk at a lower cost than would otherwise be faced if the buyers or sellers produced the analysis themselves. In practice, issuers of debt obligations buy a published rating from a credit rating agency to enable investors to compare the issuer's offering to other debt obligations that also carry a published rating. Investors use the ratings to also satisfy any number of possible needs: institutional investors such as insurance companies and pension funds may have portfolio guidelines or requirements, investment fund portfolio managers may have risk-based capital requirements or investment committee requirements. And of course, private investors lacking the resources to do separate analysis may use the published ratings as their principal determinant of the risk of the investment. Even though each rating agency is attempting to describe the same thing – the risk presented by an investment opportunity – the ratings themselves differ both in form, and in definition, as between the rating agencies. Without going into detail, Moody's ratings assigned to the derivative structured finance products that I used to rate are a measure of the “expected loss” facing an investor, given the tenor of the instrument. The determination of expected loss includes making an estimate of both the probability of a default in the promised payment, and the severity of the loss experienced given such a default. In essence, the Moody's rating assigned to a derivative instrument is a function of the probability that underlying assets will default, and a measure of how much the consequent losses due to such defaults will impair the ability of the pool of assets to pay the promised amounts due on the derivative instruments. Keep in mind that these derivative instruments are obligations backed not by a single payer, but by a pool of underlying assets, and it is the funds raised by the issuance of the derivative instruments that are used to purchase the pool of underlying assets. I.A. Structured Finance – The basics Structured finance – the ability to raise capital and reduce financing costs through the aggregation and organization of both pre-existing and future obligations – in more ways than one raised the profile and magnified the importance of the rating agencies. Almost by definition, structured finance products beget new structured finance products, with ever increasing degrees of complexity. At the first level of finance, there are corporate bonds and loans, and mortgages issued by banks to home buyers. These obligations consist of a single obligor/borrower. While a credit rating assigned Statement of Richard Michalek page 2 of 21 to the obligation might help in transferring or selling the obligation, many issuers of these obligations held them to maturity. From the home mortgage issued by the bank to a single borrower, there comes “RMBS” or residential mortgage backed securities. From the corporate bond or loan sold in a debt offering by a company, there comes “CBOs” and “CLOs” or collateralized bond and loan obligations. From credit card receivables and car loan receivables and other purchase account receivables, there comes “ABS”, or asset backed securities. And then, when you issue debt backed by a variety of these structured finance products, you have a “CDO” or collateralized debt obligation. No matter what the name is, the basic aggregation and organization is similar. An investment bank or other structurer buys the cash-paying obligation (whether a mortgage from a bank, a credit card account from a retailer, or a corporate bond or loan in the secondary market) at a discount from the bank, card issuer or retailer who offers them at a discount because they are more interested in generating fees from originating (new) assets than in the income (and risk) of holding those assets until they pay off.1 The purchase of those assets is funded by the issuance of the structured finance obligation, which itself is sold in slices (or “tranches”). The “top” tranche typically holds an entitlement to the first dollars generated every month by the pool of assets. Since there is little risk that the pool will fail to generate those 'first dollars', investors buying this tranche will accept lower returns. Following that tranche may be any number of tranches, each with a slightly lower priority and/or different entitlement to the next dollars generated by the pool. The bottom tranche of course bears the greatest risk that the pool will not be able to produce the expected dollars, either periodically or upon maturity, and investors at that level of priority will demand a very healthy rate of interest.2 Because the issuer of the structured product is free to appeal to the preferences of different investors, the tranches may differ not only in priority, but also in entitlement. Some investors may want 1 The bonds and loans are purchased in the secondary market from owners that prefer cash today – in an amount less than par, or 100% of the face value of the obligation. The amount paid for the obligations does impact the rating analysis (because paying too much risks exhausting the “issuance proceeds” without buying the entire pool needed and necessarily will result in a shortfall of anticipated future cash flows, and because paying too little could result in “excess issuance proceeds” being released to the benefit of the most junior issued class of debt and to the detriment of the senior class of issued debt). However, the analysis is performed at the pool level, and not on an asset by asset basis. So, the rating does not necessarily evaluate whether each or any particular purchase of the assets comprising the pool was at a correct or appropriate price. This embedded opportunity to “overpay” for one or more assets represented a “collateral benefit” to investment banks and structurers, particularly if they wanted to remove an asset from their balance sheet. 2 It is a well-known feature of structured finance and CDO structures in particular that investors in the lower classes of the structured finance issuance considered their investment to be equity or equivalent to equity, and accordingly, evaluated their investment opportunity by considering the expected or targeted internal rate of return (“IRR”). Because the lowest class typically was entitled to all residual cash flows after more senior obligations received their stated coupons, the investor favored a structure that “front-loaded” excess cash flows. By accumulating a significant proportion of the pool in very high yielding instruments with coupons well above the weighted average coupon of the obligations above them, an equity tranche structured finance investor could collect enough excess cash flow in the first few years to achieve the desired IRR even if – due to defaults by the riskiest assets in the pool - the class prematurely stopped receiving payments altogether. This extreme misalignment of interests as compared to the senior obligations (who very much want the assets to perform over the full life of their senior investment) represents only one of the challenges to be addressed by a structured finance credit rating analyst. Statement of Richard Michalek page 3 of 21 “interest only”, while others may want the coupon to depend on the general rate of interest in the market, while others may want a fixed rate of interest. The same structured finance instrument can offer a wide variety of promised payments, for which the risk analysis can be – and quite often is – formidable. Enter the rating agency. I.B Role of the Quantitative and Legal Analyst in Moody's Derivatives Group, and Typical Timeline for Rating a Transaction The Group's policy on staffing deals for which a rating was requested was to assign both a quantitative analyst (a “Quant”) and a lawyer. In general, the Quant was responsible for reviewing the structurer's model and confirming adherence to published and necessary required parameters.3 To the extent that the structure presented failed to include the necessary parameters or should the model produce results inconsistent with the desired measure of expected loss, the Quant would work to identify acceptable modifications proposed by the client and then confirm that those modifications did not otherwise alter the stability and robustness of the model outcomes. The lawyer assigned to rate the transaction was responsible for (1) ensuring that there was no structural risk presented due to a failure to fulfill minimally necessary legal requirements (i.e., the issuer received appropriate legal opinions addressing key elements common to all structured finance transactions such as legally isolating the assets and confirming, by receipt of legal opinions, that those assets were not subject to prior competing claims, the issuing documentation was in compliance with applicable regulations, etc.) and (2) confirming that the deal documentation accurately and faithfully described the structure that was being modeled by the Quant. Of these two tasks, the process of conforming the documentation to the terms that were modeled was the most time-consuming, and more directly responsible for the ultimate performance of the transaction. For example, while the Quant could quickly scan the 10 or 20 cells and rows in the banker's spreadsheet representing the priority of payments in the structure and then compare them to the ten bullet point lines that are represented in a banker's “terms sheet”, those same ten bullet points might result in 10 to fifteen pages of turgid prose describing, in legal language in the indenture, how the trustee is to apply cash flows each month in respect of the interest and principal payment obligations due on the issued debt obligations financing the deal. Even minor errors in punctuation or syntax, like a misplaced decimal in the spreadsheet, could significantly alter the resulting obligation, and such errors would – particularly as the timelines for rating transactions compressed – regularly turn up. While the deal counsel invariably considered the “second set of eyes” valuable, the bankers considered that level of detail a barely tolerable annoyance. Similarly, when last minute changes were introduced, the quant could sometimes confirm the consequences of the change by re-ordering two steps in a single sequence of steps in the model. Those changes might entail changes in several documents, and in a variety of locations, re-introducing the need for careful, if tedious, review. Flaws in the documentation that remained beyond the closing date would require subsequent 3 Initially, the Group did not publish or use “its own model”, but instead would take each structurer's model and examine the assumptions, test it for robustness, and independently confirm that the results shown by the structurer could be replicated. Later, Moody's made a “standard” model available (but not required) for a limited number of transactions. Statement of Richard Michalek page 4 of 21 amendments, insisted upon by “liability abhorrent” trustees, and occasionally, where the error resulted in making a change to a previously assumed element of the model, could require placing the transaction on review for downgrade. This “post hoc” consequence typically was borne not by the investment bank who structured the deal, but the investors who originally bought the transaction or by the asset manager who was (in fact) the instigator of the transaction and was, following closing, responsible for maintaining the fidelity of the transaction to the governing transaction documents. The incentives for the “fee-based” structuring investment bank were clear: get the deal closed, and if there's a problem later on, it was just another case of “IBGYBG” - “I'll be gone, you'll be gone.” (First quoted to me by an investment banker who was running out of patience with my insistence on a detailed review of the documentation.) As a practical matter, a typical engagement for the purpose of assigning a Moody's rating to a client's structured obligation would include the following steps: (1) Client (an investment bank, in the usual case) contacts one of the Group's MD's to discuss a proposed structure and to confirm, informally, whether or not the instrument can be rated by Moody's; (2) If the discussion is positive,4 a rating application is filled out by the client and analysts would be assigned to the transaction (one quantitative analyst, and one lawyer)5; (3) Initial discussions take place, first between the client (represented now by a junior banker responsible for the “terms sheet” describing the principal elements being offered to the investors in each tranche) and the quantitative analyst, for the purpose of setting up the modeling of the transaction; (4) After the terms sheet is translated into an initial model and then generally verified as having features that can be modeled, client's counsel and the Moody's lawyer may go over the term sheet, and client's counsel produces the first draft of the principal transaction documents; (5) Client and counsel would then hold a conference call with the Moody's lawyer (and often with the quantitative analyst present) to review the comments that the Moody's analysts may have on the draft documentation and to discuss the fidelity between what was described in the terms sheet, what is being modeled by Moody's; 4 In 1999 and 2000, many of today's structures had not yet been created. There were any number of occasions when clients would present proposals for transactions that were simply beyond the ability of the then-current methodology. As discussed below, the culture of the Group at that time was more conservative and it was not considered inappropriate to decline to rate a structure if the methodology for assessing the risk presented was unsettled, or if there were questions or concerns for the expertise and experience of the asset manager or if the motivation for the structure was suspiciously unclear. It was not particularly relevant, in those days, whether or not our main competitor (S&P) was rating the deal. 5 Note that the Derivatives Group in NY was, at the time I left, the only group that I knew to retain – in general - the “two analysts per deal” staffing structure, and during the last two years of my tenure an increasing number of deals were being rated in NY by only one analyst. It is my recollection that the default for all the derivative transactions rated out of the London and Paris offices was for only one analyst to be used per deal. This shift away from two analysts per deal was, in my opinion, reflective of both the desperately under-resourced status of the Group, and the predominance of the concern for cost control (despite internally reported profit margins that were nothing short of extraordinary). Statement of Richard Michalek page 5 of 21 (6) If the principal documentation is progressing well, client's counsel will begin secondary documentation drafts, including the asset management agreement, the administration agreement (for the trustee), the swap documentation and the legal opinions; (7) Second drafts of the principal documentation are circulated along with first drafts of the secondary documentation and further conference calls are scheduled as necessary; (8) Pricing the transaction and receiving comfort from a “rating committee” (these two steps generally occur simultaneously): 1. At a point in the development of the documentation when the client is confident that no more structural changes or material changes to the forms of documentation are likely to be forthcoming, the client will want to “price” the transaction, soliciting “indications of interest” from investors by issuing – when required – a preliminary offering memorandum describing all the principal terms and conditions of the offering; 2. Because the ability to commit on price means restricting or precluding significant changes, the client wants comfort that the rating is going to be assigned; accordingly, Moody's would schedule a rating committee for the purpose of confirming that the requested rating can be issued given the now evolved terms and conditions of the offering; (9) To the extent that the pricing requires certain modifications, changes have to be modeled and tested, and documentation has to be modified and conformed; (10) Once the price has been settled on, a final rating committee is convened authorizing the issuance of the rating upon the closing of the transaction based on the final terms and conditions of the offering, the satisfactory condition of the form of closing documentation and the acquisition of the underlying pool of collateral.6 The above steps could vary in time and in order. As a general guideline, the process would have taken anywhere from 4-8 weeks for CBOs and CLOs, and in some cases, longer.7 As more and more issuances took the form of “series” deals (where each deal is identical to it's immediate predecessor, but for the underlying assets in the pool or some distinguishing feature), the timeline compressed enormously. 6 In fact, in cash-based deals (as opposed to synthetic or hybrid transactions where the “pool” consists of at least partially of a list of specified reference obligations), the collateral comprising the pool would only be partially assembled as of the closing date. The asset manager is obliged in these deals with partially completed pools to complete the accumulation (“ramp-up”) of the entire portfolio within a specified period after closing. The rating issued at closing remained subject to “confirmation” upon the completion of the prescribed “ramp-up” period. While investors were initially reluctant to close on less than 100% of the portfolio having been assembled, over time, and depending on the reputation of the asset manager, the structurer and market conditions, investors accepted both longer ramp-up periods (in some cases and for some structures up to 2 years) and ever lower percentages of completion as of closing (a general guideline of “at least 80%” could be modified to 50% or lower depending on facts and circumstances). 7 As noted, the ramp-up of the complete pool for cash based structures, depended on the ability of the asset manager to actually find and purchase – at a price consistent with the modeling assumptions used in the rating process - the bonds and loans going into the pool. If this process is extended across different market conditions, or if the ratings on the purchased assets changes, substitutions might have to be made. Too high an aggregate price for the entire pool and the desired rating may not be achieved, and without the desired rating, investors in different tranches of the capital structure may withdraw their interest. It is easy to understand why synthetic CDOs, which only “refer” to the underlying assets as if they had been purchased were so attractive, given their ability to reduce or eliminate the risk of a failed ramp-up. Statement of Richard Michalek page 6 of 21 This compression of alloted time for completion unfortunately co-incided with emergence of three critical forces that I believe contributed to perception of a degradation to the quality of ratings issued by all the credit rating agencies: (1) the shift in bargaining power away from investors (and simultaneously away from the rating agencies), (2) an overcrowding into structured finance issuance and the corresponding increase in market competition for all participants, and (3) the arrival of “synthetic” structuring technology, whereby “virtual” pools of one or more assets are assembled, and where the “buyer of credit” assumes the benefits and suffers the losses corresponding to the pool of assets identified as the “reference” obligations.8 As I describe later in this Statement, I believe that for the Derivatives Group, there was a fourth, and in my opinion, dominant force to contend with: a change in culture. I.C. Quality of the Models and Modeling Performed by the Group, and a Comment on the Sources of Rating Errors It is critical to appreciate a defining limitation to the work done by the Derivatives Group: We took the ratings already assigned to the underlying assets as our inputs. We did not “re-underwrite” any of the assets proposed for inclusion in the issuer's portfolio. Errors in the ratings assigned to the underlying assets – whether representing under-estimations of the likelihood of non-performance and/or an underestimation of the losses realized due to such nonperformance – by definition compound into magnified errors in the derivative product. Broadly considered, any such errors can be classified as “asystematic” errors – created by individual negligence, or the misrepresentation or fraud by some random number of borrowers or underwriters – the effects of which were expected to be diluted by the law of large numbers; and “systematic” errors – where whole cohorts or classes of assets would be found to represent more risk than was identified by the original underwriting and the rating. The latter type of error understandably created very significant problems for the rating of derivative products. The potential for errors and limitations in the accuracy of the assumptions underlying the methodologies employed by the Group were not ignored; variance “cushions” were included in the determination of tolerable thresholds of “expected losses” corresponding to each rating level assigned. Inevitably, however, maintaining those cushions was vulnerable to market pressures. The hypothetically “perfect” model was sometimes discussed: in theory, an investor only needs 2 ratings – Aaa, and NR, which would correspond to “get everything you are promised” and “won't get everything promised”. With infinite time and resources, all obligations might be assigned these “perfect” assessments of risk. The users (and purchasers) of ratings constrain both time and available resources, and within the competitive business model, the constraint is doubly felt. 8 In fact, the synthetic investor assumes the benefits and losses accruing to either (a) the holder owning all of the individual assets, if the investor is financing the entire capital structure for the “virtual” issuer that “holds” the pool, or (b) a particular section of the virtual issuer's capital structure, in which case that investor's losses are only triggered when and if the virtual classes beneath that particular section experience losses, and whose investment is repaid only after sufficient funds are paid out of the reference obligations such that more senior sections of the virtual issuer's capital structure receive their full entitlement. This ability to offer an investor an opportunity to invest in a single class of a derivatively created structured finance issuer's capital structure without having to actually sell the entire capital structure, and without having to physically assemble an entire pool of assets represented a game-changing jump to “hyper-drive” for structurers, and for investors seeking exposure -on either the long or short side - to particular assets and classes of debt. Statement of Richard Michalek page 7 of 21 Admittedly, there was something of an inconsistency in the logic of “adding cushions”. Assets included in an issuer's portfolio would, if rated only by a competitor, be given a rating “haircut”. For example, an asset rated BBB by S&P - but not by Moody's - was assigned, for the purpose of Moody's evaluation of the risk presented by a pool of assets, a rating lower than equivalent Moody's Baa rating. In general, this was said to address a statistically observed variance in the performance of assets with single ratings from S&P. Further, the long-standing policy of the Group was that even with these “haircut” ratings, our own methodology and corresponding modeling would lose its robustness and reliability if more than a certain percentage of the assets lacked a Moody's rating.9 This practice inevitably acted to increase the use of Moody's ratings on the underlying assets. Structurers would assemble pools of assets that already had Moody's ratings, thereby placing a premium on their marketability, and consequently, placing a discount on the marketability of assets lacking a Moody's rating. Incidentally, this practice may in some part contribute to an alleged “race to the bottom” for any rating agency seeking to obtain or increase market share. If structurers (and investors) discounted the value of the asset that wasn't rated by Moody's or S&P (historically, the two main sources of ratings for most of the assets eligible for inclusion in CDOs), the “other agency” might have to offer something else to gain market share. Given the high fixed costs of the intellectual capital necessary to offer ratings on sophisticated complex derivatives, there are those that claim that the “something else” was a rating that did not fully represent the risk of the rated asset.10 While we did not unquestionably accept assets with ratings solely from other agencies, we did unquestionably accept the Moody's ratings on the assets. We continued to accept these ratings even after it was known (or should have been known) that due to increased deal flow and a lack of resources, some of the outstanding Moody's ratings on RMBS assets were extremely stale, and therefore less reliable. During the initial research in [2000 and 2001] surrounding the preparation of the methodology for rating “Multi-sector CDOs” (in which multiple types of assets were to be included in the pool of assets backing the obligations of the derivative CDO), I learned from the then-managing director of the RMBS group that many tranches were not being monitored or reviewed as often as they would like. Due to limited resources at the time (the RMBS transactions used only one analyst for both document review and model analysis, and each analyst was responsible for monitoring the performance of their own transactions and any “inherited” transactions), it was possible that some of the tranches carried ratings that had not be reviewed for more than a year and in some cases for more than two 9 To be clear, it is my understanding that Moody's competitors employed equivalent haircuts and limitations. 10 Note that any such “hidden” risk works to the benefit of the risk-seller i.e., the structuring bank. To the extent that the published rating(s) do not accurately represent the risk presented to the buyer of the asset, the seller of such assets can obtain economic rents. As the “riskiest” part of the capital structure of structured finance obligations in general, and of CDO products in particular, is often retained by the structurer until it can be sold or repackaged into another derivative product, it would be logically beneficial to the seller of any such asset if the rating assigned did not fully represent the “hidden” risk. Statement of Richard Michalek page 8 of 21 years. It was possible – I was told - that the last time a particular RMBS deal's performance was carefully reviewed – might have been when the initial rating was issued.11 It is my understanding that it wasn't until 2007 and the “mass downgrades” to RMBS backed primarily by sub-prime mortgages that Moody's CDO group publicly acknowledged that, given the sensitivity to the underlying assets and their ratings, ratings on the derivative products rated by our group would be “adjusted” in anticipation of - and prior to - wholesale changes in the Moody's ratings that were assigned to the assets in these CDOs. In effect, Moody's admitted that our own RMBS ratings were “temporarily” untrustworthy.12 To be clear, the RMBS transactions in question “belonged to particular cohorts” known to include a disproportionate amount of sub-prime and/or second lien obligations, and in many cases, had already been placed on “negative credit watch” (signaling to the market that a downward rating action was likely in the foreseeable future). But the CDO deal documentation's prescription for “adjusting” the ratings of assets on negative credit watch was admitted to under-represent the actual risk from those assets. While the deal documentation may have required the asset manager to “adjust” rating on the negative credit-watched asset by 1, 2 or 3 rating levels or “notches”, the anticipatory adjustment in the summer of 2007 indicated that such deals would immediately require – until such time as the mass downgrades were published – significantly more notches in adjustment.13 II. Culture, and Why It Matters II.A An Early History of the Group In 1999, when I was hired into the Derivatives Group at Moody's, there were only seven (7) lawyers. There was also a managing director who was a lawyer, but her duties limited her direct involvement in analyzing transactions. Of those seven, I was the fourth attorney with Skadden Arps on their resume. There was a strong emphasis on the quality of the work, and I found this reassuring and consistent with the priority given to quality both at Skadden and at my New Zealand employer. I chose Moody's as a career opportunity both because the subject matter of the work fit well with my own experience and interest, but also because of the importance the Group reported giving to “quality of life” and the “work-life” balance. Each of the “battle tested” former Skadden attorneys assured me that, while the compensation package was inferior to that being offered by first tier law 11 Keep in mind that for a single RMBS transaction, the capital structure financing such a transaction could be sold by issuing any number of classes of notes with potentially different entitlements and representing exposure to different “tranches” of the issuer's capital structure. Some individual RMBS transactions were issued with a dozen or more different rated classes each representing a potentially different position in the capital structure and each representing a potentially different entitlement (i.e., interest only, inverse interest, principal only, etc.) 12 This occurred after, if I recall correctly, a similar action by S&P. 13 By Moody's own admission, the variance in the deals and structures with exposure to the sub-prime and home equity (second lien) obligations made it an administrative impossibility – given economically available resources – to timely review each deal with exposure and determine the most appropriate anticipatory rating adjustment. Instead, a conservative, and hopefully sufficient temporary adjustment would, while possibly over-stressing many of the structures relative to the actual consequence of the downgrades when realized, create the least avoidable harm. See generally, The Impact of Subprime Residential Mortgage-Backed Securities on Moody'-Rated Structured Finance CDOs” a Preliminary Review; John Park, March 23, 2007. Statement of Richard Michalek page 9 of 21 firms, the work experience did not include the extreme stress and deadline pressures faced by mid-level and senior associates at any of those firms.14 In those early years, the relatively conservative firm culture – perhaps a by-product of Moody's being a subsidiary of Dun and Bradstreet until spun off in late 2000 – and the intellectual bias to Group's culture largely reinforced my own personal experiences and expectations. During the time I was advising investment banks and issuers as an employee at the law firms, I had numerous opportunities to interact directly with Moody's rating analysts. In my experience, the typical Moody's analyst was very thorough, specific and prepared, ready with comments at the scheduled time, with insightful and probing questions belying a full understanding of the deal's structure and documentation. I recall a particular transaction from 1996 while I was still at Skadden: a mortgage securitization sponsored by an Indiana bank. The Moody's analyst was comprehensive in his comments, and as the lead associate I felt the stress of satisfying his requests. I recall the supervising Skadden partner complaining to the analyst – at the top of his voice during an 11p.m. conference call on the night before closing – that the requested changes would be impossible to make without causing the closing to be delayed. The Moody's analyst's requests, based on my perspective as the associate responsible for drafting and modifying the documentation, were correct and even necessary (even if they were arriving at a difficult time) in order to accurately represent the structurer's recently modified cashflows being offered to the investors.15 It was not surprising that the culture I entered in 1999 was of a similar nature to the culture I indirectly experienced when opposite a Moody's rating analyst on deals being structured by my law firm clients. When I joined Moody's in 1999, I interviewed with each lawyer in the Group with whom I did not already have a former professional relationship. I learned after I was hired that it was the Group's practice to give every lawyer a chance to review every candidate seeking a position as a lawyer in the Group. It was also the practice to only consider candidates who had a minimum of 4 years postlaw school work experience, and preferably that experience would include transactional experience in structured finance.16 Continuity in the Group's rigorous and independent culture was implicit, and the hiring practices reflected the importance management gave to that continuity. The Derivatives Group that I joined in 1999 operated much less formally than the Group in existence in 2007. Policies and practices were still in the process of being invented as nearly every 14 As events unfolded, that enticing work-life balance largely disappeared. While I did not re-experience the consecutive “all-nighters” which had been something of a common practice in Skadden's structured finance department in the mid1990s, I recall only too clearly participating in deal conference calls and faxing comments to second and third drafts of deal documentation from quite a number of vacation locations and on any number of weekends. 15 Standard and Poor's rating analysts were no doubt equally competent, but would – in my experience as a transactional lawyer representing structurers and issuers – require less time for addressing their comments on the draft documentation. In part this was because (I later learned) the S&P analysts were often not responsible for the legal sufficiency of the documentation, as that responsibility was handled by their outside counsel. On that same Indiana bank transaction, S&P's analyst did not raise the issues identified by the Moody's analyst; but neither did they complain when the corresponding last minute changes went into the deal documentation. 16 Lawyers, unlike quantitative analysts, entered the Group at the level of vice-president. It is an indication of the extent of the change in culture and the pressures of business competition felt by management that by the time I left in 2007, lawyers were being hired at the level of “assistant vice president” and with significantly less work experience after being interviewed unilaterally by managers with only limited, if any, input from other lawyers in the Group. Statement of Richard Michalek page 10 of 21 product being assigned a rating by the Group was new. Much of the knowledge required for effectively working within the Group came from an informal “mentoring” that took place during the process of rating transactions. A more experienced quantitative analyst would be assigned to a transaction with an inexperienced lawyer in order to distribute the knowledge and expertise, even while producing ratings revenue.17 In addition, it was expected and encouraged that newer analysts would refer to and rely on the guidance and advice of the more senior analysts with respect to any questions they might have during the rating of any transaction. And of course, further questions would come up (accelerating the climb up the learning curve) during every rating committee. The commonly agreed upon “mentors” available to everyone in the Group included: Arturo Cifuentes, a PhD credentialed quantitative analyst who was as outspoken and demanding as he was intelligent, and who – it was apocryphally told to me – once called up a client's counsel who (probably in the interest of time) had sent Arturo a revised draft of a key transaction document without “blacklining” it to show the changes. Arturo politely told the counsel “Thank you for the updated draft. I notice that it is not blacklined. Are you listening? The next sound you will hear is that draft hitting my trash can. I look forward to receiving a blacklined draft on Monday.” Nevertheless, he was incredibly generous with his time and instruction to any new analyst who asked questions. I believe it was Arturo who co-authored the first methodology pieces explaining the Moody's use of the “binomial expansion theorem” and the importance of diversity when modeling CBOs. Isaac Efrat, a math PhD, who was incredibly fast at modeling the complex waterfalls embedded in the legal documentation. At one point it became necessary to come up with a “path dependent” simulation analysis of default risk for certain types of transactions in order to fully assess the risk; his model became the critical tool and remained in use long after he had left Moody's. Danielle Nazarian, an MIT graduate whose eye for detail bordered on the maniacal. As a quantitative analyst, Danielle had no peers in terms of the thoroughness of her reviews and the extent to which she would require exacting fidelity between what she was modeling and what the documents described; she was an inspiration to every carefully-minded lawyer.18 Stephen Lioce and Mark Froeba, both Skadden alumni, offered the deepest resources for new lawyers coming up to speed with the methodology and practices employed by the Group. Mr Lioce in particular set an important standard in terms of detailed and thorough reviews of documentation and his record-keeping and precedent library on more than one occasion helped to break an occasional impasse presented by an aggressive banker insisting on (or resisting) a particular provision. Of the aforementioned individuals, all but Ms. Nazarian and Mr Lioce have moved on. It is my understanding that Mr Lioce no longer actively assigns ratings to new transactions, but instead manages the surveillance group responsible for monitoring transactions. As for these other mentors, 17 The number of deals being rated by the Group in 1999 and 2000 (between 45 and 60 deals per year) permitted this somewhat unsystematic transfer of intellectual capital. 18 Ironically, the extent of her attention to detail may have indirectly encouraged some lawyers to “leave the detail to Danielle”, and it was no accident that she was one of the first quantitative analysts allowed to rate a transaction without having a lawyer formally assigned to the deal. Statement of Richard Michalek page 11 of 21 with the change in culture, they were either “encouraged” to move on, or were expressly asked to leave. As the culture and the pressures of work increased with the increased deal flow between 2003 and 2007, it was clear that survival – in a culture that favored and was dominated by quantity over quality – required modifying one's approach, even if it meant lowering one's standards. II.B Post-”Spinoff”, 2002 and Beyond – The Inexorable Rise of Brian Clarkson From Group Managing Director to the President of Moody's Investors Services Without question, after the spin-off of Moody's from its former parent (the Dun and Bradstreet organization), Brian Clarkson was a “rising star” in the ranks of senior management at Moody's. His reputation for aggressively re-organizing the group responsible for rating residential mortgage backed deals (and later the entire asset-backed group) had successfully garnered the attention and support of his managers, and the top tier of the senior management of the company. At my level, any “water cooler” discussion of his management style included the words “fear and intimidation.” That description was periodically reinforced by the terminations of one or another of his managing director reports who had in some way failed to fulfill the express or implied expectations. The effect was only the more chilling when he was heard to say “its not personal, its only business.” After the spin-off from DNB, Brian Clarkson assumed control of the restructured Structured Finance division, and Daniel Curry was moved to London - ostensibly in recognition of his apparent success in managing the Derivatives Group - to manage the nascent growth in structured finance in Europe. Noel Kirnon was brought back to New York and assumed Dan's responsibilities as Group Managing Director of Derivatives. Amongst the analysts of the Derivatives Group, a permanent anxiety quickly took hold: that Brian was going to “fire all the lawyers”. To my knowledge, all the other groups that may have originally used two analysts per deal had reduced or reorganized their staffing such that only one analyst was assigned per transaction. I was frequently reassured by my managing directors that they would resist moving to a 'one analyst per deal' policy, in part because the transactions were by definition complex and novel. Nevertheless, it was hardly reassuring that soon after the initiation of rating CDOs in the London and Paris offices, complex structured finance transactions (including cash and synthetic CDOs) were being rated by only one analyst, referring all legal questions to a single (thinly stretched) lawyer based in London.19 While it was quickly apparent that Mr. Clarkson subscribed to the school of “management by fear and intimidation”, I think he also wanted to change the Derivatives Group into something he was more comfortable managing. During late summer in 1999 and 2000, the Derivatives Group – under the direction of then Group Managing Director Daniel Curry – took two of its last annual “retreats” at which a few analysts would be invited to present draft research papers to the rest of the group, which at that time consisted of between 20 and 30 employees. These annual retreats were – in my experience the most formal steps taken by the Group at that time to record rating practices and policies, and they gave everyone a chance to exchange information about common issues arising in the current deals 19 It is important to keep in mind that the task faced by a quantitative analyst is formidable, even without including a detailed review of the transaction documentation. Considering that without a lawyer assisting them on the transaction, a relatively inexperienced (albeit talented at modeling) Moody's quantitative analyst would, in the event they raised a question of potential importance to the client, be suddenly on the phone with either a partner at a major law firm, or an associate from that firm with deep experience on the provisions of the documentation, it is little wonder that the number of drafting comments independently raised by quantitative analysts was low, and in some cases, zero. It is my understanding that the lawyer servicing the many “single analyst” deals in the London and Paris offices was primarily focused on keeping the deals moving and chose very carefully what (if any) issues to raise. Statement of Richard Michalek page 12 of 21 being rated. The retreats consisted of a Friday afternoon exodus to a hotel or conference center located outside New York, a Group dinner, a working Saturday where papers would be discussed, a break for lunch, and then recreational activities of your own choosing in the afternoon. As a new employee, these events created a strong impression of common mutual respect, and a very flat management style. They were also refreshingly devoid of “team building” exercises so often ridiculed by the popular media. II.C Revising and Reversing the Culture Whether as a function of cost-cutting, or style, after Brian Clarkson's appointment as the head of Structured Finance, the Derivatives Group discontinued “separate” retreats, and instead an annual “one day” retreat for the entire Structured Finance group was installed, with a venue local to the office.20 With the entire Structured Finance group in attendance, the structure of the day was by necessity designed around space limitations. Outside speakers and senior management would be invited to present to the assembled throng, and after lunch there would be recreational activities – including those proverbial “team building exercises”.21 The Customer Satisfaction Interviews I believe that Brian Clarkson was determined to make his mark on the Derivatives Group when he assumed control. I think his primary goal was to reverse the impression that Moody's Structured Finance analysts and in particular the analysts in the Derivatives Group were “arrogant” or insufficiently accommodative to the needs of our “customers”, i.e., the investment banks. After personally traveling with senior managers of the Derivative Group to all the New York offices of the Group's major clients (those that provided the most business), to introduce himself as the head of the newly organized Structured Finance division and to hear the concerns and “wish lists” of those customers, he summoned – one by one – the lawyers in the Derivatives Group to his office for a “discussion”. In my “discussion”, I was told that he had met with the investment banks to learn how our Group was working with the various clients and whether there were any analysts who were either particularly difficult or particularly valuable. I was named along with Stephen Lioce as two of the more “difficult” analysts who had a reputation for making “too many” comments on the deal documentation. The conversation was quite uncomfortable, and it didn't improve when he described how he had previously had to fire [MD X], a former leader of the Asset-Backed group who he otherwise considered a “good guy.” He described how, because of the numerous complaints he had received about MD X's extreme conservatism, rigidity and insensitivity to client perspective, he was left with no choice. He further emphasized that as difficult as it was, he quickly received “dozens” of calls from clients 20 In part, I believe this was an effort by senior management to bring the Derivatives Group into the “fold”; the Group had developed a reputation for operating somewhat autonomously, and the extraordinary fees being generated by rating complex instruments may have represented something of a “prize” to the individuals most able to establish their authority over the Group. 21 In further contrast to our somewhat staid Derivative Group retreats, these day retreats would conclude with dinner for something close to 300 followed by some drinking and dancing. The night would include some predictably embarrassing performance or activity presumably meant to show the more human side of management. Whether a carnival style “MDdunking” achieved by throwing baseballs at a bulls-eye, senior management performing (better than expected!) as ersatz Blues Brothers, or a somewhat tasteless “Sumo Wrestling tournament”, where two MD's were conscripted to face off wearing rubber fat suits and Sumo wigs, the culture being shaped was new to the Derivatives Group, and distinctly antiintellectual. Statement of Richard Michalek page 13 of 21 thanking him for finally addressing the problem. Ironically, MD X later headed the structured finance research at a large foreign bank. He then asked me to convince him why he shouldn't fire me. In my case, the conversation did have apparent consequences as I received a reduced bonus in that year, and my eventual promotion to Senior Credit Officer was somewhat delayed. More importantly, and as presumably intended, I thereafter attempted to carefully monitor, and when possible drastically reduce, the number of comments I made, and I took to heart the implicit and explicit message that “the customer comes first”. Further, my change in behavior was an important signal to the other analysts in the Group, reinforcing – and in my opinion, helping to accelerate – the change in the Group's culture. Brian closed our conversation by explaining that “we have an incredibly difficult job” because we have to find a way to provide the issuers and banks with a rating they can accept, while insuring that we do not compromise our standards. However, the primary message of the conversation was plain: further complaints from the “customers” would very likely abruptly end my career at Moody's. The Importance of “Precedent” As the deal flow began to slowly increase after the 2001 recession, the role and use of “precedent” - relying on a prior Moody's-rated transaction as a “template” for the instant transaction being reviewed – steadily became more important. A typical structured finance transaction could include anywhere from 40-60 documents. The critical documents for a Derivatives Group transaction (CBOs, CLOs, and CDOs) having the greatest impact on the ongoing operation of the deal were the “Indenture” (which would typically run from 120280 pages, and which included the critical “cash flow waterfall” and the obligations of the indenture trustee), the Collateral Management Agreement (sometimes called the “Asset Management Agreement” and which prescribed the duties and obligations of the agent responsible for asset selection, acquisition and disposition), and the “swap” documents, which (prior to synthetic transactions) pertained to the interest rate hedging critical to the transactions. Each law firm used it's own “precedent forms”. (The use of standard initial draft forms by the law firms ensured efficient allocation of their resources, enabling a second or third year associate to effectively prepare draft documentation without immediate partner involvement.) Even if two Derivative Group transactions were substantively identical, the use of different “forms” would drastically increase the amount of time necessary for reviewing the transaction. As available time became scarce, it became essential to ask the counsel drafting the documentation to “mark it against a recent Moody's rated transaction”. However, this request proved to be the starting gun in what was an accelerating race to the bottom. The “Customer Satisfaction” interviews had more than accomplished their principal goal: the lawyers of the Group all retreated drastically from the “fine tooth comb” approach to document review, and customer preferences were given the highest priority. I took what I considered to be a middle approach: I would review the first presentation of a form of document or novel structure as carefully as I could given my heightened awareness for a client Statement of Richard Michalek page 14 of 21 being able to call up my boss and having me upbraided (or worse) for slowing down the transaction over “commas”. Then, for every subsequent similar deal from that law firm, I would ask that counsel to “please mark it against precedent deal X”. In this manner, I could reduce the review of the documents to simply reviewing the pages that included changes to the form that I was confident was within acceptable parameters (from the legal perspective). Unfortunately, there were some lawyers in the Group, and one in particular, who read a different message from the handwriting on the wall: if less is more, then even “less than less” must be worth more still. It wasn't long before the investment banks learned, from direct experience or from their deal counsel, that Moody's lawyer X gave a very “cursory” review to the CDO transaction documentation, and accordingly features more favorable to the risk seller (the investment bank and their clients) could be included without jeopardizing either the timing of the deal, or the rating assigned to the deal. Lawyer X soon was named as a preferred analyst by an ever-increasing number of clients seeking a rating. Lawyer X quickly developed a reputation among the quantitative analysts and the other lawyers in the Group (particularly those who suffered having to deliver comments on a form of document that had been recently “signed off” by Lawyer X ) for skipping the review of large sections of the documentation, if not the entire document. How contagious this behavior was is hard to say. For awhile, the lawyers in the Group were divided into two camps – those that would (quietly) “fight the good fight” and attempt to restrain the more aggressive drafting changes being installed in the forms of documentation, and those who, like Lawyer X, recognized that without management support, it was rational to keep the client happy, and – if any issue was to be raised, simply focus on those issues that would attract the attention and support of the non-legally trained managing directors.22 The consequence that proved to be the final push down a very slippery slope was this: deals would be marked against Moody's-rated precedents chosen not by Moody's, but by the law firm and the investment bank. I found myself caught in a politically untenable situation: trying to explain why a Moody's precedent, very recently rated by Moody's (and increasingly with documents last reviewed by Lawyer X), might not be the best choice of precedent for the instant transaction. Goldman Sachs was well known by the lawyers in the Group for consistently producing as their 22 It is difficult to appreciate just how low the bar was being set. One example comes from a standard “termination for cause” provision found in all Asset Management Agreements. The agreement provided for removal of the Asset Manager by the investors if some level of poor behavior can be shown. After extended internal discussion, the general consensus among the lawyers of the Group was that the strong form of the provision – where “cause” could be found in case of simple negligence by the manager – was appropriate, particularly because manager performance was so critical to maintaining the fidelity between the highly specific modeled parameters of the document and the estimated risk represented by the rating. Convincing the client to accept that standard often required extensive negotiation, and many investment banks and managers refused to accept the standard “unless everyone accepts it”. Inevitably, a banker would find a recent precedent containing the weaker standard (“gross negligence”), and that precedent would more often than not turn out to have been rated by Lawyer X. Convincing your manager (a non-lawyer) of the importance of the single word “gross” often required a second, more delicate, negotiation. It did not help that the single MD in the Group who was a lawyer had already conceded the battle, recognizing before the rest of us that with Mr. Clarkson's ascension, “the customer is always right” effectively meant having first to prove that the requested rating couldn't be assigned if you wanted to say “no”. Statement of Richard Michalek page 15 of 21 “preferred form of document” the most “risk seller friendly” precedent, even if it had been drafted by a law firm other than the firm working for Goldman at the time.23 Rotating the Analysts This growing importance of precedent was roughly consistent in time with a decision by management to avoid or limit the risk of “captive” analysts. Described as a way to always ensure that “fresh eyes” were reviewing transaction documents, the managers assigned the individual lawyers to certain banks – and removed us from working on transactions with certain other banks. The plan was that after some time, there would be a further rotation.24 It is my belief that the initial “assignments” in the “rotation” scheme were primarily an attempt by the managers to address Clarkson's mission to improve the perception of Moody's as a more accommodative agency. The rotations were likely a cover in order to re-assign the “difficult” lawyers onto the deals for clients that had not previously complained about that particular lawyer. During my tenure at Moody's, I was explicitly told that I was “not welcome” on deals structured by certain banks, however, I never saw a 'list'. I was told by my then-current managing director in 2001 that I was “asked to be replaced” on future deals by Chris Ricciardi, who headed the structured group at CSFB, and then at Merrill Lynch. Years later, I was told by a different managing director that a cashflow CDO team leader at Goldman Sachs also asked, while praising the thoroughness of my work, that after four transactions he would prefer another lawyer be given an opportunity to work on his deals. I also understand that Lehman Brothers similarly requested my “rotation”, as did UBS. When Mr Ricciardi left CSFB to head the structured group at Merrill Lynch, I was co-incidentally “reassigned” off of a Merrill Lynch cashflow deal on which work had not yet begun, even though I had never worked on a Merrill CDO before. In each case, I never was assigned to that bank's cashflow CDO deals again. As painful as these “removals” were to experience, the bigger picture may offer useful information.25 For example, using basic statistical techniques, in either late 2002 or early 2003, in an attempt to determine for myself whether working as I did was actually adding any value to the ratings, I compared the rating migration of the deals I worked on relative to the deals on which Lawyer X was originally assigned, and found that Lawyer X's deals performed markedly worse. Whether this was due to my initial review of the documentation, luck, or to the adverse selection of Analyst X by those banks who were most aggressive in their structures and their negotiation of documentation, I cannot say. I 23 While Goldman Sachs was not the only investment bank that used the practice of rotating law firms, in part to gain access to the broadest selection of precedent documentation and thereby the greatest potential for finding a precedent that supported Goldman's preferred language, they were the only bank I knew of that employed someone whose primary job was – to put it politely – arbitrage the rating agencies. It was not difficult to know where Moody's stood in terms of the relative conservatism of our modeling assumptions and drafting requests; Goldman was very prompt when informing us that “S&P doesn't require that”. 24 Whether due to “higher priorities in light of new developments” or because it was never intended to be fulfilled, the second step in this plan was never fully implemented, and no further organized rotation was effected. Lawyers were once again assigned as and when available to deals as they came in. With one important caveat: if a client requested a particular lawyer, or refused to accept a particular lawyer, those requests would be honored if at all possible. 25 The request for removal very rarely included the (true) allegation that I was requiring standards that other lawyers did not, but would instead allege that my manner or style as “abrasive” or “aggressive”. Interestingly, those same manner and style complaints never seemed to be forthcoming from those banks who were less aggressive in their structures and whose documents already contained standards equivalent to what I was requesting. Statement of Richard Michalek page 16 of 21 also do not know if any formal analysis has ever been done by the company on the relative importance to rating stability due to the different rating analysts on the ratings assigned. In fact, this internal conflict dominated my – and presumably other lawyers - annual performance reviews: praising “the attention to detail” but admonishing for not “thinking of the big picture” and “losing sight of the [importance of Moody's] relationship with the banker and collateral manager”. The not so subtle instruction received in a performance review from 2001 that I should “think of a particular deal as one piece of a larger relationship (the issuer relationship)” is representative of the explicit priority management assigned to the relationship between Moody's and its customers (the investment banks), notwithstanding the possible impact to the investors in “a particular deal”. By 2004 and 2005, the culture of the Group from 1999 and 2000 and been largely replaced, and the new hires added to address the demands of an ever increasing deal flow were joining the Group at a time when one eye was constantly on the (ever-rising) Moody's share price, and the other was calculating the number of deals necessary to reach “maximum bonus”. Suggested improvements for the robustness of our processes and ratings integrity were subjected to a first order question: will they help make the workload any easier? III. Business Competition & Market Pressures The pressure from competition and the predominance of the importance of quarterly earnings increased in proportion to the growth in the derivative market. As more and more banks identified the extraordinary opportunity for profits in securitizing and re-selling structured products, the effects of competition were felt by all. Structured finance was, in my first year at Skadden in 1993 and until the time I left in 1996, understaffed but extremely busy. Partners would bill 2500 hours and up, and associates who wanted to make partner could expect to bill more hours than that. Skadden held a strong market position. It was difficult to attract junior associates into the group, however, because of the impression that the lifestyle was too grueling, and the appeal of the transactions limited. From 2003, structured finance activity, and Moody's rating activity kept on growing right until the cliff effect seen in 2007. With every increase in issuance there was a concordant increase to the number and size of the structuring departments at the banks. And with every such increase, there was automatic need for increasing deal flow to service that new investment in staff. Margins for the banks remained healthy for the dominant issuers, but the increased competition and the refinement of executing transactions “synthetically” pushed complex derivatives down the same path as RMBS and ABS issuance: towards “commoditization”.26 26 There were several natural constraints on the growth rate of cash-based CDOs. The most obvious being the finite number of assets available to use as collateral. Once the last of Bond X has been put into CBO #1, a manager can't find any more Bond X to put in CBO #2. Also, acquiring the actual assets for a portfolio that will back the obligations of CDO “A” requires time and exposes the manager and investors to “ramp-up risk” (as previously discussed). With the refinement of the legal documents necessary for synthetic transactions, both these risks were avoided. (No more Bond X on the market? Doesn't matter; can simply “refer to Bond X”. And the “ramp-up” for a synthetic CDO consists only of writing down the reference obligations that will serve as the “reference pool”.) Statement of Richard Michalek page 17 of 21 Commoditization – where deals are produced on “standard” documentation, and turnaround time becomes shorter and shorter, was an effort to preserve profit margins in an environment of extreme competition. Our customers who were looking to increase market share and/or preserve profit margins in a competitive commoditized environment had few choices. They could trim expenses (placing pressure on the vendors in the deal – giving rise to “capped” legal fees, and “shopped” ratings) or pursue “yielddriven” investors. Alternatively, some customers recognized that finding a way to service the increasingly commoditized product could be a high value-added endeavor, and those customers chose to pursue new and novel structures and instruments. All choices worked to the disadvantage of the rating agencies and served to “chill” the effort of careful legal analysis. For example, if the deal counsel was operating under a “capped” fee, they would complain to their client if the work and time on the deal was extended due to “unnecessarily detailed” rating agency reviews being performed on what was increasingly a standardized product. These same “capped fee” firms would also push the work down to the 'least cost provider' resulting in some very junior associates being tasked with managing and producing the documentation for the “nth” deal in a series of deals from the same investment bank. Unfortunately, as a deal was repeated, it would sometimes need to be revised or amended to remain competitive in an increasingly competitive environment. What was good enough for the investors in deal #1 was no longer enough in deal #5. Exactly when more attention needed to be paid to the marginal changes inserted by the structuring bank, less resource (or capable resource) was being allocated to the series. In practice, this often led to longer than usual conference calls as even harmless drafting comments had to be explained to an anxious and defensive junior legal associate whose mandate very nearly was “resist all changes”. If the client – who was likely deaf to the reasoning for the change, but highly sensitive to the legal time (cost) consumed by discussing such change – complained to Moody's management, it reflected badly on the analyst's record. Practical consequence: reduce or simply stop making the comments. Second source of the “chilling effect” from commoditization: If the rating was being shopped, pressure was indirectly applied on the rating methodologies employed. Ironically, any concession achieved by favorably modifying an assumption in the methodology implicitly increased the pressure to be careful on the legal drafting side. If the source of any rating inaccuracy – given the same quality of underlying assets – can be traced to either (1) the accuracy of risk estimated by the model or (2) the fidelity between the documentation and what was in fact modeled, any deterioration in one places more of the onus for a correct rating on the other. Unfortunately, the business environment, and Moody's changed culture, were not particularly receptive or supportive to any legal analyst who was trying to carry that burden of compensating for the overly generous adjustments to the methodology. And finally, the novel structures and instruments that arrived in the midst of an accelerating deal flow offered their own source of competitive pressure. It was always important that Moody's be on the “first impression” deals – at least to the extent there was sufficient confidence that our existing methodology could be adapted to the new structure or that any gap in such methodology not present an Statement of Richard Michalek page 18 of 21 intolerable risk to our “standards”. After that first deal of its kind is rated, subsequent structurers will – when attempting to replicate the structure and create their own product – often use the rating agency that was on the first impression deal. Since these novel structures and deals were by definition high value added, they attracted a generous ratings fee. Thus, the new structure and new instrument represented a future cashflow to the rating agency that successfully rates that structure/instrument. The managers in the Group were understandably eager to land that future cashflow (particularly in light of deferred compensation expressly tied to the improved share price). To the extent a proposed novel structure was identified as containing significant legal risks, the acceptance of those risks became a “business decision” that would require senior management review and analysis, and could potentially delay the rating and cost Moody's and the Group the chance to be “on the first deal”.27 For previously presented structures that offered legal or documentation “challenges”, the approach preferred by management was first seek out sufficient “mitigating factors” that could subjectively be thought to “balance out” the identified risk or challenge, and failing that, to “temporarily” address the potential potential problem by conditioning the rating on the issuer's or structurer's agreement not to consider this deal as a precedent.28 Market Share Pressures The rating agencies had consistently lost bargaining power during the growth of the structured finance issuance, and given the incredible rise in the share price of Moody's stock (largely due to the enormous increases in deal flow), it is hardly surprising that a focus on “maintaining market share” and preserving customer relations and loyalty coincided with the increased risk of losing deferred compensation for senior management, should the share price reverse course. Initially, Moody's and S&P completely dominated the derivatives market. I was told by Brian Clarkson during our “discussion” that the Derivatives group's “market share” was between 92 and 96% of the rated derivative transactions, and that he would like to see it stay at or above 95%. One source of this extraordinary participation was the underwriting requirement – imposed in some cases I believe by capital requirements and in other circumstances by the credit committees of the structuring banks, monoline guarantors, issuers and portfolio investment credit managers – that restricted investment to assets having at least two ratings from a “NSRSO” (Nationally Recognized Statistical Rating Organization). Combine that requirement for a rating from two NSRSOs, with the onerous consequences of assembling a pool of assets that would suffer “rating adjustments” for lacking a Moody's rating (discussed above), and you can see a strong contribution to the market dominance. 27 This created something of a dilemma for the lawyer who might point out potential legal risks; by remaining silent or at least waiting until another lawyer confirmed that the issue required addressing, the assigned lawyer could avoid risking a “false positive” and stay clear of being considered an obstacle to the rating. However, if the perceived risk turned out to be real, failing to bring it to committee and the attention of the MDs was potentially equally damaging to one's career. 28 This technique consisted of extracting an oral promise from the banker that the concession would be a “one-off” and that Moody's was free to establish a contrary or adverse policy that could possibly result in an adverse consequence to the instant transaction. As the deal flow raced faster and farther ahead of our ability to update and revise and adapt methodology, these “one-offs” became anything but. And despite strong assertions that such deals would never be used as precedent, in a surprisingly short period of time the issuer would return, seeking to do deal #2 “exactly like deal #1” and frequently the “concession” would be grandfathered into all the progeny from that precedent deal. Statement of Richard Michalek page 19 of 21 While that initial domination allowed for a certain amount of control and “market power” to reside with the rating agencies, once the “market model” (controlling market share, competing on both features and price, even if to the detriment of reputation) was fully embraced as the key element of the business plan, the fear of losing market share, and suffering an earnings disappointment in the following quarter (which of course would seriously impact the compensation packages of senior management), would necessarily become the defining principle. It was not long before the structuring investment banks and underwriters recognized that just as Fitch was forced by the threat of limited or reduced business to find ways to make their ratings sufficiently more attractive to justify not going with Moody's and S&P as the “two required ratings”, the same threat (of freezing out one agency and going with the “other two”) could be – and was applied to each of the agencies. I believe that the extraordinary 95% market share was created by the initial dealer presumption that (a) on very nearly 100% of the deals the rating will be by Moody's or S&P, and (b) on some significant percentage of the deals the rating will be by Moody's and S&P – especially if the product is new and investors are going to need assurance, and (c) on that some small percentage of deals that are more commonly understood and on which the investor discount for a Fitch rating alone was justified by the reduced costs of not having to use Moody's, the rating would be by S&P and Fitch, or Fitch alone. As derivative products evolved, and as market acceptance of a Fitch rating increased (in part the result of public complaints from Fitch regarding unfair competitive practices and the allegations of a “duopoly” by Moody's and S&P), the full effects of a “competitive landscape” were felt at both major rating agencies, and dealers more and more often challenged the presumption that every deal needed at least Moody's or S&P, or that investors would not accept a Fitch rating alone. This thesis – that competition and the inevitable commoditization will lead to an erosion of market share dominance - underscores the motivation and urgency behind the accommodative adjustments to ratings policies and methodologies, and, as discussed above, the internal pressure to revise the culture and become more “customer oriented.” With the emerging dominance of synthetic transactions (which, incidentally, attracted a rating fee that was heavily negotiated by the structuring banks and which nevertheless entailed an equivalent or near equivalent demand on analytic and legal resources), and with an over-riding culture of cost-containment, the consequence of steadily deteriorating quality is hardly unexpected.29 IV. Prescribing a “Better” Credit Rating Agency In my opinion, any legislative effort that attempts to improve the reliability and utility of the products of the credit rating agencies must begin by recognizing the limitations of a market based business model. Competition for fees, and the “efficient” allocation of scarce resources present formidable challenges if not outright obstacles to achieving something akin to the “safety and soundness” that should be associated with the practice of using an independent agency for estimating the risk presented by capital investment. 29 When I joined the Group in 1999 there were seven lawyers and the Group rated something on the order of 40 – 60 transactions annually. In 2006, the Group rated over 600 transactions, using the resources of approximately 12 lawyers. The hyper-growth years from the second half of 2004 through 2006 represented a steady and constant adjustment to the amount of time that could be allotted to any particular deal's analysis, and with that adjustment, a constant re-ordering of the priority assigned to the issues to be raised at rating Committees. Statement of Richard Michalek page 20 of 21 To be effective, the motivation and implementation of any such legislation will do well to recognize that meaningful30 ratings represent a public good, and as such some part of the function of a publicly “approved” Credit Rating Agency should acknowledge and accept the responsibility – and liability - for providing that public good. To the extent that any such “approval” is conditioned on providing ratings within legal safe harbors – hypothetically, by employing methodologies that have achieved an objectively agreed consensus as to their adequacy for measuring and describing the risk that the rating purports to measure and describe – such legislation could represent an important advance in contributing to the utility of the credit rating agencies.31 Remaining within an “issuer pays” business framework without addressing the significant and ever increasing time and resource demands imposed on rating agencies dependent on market share and marginal profits derived from processes that, in my experience in the Derivatives Group, were by definition incapable of seriously capturing any value from “commoditization,” will very possibly result in repeating the mistakes of the recent past and re-experiencing the costs of such mistakes. It is my opinion that successful legislation will include an incentive, and proper opportunity, for a publicly approved rating agency to “just say no” to market participants seeking an independent assignment of something as valuable as a marketable published rating. 32 Thank you. 30 The rating agencies are careful to narrowly limit the legal definition of the ratings that are issued, even if those legal definitions are at some odds from the commonly expected or generally anticipated meaning. For example, the measure of expected loss is a measure valid as at the time of measurement. The fact that a rating of Aaa is subsequently or rapidly downgraded to C does not per se justify a claim that the initial rating was “incorrect”. The defense to any such claim would either be “something changed since we issued our rating” or “the rating was based on the information available at the time the rating was issued, and as new information was available, we revised our rating opinion.” Despite having the ability to employ methodologies that included a measure of the stability of the rating, that very stability commonly presumed by the users of the ratings, and in particular, the users of Aaa ratings, was not – as of the time I left the Group – an explicit element of the rating analysis. On the contrary, ratings could be and were issued notwithstanding a strong probability that the rating would be downgraded – and in some cases drastically - upon any number of foreseeable and anticipated events. 31 It is my opinion that an important distinction can and should be made by the market as between structures and products appropriate for rating by the “publicly approved” agencies using tested and acceptable methodologies, and those structures and products for which no public rating can or should be issued (given the liability that would and should otherwise attach to a negligently or otherwise inadequately justified opinion). 32 Note that nothing I am suggesting precludes investors, issuers and structurers from using a private rating issued by an entity prepared to accept full responsibility and liability for errors in design, analysis and execution. Statement of Richard Michalek page 21 of 21
@FedFRASER