Auditors: Market concentration and their role - Economic Affairs Committee Contents

Examination of Witnesses (Questions 489-503)

Mr Paul Taylor, Mr Alastair Wilson and Mr Dominic Crawley

18 JANUARY 2011

  Q489The Chairman: Gentlemen, thank you very much indeed for coming in. I apologise for the fact that we are starting a bit late but, as you can see, we have tried to combine a number of different witnesses this afternoon because we are coming towards the end of our inquiry. Thank you for your patience. When you first speak, could I ask you to give your names for the benefit of the recording of this session, and to speak fairly loud and clearly? In order to not prolong the discussion, if in answer to a particular question, the person who first answers is covering the subject completely, or you are happy with the answer, please don't feel that you need to supplement it, but we are very happy to hear from you in any way you wish. Could I begin with the first question? In the same way as an audit opinion, a rating agency's rating is relied upon by parties as a form of assurance. In parallel with auditing, is it satisfactory that companies needing such a rating can both shop around between rating agencies and pay the rating agency for their work in rating? How do you preserve your independence and that of the auditors?

  Mr Taylor: I'm Paul Taylor; I'm President of Fitch Ratings. I think it's very important to start with, to go back to your word of "assured", assurance in rating. There's a fundamental difference between a credit rating and an audit report. A credit rating is a forward-looking view; we're trying to predict the future. You can't have assurance in trying to predict the future so, at some stage some ratings will always be wrong. You look at default data going back 40 or 50 years, and an investment grade rating doesn't say there's no risk; it tells you there is a level of risk and the risk should go down as the level of rating comes down. So, you can't take an assurance from a credit rating; you can use it as a guide, as one of the inputs into the process of assessing a particular situation or company or debt instrument. Perhaps I should stop now and pass over to someone else to speak.

  Mr Crawley: Dominic Crawley, I head up the Financial Services Rating Business for Standard & Poor's for Europe, Middle East and Africa. I would concur with what Paul has said, but I would just add something on the particular question of shopping around. As Paul has said, ratings are forward-looking opinions and they are derived from sourcing information from a wide variety of sources, but they are used against a framework of published, public criteria and analytical methodologies. We stand and we fall based on the criteria we use, which is public. So it isn't as if we offer different levels of service or product to different people dependent on what they might or might not pay us.

  Mr Wilson: Alistair Wilson, Moody's. I agree with both of the previous comments. I only add that whatever you think of our performance over the past two, three, four years, rating agencies have a very, very strong incentive to build and to try to maintain a strong reputation. That reputation will rest on the accuracy, however defined, of its ratings. That reputation will, in the long term, ensure that the rating agency overcomes any short-term incentives it might have to accommodate issuers.

  Q490  Lord Tugendhat: You heard what we were talking about with Lord Myners; the auditing profession is very highly regulated. You are very much less regulated. When you look at your position and the work you do and the auditing profession and the work it does, would you feel the auditing profession is over-regulated, too prescriptive as we heard earlier this afternoon, or what do you think the balance might be?

  Mr Taylor: I think your statement was correct two years ago. I think it's not correct now. Rating agencies have become very heavily regulated over the last couple of years. It started four or five years ago. We've just finished a consultancy period with the European Commission, who have just launched their third round of suggested regulation within the last couple of years. Interestingly, most of the market—the market in terms of treasury associations bank and investor organisations,—and the UK Government came out with a statement; I think I saw it today—are really pushing back quite strongly against the commission in particular, because they are starting to try to interfere with how we do the work, how we reach our decision and how we assign ratings. To give you a specific example, the commission is suggesting that we should give sovereign nations extra time to think about a decision we've come up with. So dictate that we can't announce the decision we've come to until three days after we've told the sovereign. We think that opens up all kinds of potential problems. Dodd-Frank in the US is another good example; of increased regulation. You can look at Japan and Australia and many other jurisdictions. We are very heavily regulated now. I'm not sure the sum total of that new regulation is going to add any more than the lessons we've applied ourselves from what we've learned from what's happened in this crisis that we have been living through. It helps to give a framework of regulatory confidence around the industry and some of the ideas are sensible, but a lot of it is just simply bureaucracy for the sake of it.

  Q491  Lord Tugendhat: So, you are more regulated than you used to be, but you think that the auditors are over-regulated, or not?

  Mr Crawley: For the record I would have to say that I'm certainly not qualified to make that observation.

  Q492  Lord Smith of Clifton: Gentlemen, do you consider that a rating agency's rating of a securitisation, for instance, can be regarded by auditors as a reliable guide to asset quality?

  Mr Crawley: Let me take that first. Paul has already said that ratings are forward-looking opinions of uncertain future events. Certainly, we do not carry out due diligence on assets or pools of assets; that is not the role we play in the market. What we are assessing is relative credit-worthiness; we are looking at the potential of default. We are not looking at many other factors that go into the assessment of assets in terms of liquidity, value and so on.

  Mr Wilson: It's probably worth adding that credit ratings are usually intended to reflect more than one objective. For example, credit rating agencies will try to ensure that their ratings are both accurate, but stable—which is a balance that needs to be achieved—and the ratings that come from those objectives will not be necessarily directly related to an asset's value at any one point in time. By "value" you mean its market price, because market prices will reflect a range of factors that are not germane to a credit rating.

  Q493  Lord Smith of Clifton: Could you give me some indication of what methodologies you do employ? You just don't go in and feel the seaweed and say, "up market" or "down market". What do you actually do at nine in the morning? How do you start your day?

  Mr Wilson: If by that you mean how does the credit rating process work, we have a range of methodologies that apply to industries—to particular sectors, particular types of issuer. These are intended to enable us to standardise and systematise the way in which we credit-assess issuers, whether those issuers are from the financial institution sector, structured finance or corporate finance.

  Q494  Lord Smith of Clifton: Is it largely a mathematical analysis you're going for?

  Mr Wilson: No, there's a significant degree of judgment involved in assessing credit-worthiness. The methodologies will employ some mathematical models, but even the methodologies themselves are about trying to systematise judgment, and the rating that ultimately comes from those methodologies will reflect the judgmental overlay from the rating committee. The methodologies are intended to be a starting point for any analysis of credit-worthiness.

  Q495  Lord Smith of Clifton: So there is still a lot of tea leaves and seaweed around?

  Mr Wilson: There's a lot of judgment.

  Q496  The Chairman: What view do you take, for example, in looking at a company's accounts, and what the auditor says about the accounts? Is that of great value to you in making your judgment?

  Mr Crawley: We would look at a wide range of information that we will then bring and apply against the framework of our criteria. We will look at a wide range of information that's provided directly from the company that we're engaged with; we will look at audited financial statements; we will look at market information. Certainly—in terms of the question that was asked earlier on—it is not our role and we do not carry out due diligence, but, when we look at audited financial statements, we clearly will be looking at the audit certificate and will be taking a certain level of comfort or validation from who that audit is conducted by, and the facts of the audit certificate.

  Q497  Lord Hollick: Mr Crawley, you made a couple of remarks, which I hope I noted down correctly, which puzzle me. You said that when you do a rating, you are looking at a forward-looking opinion, and secondly, you said you don't look at the underlying assets in a bond. My source on this is Mr Michael Lewis's book, The Big Short, and he describes the rating agencies giving AA ratings to bonds where the information was available by careful and diligent inquiry, that the underlying assets in the bonds—the CDOs—were already not performing properly, keys were already being handed back in, and that enabled those who did that work to make a fortune by betting against those bonds. I'm puzzled at how you can form a judgment as to a rated bond without doing diligent inquiry as to the assets that underlie it. It's nothing to do with a forward opinion; it's actually what happened today and what happened yesterday.

  Mr Crawley: My Lord, I wish to be helpful to your question, but what you are referring to is the structured finance side of the ratings business, which is not my area of responsibility or my area of specialisation so it is difficult for me to comment on that. I don't know whether—

  Mr Taylor: Let me try and answer it. We do look at the underlying assets, of course we do. If you take a mortgage bond, for example, it's a transaction which is, say, several thousand individual mortgages packaged into a cash-flow instrument. We look at the mortgages in aggregate and you rely upon the data given to you, which is generally audited so it is checked. We don't, for example, go and revalue several thousand properties; we rely upon that data being accurate. We then apply stresses to it, but we are looking at underlying data. The fact that our assumptions on the underlying data prove to be incorrect, or not stressful enough, was the issue that caused all the problems— I enjoyed reading The Big Short as well, though some of the stories are embellished. There are always a few people who take the counterargument, of course. Time has passed and it's amazing how many people now say they saw the problem coming.

  Q498  Lord Hollick: What was interesting in the book is that they did so based upon publicly available information, rather than guesswork.

  Mr Taylor: To keep things in context here, the major problem in the US housing market was the sub-prime assets, the lower quality assets. For prime assets—normal US mortgages—the typical loss number for AAA credits was something like 3% to 5%, so we assumed a loss level of 3% or 5%, which compares to less than 1% historically. For sub-prime assets, the average assumption was something like 20% to 25%, and the number was increased between 2004 and 2006. It went up by a few per cent; it went from 25% to 26% or 27%, so we were building in the declining quality of asset analysis. What we should have done is to build in a 50% enhancement level. The level of meltdown was way in excess of our worst-case assumptions.

  Q499  Lord Shipley: Just to pursue that then: what has changed following the financial meltdown in the criteria that you are using in your businesses? What is it, since people rely heavily upon the ratings that you give? What are you now doing differently in terms of those criteria?

  Mr Wilson: To pick a couple of examples, and following up on what Paul has said, I think that we understand that the assumptions that we made—particularly in the structured finance area, and particularly where US mortgages were concerned—about loss rates and correlations between assets were undercooked, and we have changed many of the key assumptions that underlie the models and the methodologies. I think, more broadly, we, like many other commentators including regulators and central banks, understand far better than we did two, three, four years ago the degree of interconnection within the financial sector and within the global economy, so that we understand better the need not to look at financial institutions in isolation from the people they lend to and the sovereigns under whose jurisdiction they sit, for example. So we understand better the need to try to look at the issues—as we rate—in their broader context.

  Mr Taylor: It's important to mention as well, when we talk about criteria, that there are dozens of criteria to apply to different types of credit. We have corporate credit criteria, we have bank criteria, and even within structured finance you have dozens of types of criteria, whether it's German auto loans or Japanese houses. Most of the structured market, which again tends to be missed through most of the dialogue, has actually performed very well in the face of incredible economic crisis. So, if we look at the UK housing market, AAA bonds have done particularly well. As has the $1 trillion European residential mortgage market, it has generally done fine. We still tweak some of the criteria in different markets but US residential mortgages, and more importantly the repackaging of those assets into CDOs or CDO2, is really where the big changes have come. The second big change, I think, is—and this is something that's ongoing at the moment—how we treat banks going forward. In the past we had a very heavy reliance on the assumption that we'd get support. We built in support to our credit ratings, and it generally worked. That's why our ratings have been fine in banks with a few exceptions but we think that's going to be different going forward and that's a huge area of debate at the moment.

  Q500  Lord Shipley: By support do you mean support from the Government?

  Mr Taylor: We are rating the debt instrument. Will you get your money back on a particular bond or security?

  Lord Shipley: Sorry, Mr Crawley, are you—

  Mr Crawley: I would certainly endorse what the others have said. Certainly correlation of risk, how that then led to very significant concentrations of risk and then, when cycles were turned, how they turned very viciously in terms of impact. That's what we saw and, frankly, all the market learned through the last three and a half years. The one thing that I would add, which has been touched upon by both Paul and Alistair, is about rating stability. That's certainly something that at Standard & Poor's we have spent a lot of time over recent months looking at, debating, drafting and now discussing with the market around how better to ensure that highly rated entities and highly rated securities are able to perform with a high level of stability demonstrable by the high ratings that they are given.

  The Chairman: Do you want to go on to the next question about Big Four?

  Q501  Lord Shipley: Do any of you employ Big Four firms to assist you with any of your work and, if so, do you find market concentration of public accounting firms sometimes unduly restricts your choice?

  Mr Crawley: The straight answer is, no, we don't. We make use of audited financial statements, but we do not engage in a contractual sense with firms of auditors.

  Mr Wilson: We occasionally use consultancy arms to carry out small projects—

  Lord Shipley: Sorry, I missed that.

  Mr Wilson: Sorry, we occasionally use consultancy arms of the Big Four to carry out projects for us, but as regards the audit practice, it's precisely as Dominic says.

  Q502  Lord Hollick: When we talked about the Big Four's concentration, we referred to concentration in other parts of market, and obviously the rating agencies are a case in point. How did the concentration come about? Do you feel it acts against public interest, that there is lack of choice? For instance, why isn't there a rating agency that has a different economic model that is paid for by investors rather than the banks themselves, which might lend it a greater degree of independence?

  Mr Crawley: There are two questions. Perhaps if I take the first one. I think how we have a fairly small number of large rating agencies probably comes from two basic characteristics. One is that our relevance in the market is built upon reputation and it's built upon performance, and therefore a long-time series track record of our ratings performance and ratings transition is important and clearly that is something that is established over a long time. I think the other issue is that a relationship between an entity and ourselves—whether it's a large corporation or insurance company, bank, local authority, whatever it is—is a fairly intense relationship. It is regular contact, a lot of dialogue and, therefore, from the client's side, so to speak, it's fairly intensive. Therefore to have a large number of similar individual types of relationships from their side probably would not make sense. I think an element of the concentration has developed as a result of those two features.

  Mr Wilson: Yes, I would say I agree with Dominic. The concentration has risen to a large extent because of the way commercial firms have responded to what they think their clients' investors want. Investors want ratings that are comparable across regions, globally comparable, comparable across time. That tends to be best delivered by large firms with large global resources. I think it's also, and this is long before my time, fair to point out that originally—certainly for Moody's; I am not sure about the other two—the model was "investor pays" until the early 1970s, but it moved away from that towards "issuer pays". As I understand it, and, as I say, this is way before my time, that was to enable Moody's to help meet investor needs and wishes for more research. So we have moved from the "investor pays" model to the "issuer pays" model.

  Mr Taylor: Sir, our current industry structure is largely a result of inertia. We always have here the Big Three; actually we're a third the size of these two guys. We're the new player in this market and we have really only existed in this form for about 10 years. We're the result of putting together some of the smaller agencies in the market to try and compete with the big two players. The inertia in our market and the inclusion in investor guidelines, for example, is immense. There are thousands of individual accounts that we have to get out to. We were doing pretty well at that until a few years ago when the whole market had its own problems. So a lot of it is inertia. There are something like 150 rating agencies in the world, of which a good number are "investor pays", so that model does exist. There's nothing stopping that model. It comes down to economics at the end of the day. It's very, very hard to build up a significant team with just an "investor pays" model, because investors are used to getting things for free. Even the new players that have been announced—there have been a number of new launches, pretty high-profile launches with significant financial support behind them—have all launched as "issuer pays" models. It comes down to us to manage the conflicts inherent in that model. There are conflicts in the "investor pays" model as well. If you have an investor who is providing a large proportion of your revenue, the investor might have an interest whether your rating goes up or down, so there are conflicts whichever way you look at it. If you use a government-funded approach, that's probably the most conflicted path you could go down. So what does the UK Government-funded agency do when it wants to take action on the UK, or a big UK bank? You can imagine the conflict. In our industry there are lots of players; the challenge you have is how many ratings does an entity want? You can push hard to get in there as an additional rating if you're the third player coming in; if you're a bank do you really want four, five, six credit ratings? The answer is, no, you don't. Really the first two places are taken and it's incredibly difficult to get rid of an S&P or Moody's rating if you are a rated entity. That was just beginning to happen a few years ago, and we've just started to see the signs of that happening again.

  Q503  Lord Moonie: Do auditors have access to any information from the rating agencies beside published ratings? Is there any dialogue between auditors and rating agencies about the quality of assets and bank balance sheets? Do you consider such a dialogue needs to be developed, or perhaps institutionalised?

  Mr Crawley: My Lord, there were two or three questions there. To answer the first one, no, they do not. Like everyone else in the market they have access to the ratings and the information and research opinions and analysis that gets published and placed on our public website. I think in response to the second question what I would say is the market knows what we do and the market understands what we do, and the market knows what auditors do and understands what auditors do. We talked earlier on about what our ratings are and how we go about constructing them, and we don't see there would be any benefit through additional dialogue. They do what they do, and they've done it for a long time, they've established processes, procedures, resourcing and everything, and that is their specialisation. Our specialisation is what we do.

  Mr Wilson: I think you could imagine a situation in which an issuer was prepared, or wanted its auditors to talk to the rating agencies in order that the rating agencies had a better understanding of, for example, the control framework in the issuer. So I can imagine a situation in which the information flow would usefully work in that direction. I think it's very difficult to imagine a situation in which information was flowing in the opposite direction because it would raise some real confidentiality independence issues.

  The Chairman: Any other questions? Gentlemen, thank you very much indeed for coming. We are very grateful to you for giving an insight from your point of view, and apologies again for a late start. Thank you.

previous page contents

© Parliamentary copyright 2011