IFR: I think it is time to move on. I am going to be the first person to mention Joint Default Analysis (JDA), which had to come. Michael, Moody's is still near the beginning of wheeling this out; were you surprised by the reaction it received?
Madelain: This has been a process that has been taking place a little over two years now. There has been extensive market outreach and discussion and I think one has to differentiate between the perception around the conceptual framework and the actual execution of the introduction of this change.
Clearly, you know, there has been a very vocal market reaction about it, I think much more than what we expected. We have a very active process at the moment of engaging with market participants and the feedback we are getting is that there seems to be some conversion. The overall framework, the conceptual framework, seems to be well accepted and understood. The concern is really around the disruption caused by the change in the ratings and the scope and timing of these changes.
So, as you said, it is at a very early stage, and we are tying to address the concerns being expressed.
IFR: Is this a situation where we can use the phrase "transparently wrong"? I mean the process is transparent but is this actually a situation where, the market is saying its transparently wrong?
Hawley: I think it is difficult to say that so unequivocally. As Michel said, it has been two years in germination and a lot of market participants have been invited to comment. But, something has gone wrong. The end result has been that market participants have looked at the methodology and formed a view and thought, "no". It has come out differently to what had been expected so there is a gap between what had been expected and what transpired, that’s come as a surprise. I am not quite sure what went wrong.
IFR: Maybe you could have done it in a different order, if perhaps, you had gone with the US to begin with, as a low support country where the ratings differences were not that great. Maybe you could have broken the market in gently.
Madelain: I think there was a bit of a surprise around the first changes that were introduced. But I think one has to really go beyond that and think about the conceptual framework and the underlying opinion and judgments that are being assessed. I think what we have tried to do is to make a clear distinction between the risk of default and the expected loss for bank paper. We think that what we should address is to provide additional transparency around those two metrics.
But I think, as you say, just by the extensive dialogue and the broad consensus, the most surprising thing that emerged from that dialogue was the implication in terms of ratings. We have seen a lot of intervention and comments and reactions.
IFR: I think the Icelandic banks, which too have employed the same method, still would not have come out Triple A, would they? So, is it an assumption placed on an assumption?
Madelain: This is interesting. We spent an hour talking about the lack of competition and the need for a different opinion. Yet effectively what you are saying through the line of question, is that there something wrong in having different views.
IFR: Which is not the same as competition is it, that is discrepancy?
Theodore: Competition of ideas and outcomes.
Madelain: Yes, we try to do the best thing we can and we try to refine analytics and provide and improve opinions and judgments. At the end of the day, they are just opinions, I think we say that very clearly. But I think this illustrates what should be expected in a situation where you have competing and different views on different assets.
IFR: At Fitch you had looked at implementing this approach and stepped back. Why was that?
Hunter: I do not know about "stepped back". The issue was that, of all the agencies, we have had the longest track record of having very clear, separate financial strength ratings and a support rating. The combination of those two giving the long-term rating.
I think our primary concern was there was an element of the tail wagging the dog. I refer to something Blaise said right at the start. It is a binary issue with ratings. We are only looking for default, which is thankfully a low frequency event. So to take those events which occur, particularly at the distinction of the higher end of the ratings scale, I think you said “assumptions upon assumptions”, there is the danger that this leads you to a very strong expression at the top end of the rating scale.
When we identified that, we had moved to, if you like, a modified form of arithmetic. We look at both the elements, and take a "higher of" approach that definitely reflects support. But that still does not take you to the point of having a number of ratings notches that you have seen. But it still allows for a differential between institutions at the top end of the scale.
There is also a fear that a lot is said about the lack of default history for financial institutions. There is always the danger that people will say, “banks will not default, because they have never defaulted.” Well, people used to say that about utilities until the 1980s and then, in addition, you had nuclear power. So there is the risk, as a number of people have already mentioned, that the likelihood of support changes throughout the developing world. The very small data-set that you are looking at, at the top end of ratings scale, to drive a significant number of ratings, certainly was not something that we felt would be an attractive idea.
Ganguin: And from our standpoint we have always incorporated the external support from the government into our bank ratings. Clearly the large banks are a systemic fabric of countries, so it is important, and it has always been important to incorporate that aspect into the analytics.
One of the things that Richard mentioned which is also true from our standpoint is that we would like to continue to differentiate at the top end of the scale with some granularity among the high Single As, Double As and Triple As, to show that there is no expectation of default at investment grade.
Having said that, I think we should welcome the competition for ideas. I think that that shows that there is space for different methodologies out there and investors have to get their own minds around that.
IFR: Sam, you were probably at Moody's when the idea was first floated; what is your take on it?
Theodore: I moved out!
IFR: Was that expressing an opinion or not?
Theodore: No it wasn’t! One year ago we published a methodology for rating banks, and we do agree with the element of support. My view is that the analysis of suspected government support is the least quantifiable part of fundamental analysis of a bank. So to be able to create a default probability for this expected support and to incorporate that percentage, whether it is 90%, 98%, or 70%, and to be able to make that decision on a fundamental basis is, in my view impossible. I cannot do it as a fundamental analyst and I do not think any model can.
So, to that extent, you have a model and analysis with which some people in the market will disagree.
Another point is that, in general, I am against trying to use a model based approach to fundamental analysis. When you talk to investors they like to have opinions expressed by real people, by people with experience, with expertise, with an analytical equation, not expressed by a certain mix and match of ratios and of ratings of the markets.
The second point, that Blaise referred to, is that when you get to the top end of the ratings scale, if you look at pure default probability models, you will see that the difference between a Triple A and a Double A plus, or between a Double A plus or a Double A, is a few basis points. So it is challenging to apply the pure prospective loss analysis for those ratings. I have difficulty in finding the fundamental analysts in this world who can say, for ‘X’, I think there is a 2bp higher default probability than ‘Y’. Therefore, the rating of ‘X’ should be Triple A and the rating of ‘Y’ should be Double A plus. It is simply impossible.
At the end of the day, ratings at the top of the ratings scale reflect value and that is what people want to know. You tell the market the bank industry is not going to default for another 15 years and then you look back and say “I was correct”. But that is not always what users of ratings are looking for. They want you to differentiate between ‘X’ and ‘Y’. That is the role of the rating agency, to measure relative creditors.
The beauty of the market is that it incorporates in one single price a lot of market information. So in a way the beauty of a rating is that it expresses in one symbol all the credit analysis and information which exists in respect of that issuer.
So to me, what is really important, is not to lose sight of the main goal. That a rating is meant to be something simple and direct to address the investor's needs. People do not have time to look at a wide range of ratings and sub-ratings and the correlation of ratings. They want to know one answer, because otherwise they would do the analysis themselves. That is why I think that rating, the flat sheet rating, should be the one used by most people. If some have time or interest to look at sub-categories, that is fine, but I do not think we should point people to things, saying, "Look at this one, do not look at that one".
IFR: Is there a danger that, if one agency's ratings are so far out of kilter with everyone else's, that they will become disregarded, or people will simply ignore them. All Triple As are not created equally so are they to be disregarded?
Theodore: I do not think they should be disregarded, because it is not only the rating which is important, but the analysis behind the rating.
IFR: But, as you say, some people may not have time or may not be willing to delve that deeply.
Madelain: But the truth will be in the default. Okay so we will not be here around this table in 25 years time. But someone will look at the default table and say, “you know what, they were Triple As”. You can see different people have different aims and objectives. When they look at the ratings, they see different views and want to serve different needs. That’s what makes this complex.
Liehr: I think at least what you did trigger is a real look at the level of support, and I think that was in itself quite positive. I mean, being independent of the rating implication that we had, it is really the first time where you have been radically different.
This is clearly something that issuers will be able to carry to the other agencies.
Aucutt: And to a degree the bank financial strength rating is a very good start around getting some good quantitative analytics. But if you overlay that with the JDA, which just automatically gives you a Triple A, the rating is less meaningful. I think you probably overestimated the extent of government support. Certainly I know you did some work on subordinated debt, but I just wonder how much. If a bank goes into default, how much does the government support the subordinated debt?
IFR: Because subordinated ratings just moved up?
Aucutt: Yes, and I would be very surprised if the preference share was supported in event of default.
IFR: There did not seem to be that much differential made between the senior and subordinated, everything just moved up lock, stock. Should there have been more difference, do you think?
Ganguin: I think when we are talking about subordination, it is significant to talk about loss and default in general. When we discuss subordination we are moving away from the default risk, but we are more looking at the priority ranking, or the waterfall system in structured finance. We are looking at two very separate things.
In the past, we were differentiating the relative positions of two debt holders within the same capital structure through the notching system. Since 2003 we introduced an alternative scale that is totally separate from the Triple A to Single D scale. A break from the notching approach, as we felt that there was the risk that, particularly in the sub-investment grade zone, we were comparing apples and oranges.
I think when we are talking about new avenues of development for rating agencies, this is definitely a key part of it. Credit net loss is definitely part of credit, this is a totally new area that everyone is looking into. I would argue, having worked on that particular topic for the past six or seven years, that no-one really can claim to have a good handle on it, simply because it is still too new a topic. But I think everyone is really looking into it just to understand the difference between default and loss-given default. And also to be able to benchmark their own analytics to figure out what is going to be the actual loss.
IFR: How do you approach recovery ratings, Richard?
Hunter: We have a recovery ratings scale and we roll the recovery ratings out as well. But, as Blaise says, it is difficult because you come back to this issue of what actually is the observed data. There is very little in the way of publicly consistent and observable data on the topic. That said, we ended up rolling them out not just to the usual suspects but to all the countries where we have international scale ratings.
What was interesting was the relationship between a bond's ratings and its recovery rating. Something had happened to make the company more or less valuable in the post-default value. They had done something to the capital structure, perhaps by raising more secured debt, or paying secured debt. So the whole concept is not at all straight-forward, I agree with Blaise, this really is a new frontier.
One area where it gets less pressing, but where it is also relevant, is on the structured side. We have recovery ratings for structured instruments at the very low end of the scale of structured transactions. But the rest of our structured finance ratings are essentially first loss. What we are seeing at the moment is an increasing number of very highly rated senior structured products where there is a reasonable element of loss-given-default (LGD).
In the past a Triple A tranche of an RMBS with a 10% LGD would have been extremely unlucky. But, on the synthetic side you could have a 90% LGD. This is not something that is really measured in a Triple A symbol.
IFR: Sam, what’s your opinion on recovery ratings?
Theodore: We do not have recovery ratings. Again, we try to stay away from complicating the game too much. You introduce too many ratings scales and people get lost and they do not have time. The investors use ratings to simplify their life, not to complicate it. So we try to address these issues through the underlying analysis rather than forcing people to look at different ratings and make a correlation themselves. We do not ask them to draw their own conclusion based on a very complicated game, we try to stay away from this.
But, coming back to the issue of subordinated versus senior. At the senior rating, the likelihood of the subordinated debt behaving differently is extremely low. No rating agency would like to see a Triple A bank defaulting at the Triple A level. So, if we assume that the rating would scale down as fundamentals worsen, the likelihood of expected support would also worsen to some extent. As such, I assume we would probably see a wider notching than would be seen in subordinated debt. So I would not be overly concerned from seeing subordinated debt at a higher level, as long as some notching has been expected there.
IFR : Moody's was considering introducing incremental notching, was it not, of some subordinated paper until recently and then it was pulled?
Madelain: There was a request for one that was published and the discussion was around adding notches for different payments. We had a fairly extensive market consultation. After we put out the proposal, we had huge feedback with probably had the largest number of conference calls ever. We took the view that the proposal, effectively recognized a difference that did not necessarily need to be translated into notching. The question is, do you want to flag a difference and should you flag that through a difference in notching?
This was a case where there seemed to be a consensus. We recognised that and moved on and published a revised methodology that incorporated that feedback.
We have not talked too much about what has changed in this industry for the last five years, but to me there are two main things that have changed. One is the degree of transparency in terms of methodologies, and I would say the second is the level of consultation that takes place with the markets, which has improved tremendously, and we see a lot of initiative around that.
The third is innovation. We all welcome recovery ratings. The three agencies have different approaches, and they all expect their approach to eventually become, if not the market standard, at least the one that is benchmarked as providing the most value. In structured finance, there is constant competition for innovation and development of analytics, so it is a very lively industry from this perspective.
IFR: So what we now have is a transparent, innovative market?
Hunter: I think we missed the adjective ‘lively’.
Theodore: And ‘competitive’.