A Framework for Valuing Financial Hybrids - …quantlabs.net/academy/download/free_quant_instituitional_books_/[JP... · A Framework for Valuing Financial Hybrids Introducing and

European Credit Research 15 September 2006

A Framework for Valuing Financial Hybrids

Introducing and Applying our Hybrids Valuation Framework to Bank Tier I Debt

Financials Credit Research

Roberto Henriques,CFAAC

(44-20) 7777-4506 [email protected]

Credit Derivatives & Quantitative Credit Research

Jonny GouldenAC (44-20) 7325-9582 [email protected]

Andrew Granger (44-20) 7777-1025 [email protected]

J.P. Morgan Securities Ltd.

See page 42 for analyst certification and important disclosures, including investment banking relationships.JPMorgan does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that the firmmay have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor inmaking their investment decision.

• We introduce a comprehensive valuation framework for traditional ‘step’ and ‘non-step’ Tier I instruments, encompassing both the Euro and Sterling-denominated markets.

• Our valuation framework allows us to account for the non-standard features of Tier I instruments – subordination, coupon deferral and call risk - within a transparent credit pricing model

• In the ‘step’ Tier I market, our framework concludes that:

- instruments issued by highly rated entities (Aa3 and above) are currently cheap, whereas;

- instruments issued by lower rated entities (A2 and below) are currently expensive.

• The valuation framework implies that on aggregate, ‘step’ Tier I as an asset class is marginally cheap, with the Sterling market trading more cheaply than its Euro-denominated counterpart.

• The resulting sensitivity analysis provides insight into which of the non-standard features of Tier I has the most impact on valuation and enables us to stress valuations where credit quality is changing.

Figure 1: € Tier I Banks - Rock-bottom Relative Value Through Time bp

-50

0

50

100

150

200

250

Oct-01 Apr-03 Oct-04 Apr-06

Weighted Ave T1 Spread (ASW spread, bp)Market Excess Spread (Market - R BS Spread, bp)

Cheap

Dear-50

0

50

100

150

200

250

Oct-01 Apr-03 Oct-04 Apr-06

Weighted Ave T1 Spread (ASW spread, bp)Market Excess Spread (Market - R BS Spread, bp)

Cheap

Dear

Source: JPMorgan

2


Roberto Henriques,CFA (44-20) 7777-4506 [email protected]

Jonny Goulden (44-20) 7325-9582 [email protected]

Table of Contents 1. Executive Summary .............................................................3 2. Overview ...............................................................................4 Recap - Building Block Approach ...............................................................................4 Overview of our new Valuation Framework ...............................................................5 The Rating Transition of Banks versus Industrials ......................................................6 Valuation in Brief ........................................................................................................8 3. The Valuation Framework in Detail .....................................9 Subordination...............................................................................................................9 Rating States as Coupon Deferral and Extension Triggers ........................................10 Coupon Deferral ........................................................................................................11 Mandatory Coupon Deferral ......................................................................................11 Voluntary Coupon Deferral .......................................................................................12 Valuing Coupon Deferral...........................................................................................14 Extension Risk ...........................................................................................................14 Mandatory Extension .................................................................................................15 Voluntary Extension ..................................................................................................15 Assumption for extension risk ...................................................................................16 Pricing the Hybrid......................................................................................................17 4. Results from Valuation Framework ..................................18 Steps-ups: Rock-bottom or ‘Fair Value’ by Rating ...................................................18 Steps-ups: Rock-bottom or ‘Fair Value’ by Spread ...................................................20 Step Tier I Market: Overall Pricing ...........................................................................21 5. Sensitivity Analysis............................................................22 Sensitivity to Initial (Start) Ratings ...........................................................................22 Sensitivity to Coupon Deferral Trigger Rating..........................................................22 Sensitivity to Call Trigger Ratings.............................................................................23 Sensitivity to Recovery Rates ....................................................................................24 6. What about the Non-steps?...............................................26 So, will they be called? ..............................................................................................26 Refinancing Costs ......................................................................................................27 Equating funding costs to rating states ......................................................................27 Results for the € non-steps.........................................................................................28 Non-Step - Conclusions .............................................................................................29 7. Historical Analysis .............................................................31 BAVB 6.988% €11 ....................................................................................................31 CRDIT 5.396% £15 ...................................................................................................32 Appendix A – Full Results of the Valuation Framework .....33 Appendix B – Detailed Workings of Our Hybrids Valuation Framework ..............................................................................35

3




1. Executive Summary This note introduces and applies our hybrids valuation framework to financial hybrids. Our valuation methodology for Tier I takes into consideration the specific credit risk features of the asset class – namely: subordination, coupon deferral and extension risk – which introduce an element of uncertainty into future cashflows. Our approach is based on the assumption that the cashflows are dependent on the financial state of the issuer which can be represented by the senior rating attributed to the issuer. We can therefore derive a ‘fair value’ for Tier I using the JPMorgan’s Rock-Bottom Spread framework, by using the issuer ratings and sector-specific ratings transition matrices.

The application of the valuation framework to benchmark Tier I bonds, both in Euro and Sterling, implies that the asset class is cheap relative to current market levels. While the overall asset class appears cheap in line with the ‘fair values’ derived by our valuation methodology, we note that there is a bifurcation in the ‘fair values’ attributed to the instruments issued by highly rated issuers versus lower rated issuers. According to the results of our valuation framework, Tier I issued by highly rated entities (senior rating of Aa3 and above) tends to trade cheaply in the market, i.e. the ‘fair value’ (in terms of spread) is tighter than current market spreads. However, Tier I issued by lower rated entities (senior rating of A2 and below) tends to trade expensively in the market, i.e. the ‘fair value’ (in terms of spread) is wider than current market spreads.

The framework we use for valuation is intuitive and transparent, allowing investors to implement a version and to see the impact of their different views on the pricing inputs.

Our valuation framework allows us to develop a sensitivity analysis in terms of identifying the variables that most impact ‘fair value’. In addition, the sensitivity analysis also demonstrates what the impact on the ‘fair value’ is, in the event of different ratings triggers being selected either for coupon deferral or extension risk. We will see that recovery rate assumptions play a limited role in the derivation of the fair value.

We apply our valuation methodology to value the non-step Tier I instruments, an area where there has been a lot of debate with regard to valuation approaches. For the non-steps we assume that the decision to call these instruments will be a function exclusively of refinancing costs, with the issuer only calling the bond if it can be refinanced at a cheaper level. In terms of the valuation framework, we therefore flex the assumption with regard to extension risk, by making this dependent on expected refinancing costs.

The availability of time-series data for Tier I has allowed us to perform some limited historical validation of the valuation framework in order to test the robustness of the valuation approach. Given that the valuation approach is driven by the issuer ratings, we have selected a sample of issuers which have been impacted by a material level of rating volatility over the recent past. We present the results of this analysis for BAVB (€ Tier I) and CRDIT (£ Tier I).

We extend our hybrids valuation framework to value Tier I

Highly rated Tier I looks cheap, lower rated Tier I looks expensive.

Sensitivity analysis allows us to identify drivers of Tier I valuations

Non-step Tier I valuations driven by expected refinancing costs

We perform back-testing for a select group of Tier I instruments

4




2. Overview The present note represents both a continuation of our work on developing a valuation framework for Tier I and a continuation of our work in valuing hybrid capital in general. In our research note “Are Tier I Valuations Stretched” (Jan 2005), we proposed a building block approach which separately valued the different components of Tier I with a view to obtaining a comprehensive valuation for the asset class. Whilst this approach shed insight on the absolute value of the different components, the emergence of a corporate hybrid sector and the implementation of our valuation framework for these has led us and many investors to look to value financial hybrids using this approach. Before presenting our new valuation framework, we will briefly recap on our original building block methodology.

Recap - Building Block Approach As suggested by the name, our previous building block approach to valuing Tier I debt was based on valuing the individual components or features of the asset class and using these to derive an aggregate valuation. Using senior market levels as the starting point, the valuations derived for the specific features of Tier I, such as subordination, coupon deferral and extension risk, were added to determine a theoretical ‘fair value’ which would then serve as a term of comparison for prevailing market spreads. This approach had the advantage of being extremely intuitive, whilst also providing some insight into the valuation differential between the different classes of bank capital.

Figure 2: Overview of the building block approach to valuation Bp

0

10

20

30

40

50

60

70

1 Yr 2 Yr 3 Yr 4 Yr 5 Yr 6 Yr 7 Yr 8 Yr 9 Yr 10 Yr

CDS 40% Rec Rate

0% Recovery Rate

Coupon Deferral Option

Extension Risk

Source: JPMorgan

However, two important critiques of this approach have encouraged us to adopt an alternative valuation framework. First, by valuing the various features of Tier I separately, the interaction between these features on valuation is ignored. Plainly put, we believe there is an inherent interaction on value between how likely coupon deferral is and what the call 'option' is worth. An example of the extremes should illustrate this. Say we have a financial hybrid with the first call at year 10 where we believe the company will defer coupons with a high degree of certainty. This will

Previous valuation approach based on 'sum-of -the-parts' methodology

‘Sum-of-the-parts’ approach ignored interaction between features of Tier I on valuation

5




either look like a zero-coupon 10yr or a zero-coupon perpetual, depending on our assumption on the company calling. The call feature will dominate value as it will be the chance of getting at least some cashflow from the hybrid at year 10 (i.e. par if it is called) or getting nothing. If we now assumed we had a similar hybrid where we assessed the company will not defer coupons with near certainty, then we will have a straight 10 year bond or a perpetual depending on our call assumption. The value of this call trigger in this second hybrid will be different to that in the first, as we now will get a lot more value in the coupons and could even want the hybrid not to be called if the post-call spread is attractive. This shows there is an inherent interaction between the value of the call feature and the likelihood of coupon deferral (and vice versa).

Therefore, by considering the coupon deferral and the extension risk options in isolation, we ignore the interaction between the likelihood of coupon deferral and likelihood of the call and how this will impact valuations.

Second, the building block approach itself is implicitly based on market levels, given that our starting point is senior market levels, from which we then build on to ultimately derive our valuation. Our new approach represents a more integrated framework based on calculating the present value of uncertain cash flows. This will give us a valuation for financial hybrid capital that is also 'independent’ of current market levels.

Overview of our new Valuation Framework The valuation framework which we will introduce for Tier I is the framework which has been previously applied to the valuation of corporate hybrids (see 'A Framework for Pricing Corporate Hybrids', Goulden, Keenan, Sep 2005). In essence, this framework is based on the calculation of the present value of a set of cashflows in order to derive a fair value for the financial hybrids. However, the specific features of Tier I introduce an element of uncertainty into the likelihood that cashflows will be paid and hence the probability (and size) of these cashflows has to be factored into the derivation of the fair value of these instruments. The single most important assumptions that are made in constructing the valuation framework are that: a) the probability of cashflows being received is linked to the financial state of the issuer and b) we can represent the financial states of the issuer by its ratings. This translates into defining ratings triggers for each of the mechanisms in Tier I structures as a function of the underlying financial state of the issuer. Under this approach, it is the financial state of the issuer, as defined by the ratings, which will determine probability of the cashflows.

Having made the link between the rating states of the firm and the future cashflows of the instrument, we now only require a way of creating a future distribution of the financial states of the firm and which will allow us to assign probabilities to each state of the firm (and hence the cashflows). Such a distribution of financial states of the firm would, in our opinion, have to be transparent and replicable to investors, as well as maintaining some level of market independence. We use the empirical evidence available of the probability of being in each rating state, as represented by a ratings transition matrix. This is an aggregate history (collected by the ratings agencies) of the probability of a firm with a given rating being in each rating state over a period. From this we can create a forward-looking probability distribution

Previous valuation approach was not market ‘independent’

Valuation framework assumes financial condition of firm is the driver of Tier I cashflows

Ratings transition matrix provides probability distribution of future rating states

6




which will tell us the probability of the firm being in each rating state at each cashflow point and hence the likelihood of the cashflow.

Conceptually this way of looking at the value of a firm and its debt through ratings states is also found in other credit valuation frameworks such as the CreditMetricsTM framework. This framework calculates bond prices and risk by using the issuer’s ratings and transition matrix to create a future probability distribution of values (assuming leverage is fixed, see Figure 3). We will therefore use rating states and transition matrices as the basis for our valuation framework.

Figure 3: CreditMetricsTM View of Rating States of a Firm as Linked to Asset Value Distribution

-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0Asset Value (changes, %)

AAA

A

AA

BBBBB

B

CCCD

-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0Asset Value (changes, %)

AAA

A

AA

BBBBB

B

CCCD

Source: JPMorgan, CreditMetrics

Our general valuation approach is therefore to make future cashflows for Tier I instruments dependent on rating states. This means that we can ascribe a probability to a cashflow based on the probability of the issuers being in a rating state as determined by the transition matrices. Setting the triggers in line with predetermined rating states (which reflect a determined underlying financial state), we can obtain a probability distribution of these cash flows which ultimately will allow us to value Tier I.

The Rating Transition of Banks versus Industrials The most commonly used and available ratings transition matrices contain data from both financial firms and industrial firms. However, most market participants would recognise that banks have a very different ratings transition and default experience to that of corporates1. We have created a ‘banks-only’ transition matrix from the underlying ratings transition and default data available2. This is a key tool in being able to value financial hybrids.

It can be instructive to view what difference this makes to the underlying distribution of the firm.

1 The bank regulatory environment plays a key role with close regulatory oversight and the strategic importance of the sector resulting in a materially lower probability of default versus corporates. 2 Specifically, we create an average annual transition matrix from the Moody's historical default and transition database. We have only included Banks and the period we have used is from 1983 to 2005.

Banks and corporates have different default and ratings transition experiences

7




Figure 4 shows the distribution of rating states of an A3-rated bank and that of an A3-rated corporate. These distributions use the differing ratings transition matrices to calculate the probability of being in each rating state at year 10 (chosen as most Tier I instruments have the first call at year 10). The height of the bars indicates the probability of being in each state. The (light blue) corporate distribution shows a mode of A3 rating (the initial rating) and a somewhat normal distribution, skewed to the downside with a ‘fat tail’, specifically with more chance of being in default. If we think of these rating states as asset returns, this fits with how we largely think of credit return distributions.

The (dark blue) banks distribution is somewhat different. Here the mode is actually higher rated (A2), with a large part of the corporate distribution between Baa1 and Baa3 being 'shifted' higher up the probability distribution. In plain language, an A3 rated corporate has a large chance of being BBB in 10 years, whereas with banks, the probability of being BBB is much smaller ie they have a higher chance of being more highly rated. This seems rational, as banks tend not to linger in the low investment grade ratings, with a common pattern being thatr a bank which experiences material financial distress and subsequent ratings downgrades is acquired by a more-highly rated bank and benefiting from a positive ratings upgrade as a result.

Figure 4: Banks’ and Corporates’ Distribution Compared Probability (%) of Being in Each Rating State at Year 10 as derived from Corporate and Bank-only transition matrices.

0%

5%

10%

15%

20%

Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa D

0%

5%

10%

15%

20%


CorporatesBanks

0%

5%

10%

15%

20%


0%

5%

10%

15%

20%


CorporatesBanksSource: JPMorgan, Moody's

8




Valuation in Brief Having provided an overview of the valuation methodology, we will now look to value Tier I instruments as a set of cashflows with four distinguishing features:

1. Default risk – the firm can default over the life of the bond. Default probabilities will be taken from the transition matrices and a recovery rate will be paid on default.

2. Subordination – As Tier I is subordinated to senior debt, we will set the recovery rate lower than for regular bonds (our base case is to set it to zero).

3. Coupon deferral – Coupons on Tier I can be deferred in certain circumstances and we set a rating state at (or below) which the issuer will defer its coupons.

4. Extension risk – Tier I has call features and we set a rating state at (or above) which the company will call the bond after 10 years (or quarterly thereafter).

To move from this set of assumptions to a valuation for Tier I requires a valuation methodology for bonds which uses ratings and transition matrices to determine fair value. JPMorgan’s Rock-Bottom Spread methodology looks at valuation of corporate bonds using the cashflow structure along with ratings and transition matrices as inputs. It can be adapted to account for the additional cashflow possibilities that can occur with Tier I.

The next section goes into more detail on how we value each feature of financial hybrids. Readers who wish to skip this can go straight to section 4., which has the results of our valuation applied to the current Tier I market. Appendix B provides a more comprehensive step-by-step guide to our calculations, with workings.

Valuation Framework Summary

Tier I cashflows made dependent on financial states of the firm

Financial states of the firm represented by ratings

Probabilities of ratings and hence states of firm given by ratings

transition matrices

Tier I can be priced using Rock-Bottom Spread framework using

future cashflows and probabilities of each cashflow

9




3. The Valuation Framework in Detail In this section we individually review each of the non-standard features of Tier I and articulate the assumptions which we will be making within the scope of our valuation framework. We will therefore detail what our assumptions are in terms of subordination, coupon deferral and extension risk and at what level we will set the ratings trigger within our valuation framework. We would also point out that the sensitivity analysis that we develop in Section 5 will allow us to see what the impact of changing these assumptions will be on the 'fair value' of the Tier I instruments.

Subordination The first feature distinguishing Tier I from regular senior debt is the subordination that the holders of these instruments have in relation to senior creditors. In the event of default, the deeper subordination will impact the recovery rate achieved by holders of Tier I and may imply that the recovery rate is significantly lower than that of senior debt holders. An investor should therefore require additional compensation for holding deeply subordinated debt and we can build the subordination feature into our valuation framework by defining an assumption with regard to the expected recovery rate on Tier I.

Our base-case assumption is a 0% recovery rate for Tier I against 40% (generally assumed in the CDS market) for senior debt. By assuming a 0% recovery rate we are, in effect, assuming a worst-case scenario for Tier I We believe that this assumption (which may appear excessively conservative) is the most appropriate given the scant history of recovery rate data in the European banking sector, where Barings and BCCI are relatively rare examples of large-scale bank failures. In addition, we would also point out that despite the conservative nature of our recovery rate assumption, a sensitivity analysis which we develop as part of the valuation framework indicates that the recovery rate is not a key variable in terms of deriving ‘fair value’. This stems from the fact that for banks, the actual probability of default, as reflected in the sector-specific transition matrices, is extremely low, hence the eventual impact of variable recovery rate assumptions is limited by the reduced probabilities of an actual default.

Therefore the deeper degree of subordination of Tier I is incorporated into the valuation framework by assuming a lower recovery rate.

Figure 5 illustrates how we deal with subordination within our framework. We compare a Baa2-rated 1-year senior bond to a Tier I instrument, with a 5% coupon. The probabilities of being in each Default / No Default state are taken from a 1-year 18-state banks transition matrix and we have used a risk-free discount rate of 3.50%.

Base-case assumption for recovery rate on Tier I is 0%

10




Figure 5: Valuing subordination in Tier I

Source: JPMorgan.

Essentially, we work through the value in each state (default / no default) and the probabilities of being in each state as given by the ratings transition matrix (the actual model introduces additional calculations as the Appendix shows). The difference in fair value price for this simple 1-year bullet bond with standard recovery rate of 40% and the simplified hybrid Tier I with 0% recovery rate is 0.06 (101.34 – 101.28).

Our base-case assumption in our valuation framework is that Tier I has a 0% recovery rate.

Rating States as Coupon Deferral and Extension Triggers For our valuation framework, coupon deferral and call triggers will depend on the firm being in various financial states. We will look to define issuer senior ratings which we think will represent these financial states and which can therefore be used as triggers for either coupon deferral and extension risk3.

Our valuation framework also allows investors to express a view on the ratings of a particular issuer. We think this is where the model is particularly useful in determining the valuation bounds on the instrument, i.e. where we think it should move to if the ratings are changed. In practice, we would expect to be using a different senior rating well before the rating agency actually announces a ratings change. We will explore this point in Section 5 - Sensitivity Analysis, where we indicate what the impact is on 'fair value' of a specific Tier I instrument if different assumptions are made with regard to recovery rate, coupon deferral and extension risk. Fundamentally, we believe that the valuation framework should be transparent, so as to allow investors to see the impact of the respective inputs. Importantly, the assumptions should also be debatable and able to be changed in order to suit investors’ own views.

3 The assumptions which we will be making are applicable to major Western European banks which generically are a mix of wholesale and retail-funded institutions. This is an important consideration given that we will be making assumptions as a rating state which would correspond to a particular level of solvency, and which would not be applicable to emerging market institutions, where frequently banks at the lower end of the ratings scale may indeed have Tier I ratios which are comfortably in excess of regulatory minimums.

Ratings triggers within valuation framework defined in terms of senior issuer ratings

Valuation framework should be flexible enough to adapt to investor's own views

State

No default

Default

105

40

Hybrid Tier I

Baa2 101.34

99.83%

Prob of State

Baa2 101.28

No default 105

0.1700%

Prob of State

99.83%

Senior Bond

0.1700% 0 Default

Value in State Value in State State

11




It is important to note that for the purposes of defining rating states as trigger levels within our valuation framework, we will be using the issuer senior ratings. We will therefore disregard the subordinated or issue-specific ratings which are themselves merely notched in relation to senior ratings. We believe that senior ratings are the best indicators of an issuer's credit quality, rather than the subordinated ratings which are merely the result of rating agency convention and which, in our opinion, do not necessarily capture the incremental risk of the more deeply subordinated capital instruments. For the purposes of the valuation framework we will use the lower of the S&P and Moody's senior per issuer.

Coupon Deferral As we have previously outlined, we model the probability and impact of the specific features of Tier I by making them dependent on the financial state of the issuer, and which we assume will be indicated by its rating. Specifically with regard to coupon deferral in Tier I, we will have to define a rating which will correspond to a financial state in which this event would occur. For traditional Tier I structures, there are two main types of coupon deferral events, namely coupon deferral, which occurs due to regulatory intervention, and coupon deferral, which occurs at the issuer’s initiative. Given that we believe that these coupon deferral events will not occur at the same ratings state, we will discuss them separately in terms of making assumptions as to their respective trigger levels.

Mandatory Coupon Deferral The main objective of bank regulators is to maintain confidence in the banking system and to protect the interests of depositors. A key element of regulatory oversight is the guidelines which are set in terms of minimum solvency requirements, with bank capital being the main protection that is afforded to depositors. Hence, in keeping with its objectives, we think the regulator is likely to intervene with regard to coupon deferral, only in a situation where a bank’s compliance with minimum capital ratios are under threat and the payment of coupon could further undermine regulatory solvency. Therefore to model the coupon deferral trigger in our valuation framework, we have to equate the minimum level of regulatory solvency (Tier I of 4%) to a corresponding rating state.

The approach of equating a particular rating to a trigger for one of the features of Tier I is, in our opinion, justified by the consistent approach that ratings agencies will apply in terms of their rating methodology. By identifying a particular rating, we can be sure that this will consistently correspond to an institution which has characteristics that will make coupon deferral highly likely. However, as the ratings agencies will correctly note, a rating is a blend of various characteristics of a particular issuer and which will be dependent on a wide series of factors ranging from asset quality, profitability and even some broader concepts such as franchise quality. Despite this, we think that it is defensible to equate the coupon deferral trigger level with a rating state which theoretically corresponds to minimum regulatory capital ratios defined by the regulators. We think that the main rationale for this is that at the lower end of the ratings scale, regulatory capital ratios will typically have greater importance as the buffer which ultimately protects depositors from losses.

Trigger levels for coupon deferral and extension risk will be set as a function of senior ratings

Regulatory event on coupon deferral likely if Tier I ratios at or below 4%

Consistent ratings approach by agencies supports use of ratings triggers in our framework

12




In terms of recent empirical evidence with regard to lower levels of capitalisation, we would highlight Banca Popolari Italiana (‘BPI’), which in Feb’06 suffered negative ratings action on the back of concerns with regard to its weakened financial profile. BPI was downgraded to Baa2 by Moody’s which cited amongst others ‘despite recent restoration of capital at regulatory levels, economic capitalisation remains modest’. In our opinion, the recapitalisation operations previously undertaken by BPI effectively saved it from having capitalisation ratios which would have been below the regulatory minimums, given that it reported a FY’05 Tier I ratio of 5.55%. Hence, from this example, it would appear that a bank with capital ratios in line with regulatory minimums would be represented by a senior rating below that of Baa2, using BPI as a recent, albeit rare, example of a European bank that has experienced solvency issues.

Therefore, in terms of adopting a conservative approach for our model, we will assume that the ratings state which corresponds to a Tier I capital ratio of 4% will be Ba1. We would again highlight that equating a rating state to a specific level of regulatory solvency may appear a restrictive interpretation of ratings given that they capture a broad range of financial characteristics, in addition to solvency levels. However, we would maintain that by defining Ba1 as a trigger level, this will also reflect other features that would increase the likelihood of the issuer not being in a position to pay coupons (weak financial profile, low levels of profitability or possibility of losses etc).

Therefore our base-case assumption for coupon deferral based on regulatory intervention is that for an issuer with ratings equivalent or below Ba1, regulatory intervention will occur and the issuer will not be allowed to pay coupon on Tier I instruments.

Voluntary Coupon Deferral In addition to coupon deferral which is enforced by regulatory action, we also note that issuers may, under certain circumstances, elect not to pay coupons. Such situations may arise when, for example, the terms of the Tier I instrument state that in the event of the issuer experiencing a loss in the year prior to coupon payment or if it has insufficient distributable resources, the issuer may elect to defer coupons. These ‘voluntary' coupon deferral events are more complex to represent within the valuation framework, given that regardless of the rating profile of an issuer, an unexpected loss may suddenly result in the conditions being met for the issuer to defer coupon under the terms of its Tier I issue. Naturally the probabilities of such conditions being met are greatly increased at the lower end of the ratings scale, however to define a single ratings level which would capture this outcome, is more problematic than coupon deferral motivated by regulatory intervention.

In our opinion, issuer optionality in terms of voluntary coupon deferral is actually greatly reduced and will not only depend on the objective triggers such as the lack of distributable resources in any one period, but will also be driven by a broader range of considerations which makes voluntary coupon deferral an option which we think will only be exercised in extreme circumstances.

To demonstrate our view, we will highlight two examples in which the issuers could, under the terms of their Tier I instruments, have deferred the coupon on Tier I debt but chose not to do so. The first example is the already cited BPI, which after having

Empirical evidence suggests that a bank with Tier I of 4% will have rating below Baa2

Our base case for mandatory coupon deferral is a rating of Ba1 or lower

Empirical evidence indicates low likelihood of 'voluntary' coupon deferral being exercised

13




reported a material loss for FY’05, could have, under the terms of their Tier I (BPIIM 6.742% €15), deferred on the coupon. However, with the announcement of the FY’05 losses, the issuer was very clear in communicating to the market that it would honour the coupon payment on its hybrid capital instruments. We saw a similar line adopted by HVB following record losses in 2004, when it promptly stated that it would be paying the coupon on all Tier I instruments. In our opinion, these two examples demonstrate that there are a series of considerations which effectively constrain an issuer's ability to defer coupon and which we will now consider.

The motivations for an issuer electing not to defer coupon are, in our opinion, linked to the hidden costs that such an action would imply for the issuer over the longer term. We think that the main consideration would be more restricted future market access for the issuer, given the likely reluctance of investors to participate in future Tier I deals, or only at a substantial premium. In addition, we would expect higher funding costs not only for the future issuance of Tier I instruments, but across the entire wholesale funding structure. We would expect that these costs would vary across issuers according to their degree of relative dependence on wholesale market funding, i.e. a predominantly retail-funded institution would be less vulnerable to the added costs, whilst a mainly wholesale-funded bank would incur greater increments in funding costs.

Given that we have looked at the ‘hidden’ costs for the issuer of coupon deferral, let us consider now the benefits of an issuer deferring coupon on its Tier I debt. Undeniably the major benefit will be afforded by the financial flexibility of not having the obligation to make a coupon payment when the issuer is likely to be experiencing a situation of financial distress4. In our opinion, the cost-benefit analysis of ‘voluntary’ coupon deferral will most likely imply that issuers will be reluctant to use the financial flexibility afforded by this feature of Tier I debt.

Therefore our base case assumption for ‘voluntary’ coupon deferral is that this would correspond to an issuer with ratings very close to default. We therefore think that the ratings state at which an issuer will exercise ‘voluntary’ coupon deferral will be much lower than the level at which a regulator would intervene to block coupon payment.

4 In terms of quantifying this benefit, we have used JPMorgan’s € Tier I SUSI index according to which the average coupon is 5.87% and the average issue size is €670mn, which would imply an average coupon payment of €39mn. Such an amount would strike us as scant relief for incurring the costs we have noted above.

Coupon deferral likely to imply increased costs in terms of future market access for issuer

We think coupon deferral provides limited financial flexibility for issuers

Our base case for voluntary coupon deferral is a rating close to default and lower than the mandatory coupon deferral level

14




Valuing Coupon Deferral We have defined both for mandatory and voluntary coupon deferral, the ratings state at which we believe these events will occur, with the assumption that mandatory coupon deferral occurs at Ba1, whereas voluntary coupon deferral would occur at a rating state very close to D. Therefore, in terms of capturing potential coupon deferral events, we will adopt the higher of these two trigger levels, namely the Ba1 rating state, which corresponds to a level at which the issuer will defer coupon due to the imposition of a regulatory event.

As our base-case assumption we set the Coupon Deferral trigger at the investment grade / sub-investment grade boundary, which is Ba1 in the 18-state ratings transition matrix.

Illustrating Valuation of Coupon Deferral We model coupon deferral on the basis that the Tier I will not pay coupons if its rating is at or below a ratings trigger. In Figure 6 we illustrate this using an 8-state ratings scale and have set the ratings trigger at BB, meaning that the bond will not pay a coupon if its rating is at or below BB. The bond is a 1-year bullet bond (senior rating = A) with a coupon of 5.00% and we have used a risk-free discount rate of 3.00%.

Figure 6: Valuing coupon deferral in a 1 Year Tier I bond with coupon deferral

Source: JPMorgan.

We can see how this will affect valuation, as an investor will now face an additional ‘credit risk’ to default - Coupon Deferral - which can happen earlier than default. The difference in rock-bottom price between this bond and a regular subordinated bond with no coupon deferral is 0.05 (101.63 – 101.58). Given the risk of coupon deferral, an investor will demand additional compensation for holding this bond. We can therefore value the impact of Coupon Deferral on the price of hybrid Tier I using our framework.

Extension Risk Extension risk is the risk that at the first call date or quarterly thereafter, the issuer will extend the bond by not calling it. As in the case of the coupon deferral option, we will again assume that the financial condition of the firm will be the key

Base-case assumption for coupon deferral trigger at Ba1

AAA

AA

A

BBB

BB

B

CCC D

Par + Cpn

Par + Cpn

Par + Cpn

Par + Cpn

Par + No Cpn

Default

Rating State Bond State Cashflow

105

105

105

105

100

100

100 0

Prob of State

A 101.58

Par + No Cpn

Par + No Cpn

0.04%

2.48%

90.87%

5.56%

0.72%

0.21%

0.01% 0.11%

15




determinant of extension risk. Given that we believe that mandatory extension will occur at a rating state different to where voluntary extension will occur, we will analyse each event separately.

Mandatory Extension In the section on coupon deferral, we have articulated our view on the ratings state which we believe would correspond to a point where a regulator would intervene to prevent any action by the issuer that would further undermine its regulatory solvency levels. We have defined this rating state as being equivalent to Ba1, and in terms of maintaining consistency with our earlier assumption, we will again define this as the trigger level at which a regulator will intervene and not allow an issuer to redeem a Tier I instrument at a specified call date.

Our base case assumption for mandatory extension for a Tier I instrument is that this will occur for an issuer with ratings equivalent to or below Ba1. At this point we believe that the regulator will intervene and not allow the issuer to redeem the Tier I instrument.

Voluntary Extension In contrast to coupon deferral, there is no requirement that has to be met for the issuer call option to be triggered, i.e. insufficiency of distributable funds. We would also contrast the extension option as one which is less damaging to investor interests than coupon deferral, which results in a clear situation of loss for the investor. By contrast, the fact that an issuer may not call a hybrid Tier I instrument, might not in our opinion result in a clear situation of loss for this investor, given that for a highly rated issuer, the investor will enjoy the benefit of a higher coupon stream due to the step-up post the first call date. An example of this would be any of the hybrid Tier I instruments which were issued in 2000/01, and which carry post-call spreads over 3M Libor in the 300-400bp range. We doubt that in the current low spread environment, many investors would object to a ‘AA’ rated institution not calling a bond and which would then imply a significant increment in yield. However, we doubt issuers would be willing to provide a ‘free-lunch’ to investors, and would most likely refinance these instruments at the prevailing tighter market levels.

Although the relative cost of refinancing is a worthwhile consideration, we note that for the traditional ‘step’ or innovative Tier I instruments, there is a far more powerful consideration which will in our opinion drive the issuers to call these instruments at the first call date and which is routinely referred to as the moral pressure to call. In our view, the step-up Tier I instruments have always been marketed as an instrument that the issuer intends to call. This understanding undermines the apparent flexibility that an issuer would at first sight appear to have and implies that rather than the relative cost of refinancing, the only driver that will lead an issuer to extend a hybrid Tier I instrument will be the financial state of the firm i.e. an issuer will always call a Tier I instrument as long as its rating and financial condition permit refinancing.

In the scenario where there has been a decline in the credit quality to the extent that the issuer does not have the resources to call the Tier I instrument, then the instrument is likely to be extended. This risk will be heightened by the fact that a decline in issuer’s credit quality which will (hopefully) be reflected in lower ratings, may make refinancing of the Tier I bond prohibitively onerous. We would therefore

Our base case for mandatory extension is a rating of Ba1 or lower.

'Moral obligation’ to call traditional step Tier I remains a key consideration

Tier I issuance for banks rated Baa2 or lower becomes excessively onerous

16




postulate that the ratings state at which debt refinancing becomes excessively onerous would be Baa2. For an issuer with a senior rating of Baa2, the refinancing of any Tier I instrument will imply that the ‘new’ issue will carry a sub-investment grade rating. given the rating agency approach of notching Tier I debt two-notches below the senior rating. We also believe that the issuer would in fact be accessing an investor base that would not in all likelihood be its traditional investor constituency, which would likely be unable to invest in non-investment grade instruments. We think that such a scenario is one that issuers would seek to avoid.

Therefore our base-case assumption at which voluntary extension occurs is a ratings state equal or below Baa2, where we believe the issuer will find it excessively onerous or even prohibitive to refinance.

Assumption for extension risk We have defined both for mandatory and voluntary extension the ratings at which we believe these events will occur, with the assumption that mandatory extension occurs at Ba1, whereas voluntary extension occurs at Baa2. Therefore in terms of capturing the extension risk outcomes, we will adopt the higher of the two trigger levels we have previously outlined, namely the Baa2 rating state which corresponds to a level at which the issuer will opt extend the Tier I bond. Contrary to the assumption that we have made for coupon deferral, we assume that voluntary extension will occur before mandatory extension. As our base-case assumption, we therefore set the extension trigger at Baa2.

Valuing the Extension ‘Risk’ As an illustration of how we factor in the possibility of no call / extension we can show a generic A-rated Tier I bond using an 8-state ratings matrix, with a year to go to the first Call date. For ease we will assume no coupon deferral in this example. If, in this illustration, we set the call state at BBB and above, then we can see what the cashflows will be in each rating state and the probabilities of each of these. This will allow us to calculate the 'value' of a single-A rated hybrid Tier I with a year to the call.

Figure 7: Call Trigger Illustration with call trigger at BBB rating

Source: JPMorgan

Base-case assumption for extension trigger is Baa2

AAA

AA

A

BBB

BB

B

CCC D

Call

Call

Call

Call

No Call

No Call

No Call Default

Rating State Bond State Cashflow

Par + Coupon

Par + Coupon

Par + Coupon

Perp + Coupon Recovery Rate

Prob of State

Par + Coupon

Perp + Coupon

Perp + Coupon

0.04%

2.48%

90.87%

5.56%

0.72%

0.21%

0.01% 0.11%

A

17




Pricing the Hybrid Having accounted for each feature of Tier I separately, we now move on to the simpler task of putting these together to show how we value each instrument. As with rock-bottom spreads, we start at the furthest maturity point and work our way back period-by-period through each rating state and each cashflow to arrive at a present value.

We first price the perpetual instrument using our adapted rock-bottom framework as a 50 year5 floating rate bond with deferrable coupons, a probability of calling at each quarter and a probability of defaulting and arrive at a set of values for the Tier I at the end of year 10. We then price the 10 year Tier I to calculate the current ‘fair value’. These two steps are illustrated in Figure 8.

What we are actually calculating is the Rock-Bottom Price, or maximum price an investor should be willing to pay given the multiple credit risks in this instrument. As investors are used to speaking in the language of spreads, we then convert this price into a spread to the first call date using a standard Asset Swap or Z-Spread calculator.

If the current spread of the Tier I is below the ‘fair value’ spread, then the bond looks expensive according to our valuation framework. If the current spread of the Tier I is above ‘fair value’ spread, then the bond looks cheap.

Figure 8: Valuing Tier I in Two Stages (right to Left)

Source: JPMorgan

The pricing steps are explained in greater detail in Appendix B.

In this section we have seen how we value Tier I using our framework, building in our assumptions for recovery rate, coupon deferral and call triggers. We now move on to see the results from this model and we can use these.

5 See Appendix B for an explanation of why the perpetual is modeled in this way.

AAA

AA

A

BBB

BB

B

CCC D

Rating State

Par + Cpn

Par + Cpn

Par + Cpn

BBB Perp + Cpn

BB Perp

B Perp

CCC Perp Recovery

AAA

AA

A

BBB

BB

B

CCC D

Rating State

2. Price the 10 year bond with final state values of Call, Value of Perpetual or Recovery Rate

1. Price the Perpetual to find the initial values of the perpetual at year 10

Call + Cpn

Call + Cpn

Call + Cpn

Perp + Cpn

Perp

Perp

Perp Recovery

Yr 10 State Yr 10 Value

A

18




4. Results from Valuation Framework We have applied our valuation framework with the assumptions which we outlined previously for the Tier I bonds included in JPMorgan’s SUSI, both for Euro and Sterling markets. In terms of the results, we would highlight the following trends evident in both the Euro and the Sterling markets:

• Tier I issued by highly rated entities (Aa3 and above) tends to trade cheaply in the market, i.e. the ‘fair value’ (in terms of spread) is tighter than prevailing market spreads.

• Tier I issued by entities rated A1 on aggregate are fairly priced, with the average 'fair value’ being in line with the prevailing market spreads

• Tier I issued by lower rated entities (A2 and below) tends to trade expensively in the market, i.e. the ‘fair value’ (in terms of spread) is wider than prevailing market spreads.

• On aggregate Tier I as an asset class is still slightly cheap compared to prevailing market values.

Steps-ups: Rock-bottom or ‘Fair Value’ by Rating In Figure 9 and Figure 10 we have plotted the relative cheapness/dearness of the step Tier I (for Euro and Sterling) versus the senior rating of the issuer. The relative dearness/cheapness of the bonds is determined by the difference between the current level of market spreads and the ‘fair value’ as determined by our valuation framework.

Figure 9: Euro - Fair Value versus Rating ‘Market Spread (bp) minus Rock-bottom Spread (bp)’, y-axis; Rating, x-axis

-140-120-100

-80-60-40-20

020406080

0 1 2 3 4 5 6 7 8 Source: JPMorgan, Moody’s

Figure 10: Sterling - Fair Value versus Rating ‘Market Spread (bp) minus Rock-bottom Spread (bp)’, y-axis; Rating, x-axis

-80-60-40-20

020406080

100120

Source: JPMorgan, Moody’s

We can see that Tier I instruments issued by higher rated entities are currently over-compensated for the risks they entail, i.e. they look cheap in our model. As we go down the rating spectrum, they become more expensive, with A3-rated Tier I in Euros, for example, not offering enough compensation to investors for the inherent credit risks. This pattern is consistent with credit risk pricing across the ratings

Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 A3 A2 A1 Aa3 Aa2 Aa1 Aaa

19




spectrum in the vanilla bond markets where lower rated (sub-investment grade) issues tend to look expensive relative to their Rock-bottom spreads.

In terms of the output of the valuation framework, we note that it implies that for highly rated issuers there is a very limited probability that triggers we have defined will actually be hit. In our opinion, this is due to the lower degree of ratings volatility which is displayed by banks versus the corporate sector and which we have already previously highlighted. Hence, for a highly rated issuer, the valuation framework implies a low probability of coupon deferral or extension risk outcomes as determined by the sector-specific ratings transition matrices. The fact that we derive a 'fair value' for the Tier I instruments which appears cheap relative to prevailing market levels, indicates that the market is potentially overstating the risks associated with these higher-rated instruments. Conversely, for Tier I instruments issued by lower-rated institutions, the output of the valuation framework suggests that the market is understating the risks and hence these would appear relatively expensive relative to the 'fair value' we have derived for these instruments.

Table 1: Fair Value Pricing for Benchmark € Tier I Bond Step-up

Spread (bp)

Rating Current Spread

(bp)

Rock-bottom

Spread (bp)

Excess Spread

(bp)

Relative Value

Euro - Top Five ‘Cheapest’ CDEE 4.625% €15 153 Aa2 78 23 56 Cheap CDEE 5.25% €14 184 Aa2 68 19 49 Cheap DEXGRP 4.3% €15 173 Aa2 72 24 48 Cheap LLOYDS 4.385% €17 168 Aa2 76 29 47 Cheap FORTIS 5.125% €16 200 Aa3 89 47 42 Cheap Euro - Top Five ‘Dearest’ HVB 7.055% €12 257 A2 55 110 -55 Dear CMZB 5.012% €16 215 A3 117 206 -90 Dear MTFG 4.85% €16 205 A3 112 205 -93 Dear ESPSAN 5.58% €14 265 A3 97 206 -109 Dear EURHYP 6.445% €13 367 A3 87 204 -117 Dear Sterling – Top Five ‘Cheapest’ BNP 5.945% £16 113 Aa2 167 65 102 Cheap BNP 5.945% £16 113 Aa2 167 65 102 Cheap HBOS 8.117% £10 385 Aa3 125 50 75 Cheap RABOBK 5.5656% £19 146 Aaa 118 48 69 Cheap HBOS 6.461% £18 285 Aa3 162 93 69 Cheap Sterling - Top Five ‘Dearest’ BKIR 6.25% £23 170 A1 111 139 -27 Dear BRITNA 5.5555% £15 205 A2 136 164 -28 Dear MQB 6.177% £20 235 A2 147 178 -32 Dear FSPAA 5.75% £16 192 A2 129 165 -37 Dear ANGIRI 7.625% £27 240 A2 132 185 -53 Dear Source: JPMorgan, Price as at COB 7th Sep 2006. Source for ratings: S&P & Moody’s.

We have presented a selection of results of the valuation framework in the table above for the Euro and Sterling Tier I market. For each of these markets we have listed the five 'cheapest' and 'dearest' instruments as determined by the difference between the 'fair value' and the prevailing market spread. In both markets we note that the Tier I instruments issued by highly rated issuers trade cheaply and have the most excess spread over the 'fair value' we have determined by using the valuation framework. In Appendix A we show the 'fair values' derived by our valuation framework for all the bonds included in the JPMorgan SUSI Tier I benchmark indices for the Euro and Sterling markets.

20




The Call and Coupon Deferral Values as an Option Price An intuitive way of looking at the output of the valuation framework would be to consider the value of the options which an issuer has under the terms of its Tier I instruments. Assuming that the options to extend the bond beyond call date or defer coupon will only be exercised in situations of financial distress, we can consider from that from an issuer’s perspective, the options will be more valuable the closer it is to a state of financial distress as indicated by its ratings state. Hence for a very highly rated issuer, the coupon deferral options would be out of the money, and will therefore be almost worthless, having only limited time value. We would therefore expect that for a highly rated issuer, the coupon deferral and extension options would be worth very little and potentially a Tier I would have a value in line with other instruments that do not have such features. A good example of this is the BNP 6.342% €12, which according to the valuation framework has a ‘fair value’ of L+9bp, and which is only moderately wider than the BNP 5.25% €12 (LTII), which currently trades at L+8bp. This would imply that the value of the extension and coupon deferral options from the issuer's perspective are out of the money and hence the limited spread differential between the Tier I and LTII (LTII does not have coupon deferral or extension options).

At the lower end of the ratings spectrum, the valuation framework derives fair values for the Tier I instruments which indicates that these are relatively dear in relation to prevailing market levels. Cleary, the valuation framework ascribes a greater probability than the general market to adverse outcomes in terms of extension and coupon deferral, as indicated by a ratings state which is closer to the trigger levels that we have defined. An example of this would be the BPIIM 6.742% €15, which with an issuer rating of Baa2 would be very close to the ratings triggers that we have defined for coupon deferral and extension risk.

In our opinion, the difference between the valuation framework’s 'fair value' and the prevailing market levels is the result of the framework having a more objective and quantitative approach than the market may be using. We would not maintain that the market is incorrect in its assessment of the likelihood of an AA-rated bank deferring on a coupon as being very low, however we think that this assessment would in most instances be driven by market conventions rather than objectively quantifying the inherent risk to these instruments. This is where we think that our valuation framework provides a valuable insight into the relative value of Tier I, where instead of using market conventions as to what the fair value for a Tier I issued by AA-rated bank should be, the model uses the objective and market independent approach of quantifying what would normally be assessed a ‘very limited’ risk.

Steps-ups: Rock-bottom or ‘Fair Value’ by Spread We can also look at the relationship between actual spread levels and the cheapness or dearness of Tier I instruments. We would hope to find no systematic relationship here, as that would indicate some auto-correlation or in-built prejudice in the model that would make the cheapness or dearness dependent on the actual spread.

Figure 11 and Figure 12 show the ‘excess spread’ versus the current spreads. The 'random’ nature of the distribution here shows there is no in-built model bias and that the relative value is not purely a function of the spread. One trend we may discern is that the dispersion around the mean (how far the high and low dots are from the average horizontal line) gets larger as spreads get higher. This would indicate that the

For highly-rated issuers, coupon deferral and extension options resemble ‘out-of-the-money’ options

Valuation framework objectively quantifies events which have a perceived low level of probability

Tier I instruments with higher absolute spreads have greater degree of 'mis-pricing' by the market

21




market is less ‘good’ at determining value for these higher spread instruments, where the market thinks (by definition of giving a higher spread) that there is more credit risk. We would argue that this is a reasonable outcome and moreover it is precisely where the market may struggle with how to value the credit risk in these instruments, and it is here that our model can add insight.

Figure 11: Euro - Market ASW Spread versus Fair Value ‘Market Spread (bp) minus Rock-bottom Spread (bp)’, y-axis; Spread (bp), x-axis

-140-120-100

-80-60-40-20

020406080

0 20 40 60 80 100 120

Source: JPMorgan

Figure 12: Sterling - Market ASW Spread versus Fair Value ‘Market Spread (bp) minus Rock-bottom Spread (bp)’, y-axis; Spread (bp), x-axis

-80-60

-40-20

020

4060

80100

120

0 50 100 150 200

Source: JPMorgan

Step Tier I Market: Overall Pricing Overall, the Euro market looks 9bp 'cheap' to our rock-bottom fair values and the Sterling market looks 41bp ‘cheap’. We would usually expect some excess spread to the 'rock-bottom' premium (which is the minimum spread that an investor should accept), which we might attribute to extra liquidity premium. From the standard corporate market, this is typically around 5bp. The results of the valuation framework are not entirely surprising, given that the Sterling market is comprised of generally higher rated entities than the Euro market. Hence this will imply that given the trend identified by the valuation framework that Tier I instruments issued by higher rated entities will generally look cheap, the Sterling market will look cheap relative to Euro market.

Figure 1 on the front cover shows this relative value over time. In 2001 to 2003 Tier I as a nascent asset class showed that investors required a large ‘excess spread’ to the Rock-bottom value. As the asset class became more established (aided by benchmark index inclusion), this excess value was eroded as spreads tightened to the start of 2005 when Tier I was actually expensive when it traded at its historical tights. Given the re-pricing in the sector from May 2006 onwards (driven by a surge in issuance, in our view), this excess spread has been increasing, making Tier I € look moderately cheap at these levels, according to our valuation framework.

On aggregate Euro and Sterling Tier I markets are cheap according to valuation framework

22




5. Sensitivity Analysis One of the advantages of our valuation framework is the ability to perform useful sensitivity analysis. This can be used both to identify which of the non-standard features of Tier I most contribute to the valuation of the asset class and to assess the impact on the 'fair value' if different assumptions were to be made relative to the ratings triggers. For example, an investor who would have a different view as to the rating triggers that would correspond to coupon deferral or extension risk, can use the sensitivity analysis to see the impact of making different assumptions with regard to these Tier I features. For the purposes of the sensitivity analysis, we have selected the AAB 4.31% €16 bond, which will demonstrate the interaction of the various inputs of our model and how these drive the resulting valuation.

Sensitivity to Initial (Start) Ratings Given the reliance of our valuation framework on ratings, we first look at the sensitivity of the ‘fair value’ of the Tier I instruments to the initial senior issuer rating. The sensitivity analysis indicates a large degree of sensitivity to the initial senior rating of the issuer, as shown in Figure 13. For example, a one-notch downgrade for AAB (senior rating) will imply the Tier I spread moving from 76bp to 124bp. The current rating is important, as it determines the distribution, i.e. the probability of being in each state. A one-notch downgrade will imply less chance of call, more chance of coupon deferral and more chance of default. In this sense, hybrid instruments are fundamentally higher beta, as any decline in credit quality has this 'triple whammy' effect on the different credit triggers, whereas a standard bond will just get closer to default. Investors who disagree with the rating agencies can use whatever rating they believe is most appropriate and can use this to see where the bounds on valuation should go to if a bank is upgraded or downgraded by the rating agencies.

Figure 13: AAB 4.31% €16 Stress Test of Initial Ratings, Deferral Trigger at Ba1, Call Trigger at Baa2. Rock-bottom Spread (bp) for different senior ratings. Current senior rating = A1 (dotted box).

4 17 24 44 76124

199279

407458

0100200300400500

Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3

Source: JPMorgan

Sensitivity to Coupon Deferral Trigger Rating The coupon deferral trigger is the rating at or below which the company will exercise coupon deferral on a Tier I instrument within our valuation framework. We can see the Rock-bottom Spreads for the AAB 4.31% €16 in Figure 14 with different coupon deferral trigger assumptions. Raising your view at which rating state the company will defer coupon has a large impact on the ‘fair value’ spread. If we set the coupon deferral at a very high level, for example at Aa1, then this will imply that there is an

Initial issuer senior ratings are key driver in 'fair values' for Tier I instruments

23




extremely high degree of certainty that the issuer will indeed defer coupon on the Tier I instruments, which in turn will result in a low valuation in terms of bond price and an associated wide spread for that instrument. Figure 14 shows that the Rock-bottom spread for a coupon deferral trigger of Aa1 is 796bp – this is because this hybrid would be almost certain to not pay coupons and the value would then just depend on getting par at a call date.

On the other hand, if we set the coupon deferral at a relatively low rating state of Ba1, then this decreases the likelihood of the issuer deferring coupons, which will tend to result in a higher valuation (bond price) and lower associated spread. The output of the sensitivity analysis allows us to determine the relative importance of coupon deferral as a driver of ‘fair value’ within our valuation framework. In addition, the sensitivity analysis allows us to see what is the impact of making a different assumption with regard to the coupon deferral trigger level, whilst maintaining the same assumptions for the call trigger and recovery rate. Hence, if we had assumed a coupon deferral trigger level of B1, then our 'fair value' spread would have been 45bp, rather than the 76bp derived for our base assumption of coupon deferral at Ba1.

Figure 14: AAB 4.31% €16- Stress Test of Coupon Deferral Triggers, Call Trigger at Baa2 Rock-bottom Spread (bp) for different coupon deferral trigger ratings. Our base case = Ba1 (dotted box).

796 796 791 774668

243158 110 93 85 76 62 51 45 39 33 31

0

300

600

900

Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1

Source: JPMorgan The shape of the sensitivity analysis curve derived above demonstrates that the lower we set the ratings trigger for coupon deferral, the less likely the event becomes and hence less compensation the investor will require for this risk. At the lower end of the ratings state, the differences between setting the triggers have a lower marginal impact given that the likelihood of a highly rated issuer reaching these levels is already so limited that the resulting pricing differential is therefore equally limited.

Sensitivity to Call Trigger Ratings The shape of the sensitivity analysis curve for the call trigger reflects the possible valuation scenarios between the two extreme outcomes i.e. the bond extends with certainty or the bond is called with certainty. We have shown this for our AAB example in Figure 15. In the first scenario, by setting the call trigger at a very high rating state (e.g. Aaa), we have a higher degree of certainty that the bond will actually extend. Hence, setting the call trigger at ‘Aaa’ implies that we will therefore be pricing a perpetual bond. This will have a marked impact on the output of our valuation framework and will result in a 'fair value' spread which is wider than the ‘fair value’ derived for our base case of Baa2. If we set the call trigger at a lower rating state, such as Caa1, then there is a high degree of certainty that the bond will be called, in which case we are merely deriving a ‘fair value’ for a ten-year bond. Therefore the shape of the curve merely reflects the fact that as we lower the level of

24




call trigger, we are increasing the probability of the bond being called, until at the very limit we are merely pricing a ten-year instrument.

Figure 15: AAB 4.31% €16 - Stress Test of Call Triggers, Deferral Trigger at Ba1 Rock-bottom Spread (bp) for different call trigger ratings. Our base case = Baa2.

156 159 166 169151

12298 86 76 70 55 45 39 33 31 30 29

0

100

200

Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1 Source: JPMorgan

The sensitivity analysis also allows us to rank the non-standard features of Tier I in terms of which has the most impact on 'fair value', and which should therefore command a greater risk premium from the investor's perspective. Contrary to coupon deferral, bond extension does not appear to be unequivocally negative from the investor's perspective, given that with the associated step-up in coupon, investors would potentially be in the enviable position of receiving a higher coupon, provided that this is not accompanied by a marked deterioration in the issuer's credit quality. Coupon deferral, on the other hand, is an unequivocal negative from the issuer's perspective given the loss of a non-cumulative coupon and where there is no prospect of compensation, such as the coupon step beyond the first call date.

As we have argued, there is an interaction between our assumptions on coupon deferral trigger and call trigger that need to be considered. This means that ‘valuing’ how much each feature is worth on its own necessarily means making an assumption about the other6. We could represent this as in Table 2, where for each stress test we keep our other assumptions constant to isolate what the value of a 1-notch stress is for each feature. With our base-case assumptions, changing our view on what the company's senior rating should be (i.e. how we see its credit quality now) has the largest impact on the value, with an additional 48bp of spread needed for a 1-notch move in the senior rating.

Table 2: Stress Test Summary, With Base-Case Assumptions for AAB Tier I. Snr rating = A1. Rating Rock-bottom Spread (bp)

Stress Test Base Case 1-notch

worse Base Case 1-notch worse Difference Senior Rating A1 A2 76 124 48 Coupon Deferral Trigger Ba1 Baa3 76 85 9 Call Trigger Baa2 Baa1 76 86 10

Source: JPMorgan

Sensitivity to Recovery Rates The recovery rate assumptions have an impact on the 'fair value' derived in the situation of default for the issuer. Given that we have already highlighted the fact that banks generally have a low default probability, we would therefore expect the

6 In fact, we will also have to keep our assumption about Recovery Rates constant as well.

25




recovery rate to have a relatively limited impact on the ‘fair value’. Hence, varying the recovery rate within the valuation framework will only have a relatively limited impact on the range of value ‘fair value’ that will be derived in the process. This is shown in Figure 16, where we can see that a 10% rise in recovery rate has around a 4bp effect on ‘fair value’. Bearing this in mind, we are comfortable in making an assumption of a 0% recovery rate within the valuation framework, which might be conservative but which we know will also have a rather limited impact on the 'fair value' produced by the valuation framework.

Figure 16: AAB 4.31% €16- Stress Test of Recovery Rates for 1 (Snr A1), Deferral Trigger at Ba1, Call Trigger at Baa2

76

72

68

64

61

57

50

60

70

80

0 10% 20% 30% 40% 50% Source: JPMorgan

In summary, the ability to look at the sensitivities of our valuation framework and stress test each hybrid is a great strength. This is particularly useful as we change our view about the fundamentals of a company and need to see where we believe ‘fair value’ in the Tier I bonds should go to. We will see more of this in section 7, where we look at Historical Analysis.

26




6. What about the Non-steps? One of the more recent developments in the bank capital market has been the emergence of the non-step Tier I instruments, which differ from the traditional Tier I structures in that they contain no economic incentive for the issuer to call them at the pre-determined call dates. Crucially, the non-steps allow issuers to overcome the ‘15% of Tier I’ ceiling which is applicable to traditional ‘step’ Tier I instruments. As a result, we have seen a steady increase of non-step issuance, as banks increasingly seek to leverage up their capital bases in an efficient manner.

One of the advantages of the valuation framework that we have presented thus far is that we can apply it in the derivation of ‘fair values’ for the non-step Tier I instruments. This can be done by changing the assumptions we initially made for the 'step' Tier I instruments so that the valuation framework can be adapted to different structure of the non-steps, as well as the different expectations that the investor community has with regard to this asset class. Specifically, we will only adjust our assumptions in relation to extension risk, whilst maintaining unchanged our initial assumptions with regard to subordination and coupon deferral, as these will not, in our opinion, be impacted by the absence of a step feature.

So, will they be called? Before we set our stall in terms of our assumptions for the non-step Tier I instruments, we will review an important consideration with regard to this increasingly prevalent asset class – that issuers have, in our opinion, a much lower 'moral obligation' to call these instruments. For traditional ‘step’ Tier I instruments, there has always been an understanding that the issuers would in fact call at the first call date.

In our view, the reality for the non-steps is completely different, with the issuers having a much lower ‘moral obligation’ to call these instruments. Not only do these contain no economic incentive to be called, but in addition the issuers have not been allowed to assume any commitment, either explicitly or implicitly, that they would in fact call these instruments. This is a key consideration in our analysis and valuation of these instruments within the framework – we will assume that in the absence of any moral obligations, the issuer’s decision to call a non-step will be exclusively driven by refinancing costs. Simply put, we believe that an issuer will only call a non-step Tier I instrument if it can be refinanced at a more attractive level. By assuming the absence of any moral pressure to call the non-step instruments, this implies that the issuer will not incur any of the ‘collateral costs’ in terms of future market access and which represented a strong catalyst to call the traditional ‘step’ instruments.

Our base assumption is therefore that the probability of a non-step instrument being called will purely be a function of the relative refinancing costs: if an issuer can refinance more cheaply, then he will call the bond, and in a scenario where refinancing costs are higher, the issuer will extend the non-step beyond the call date.

Issuers face limited 'moral pressure' to call the non-step Tier I instruments

We assume decision to call non-step Tier I instruments driven only by refinancing costs

Base-case assumption: non-steps called only if they can be refinanced more cheaply

27




Refinancing Costs The refinancing costs for an issuer at a particular point in time will, in our opinion, be driven by two factors, namely

• The issuer’s financial condition at that point, as indicated by its credit rating.

• The prevailing market spread level at that point.

The interaction of the above two factors will determine whether an issuer will call a non-step and issue a replacement instrument, or whether it will be cheaper to merely extend the current non-step beyond the call date. In the event that an issuer opts to extend, the absence of any moral obligation implies that there should be no ‘collateral costs’ in terms of future market access for that issuer.

In terms of incorporating the impact of refinancing costs into our valuation framework, we will have to represent an issuer's cost of refinancing as a function of its senior ratings. In order to do this, we establish what the mid-cycle refinancing costs are for the non-step Tier I instruments by each senior rating category. As an example, if we establish that the average mid-cycle funding costs for a non-step Tier I issued by an Aa2 rated issuer corresponds to L+120bp, then we can infer that an issuer with a non-step Tier I with a post-call spread of 130bp will only call this bond if its senior rating remains equal to or above Aa2. In the event that the issuer's senior rating should fall below Aa2, this will imply the company would pay more than L+130bp to refinance the non-step Tier I and according to our initial assumption, the issuer will merely elect to extend the existing non-step.

Hence, in applying our valuation framework to the non-step Tier I instruments, the ratings triggers for extension risk will be determined by a combination of the post-call spread and the mid-cycle refinancing costs. This approach is different to our approach followed in valuing the traditional step Tier I instruments, where we assumed a consistent ratings trigger for extension risk. For the non-step Tier I instruments, the ratings trigger will be set as a function of the post-call spread and which will vary per instrument7.

Our base-case assumption for extension risk with the non-step Tier I instruments is that the probability of being called is merely a function of refinancing costs and which will be determined by the post call spread and prevailing market conditions. Hence, each individual bond will have its own trigger level for extension risk and which will be determined by its post-call spread.

Equating funding costs to rating states In terms of the assumptions we have outlined, the mid-cycle refinancing costs for non-step instruments are required as an input to the valuation framework. This is not without its challenges, given that the non-step market is still relatively nascent and the available spread data would not necessarily be representative of mid-cycle levels. 7 The treatment of the non-steps is therefore much more similar to the way we view corporate hybrids, where we assume the call is purely an economic decision and depends whether the post-call spread is higher or lower than the expected refinancing rate.

28




Additionally, we note that the issue population remains relatively reduced, which makes the task of deriving mid-cycle refinancing costs across the various rating states more challenging. (This will improve over time as the market develops and more data is available).

Despite these limitations, we have used the available historical data for the non-step Tier I instruments and determined average funding costs for each of the ratings categories – this is shown in Table 3: Average Non-Step Tier I funding levels by Issuer Senior rating. Due to the restrictions imposed by only having a limited data set available, we have in certain instances used estimates or interpolation to derive average funding levels across all the ratings categories.

Table 3: Average Non-Step Tier I funding levels by Issuer Senior rating bp Rating Average Funding Costs - Euro Average Funding Costs - Sterling Aaa *80 *90 Aa1 105 110 Aa2 120 128 Aa3 *127 134 A1 135 141 A2 160 166 A3 187 *185 Baa1 *210 *200 Baa2 250 *230 Baa3 *300 *270 Source: JPMorgan. * Estimated levels due to lack of data points.

Results for the € non-steps Having applied the above financing costs per rating category to the non-step Tier I instruments in order to determine the specific ratings trigger, the valuation framework has produced the results in Figure 17and Figure 18, for both the € and £ markets. For both the € or £-denominated non-step market, the valuations produced by our framework do not indicate that the market is cheap or dear ‘per se’, but rather produces a mixed set of results which are driven by a combination of the issuer’s senior ratings and the level of the post-call spread of the instruments. Hence, issues which have relatively generous post-call spreads and/or where the issuer is relatively highly rated (DRSDNR and CDEE) appear cheap according to our valuation framework. However, non-steps which have been issued by relatively lower-rated entities and have relatively tight post-call spreads (CAVALE and AIB) appear relatively dear according to our valuation framework.

29




Figure 17: € Non-Steps Valuation Results Market Spread minus Rock-bottom Spread (bp)

-93

12

53

14

69

-120.00

-80.00

-40.00

0.00

40.00

80.00

AIB 4.781%€14

BACR 4.875%€14

BACR 4.75%€20

CRLOG 4.60% DRSDNR6.35% €17

Source: JPMorgan

Figure 18: £ Non-Steps Valuation Results Market Spread minus Rock-bottom Spread (bp)

-46 -41

38

112

434 48

-80.00-40.00

0.0040.0080.00

120.00160.00

ALLNCE6.222%

£19

AIB6.271%

£16

ANGIRI6.25% £15

BACR 6%£17

HBOS6.0884%

£15

HBOS6.3673%

£19

NRBS6.8509%

£16 Source: JPMorgan

Non-Step - Conclusions We fully expect that our valuation approach applied to the non-steps and the resulting conclusions are likely to be the most controversial aspect of the valuation framework which we have presented in this note. This, in our opinion, would be a reflection of the fact that even amongst the investor community, there is no consensus as to whether the non-steps will be called or not. Therefore, such a divergence in views will necessarily imply an equally divergent outcome in terms of valuations.

The advocates of the view that the non-steps will not be called refer several reasons for their argument, namely;

a) Moral pressure to call: Traditional ‘step’ Tier I have always been marketed as instruments that the issuer intends to call, whereas this is not the case with the non-steps, where the regulator forbids the issuer from making such commitments. Hence, there will be no moral pressure on the issuer should he choose not to call a non-step Tier I at the first call date.

b) Refinancing levels: The post-call spread on the majority of the non-steps represents very attractive funding levels for the issuers, and which is a result of the issuance having been done in a relatively tight spread environment. Hence, the probabilities of the issuer being able to refinance the non-step more cheaply at a future point in time would appear somewhat limited.

c) Flexibility: The majority of the non-steps have an issuer call every three months following the initial call date. This flexibility gives corporate treasurers a very attractive option, which allows them to opportunistically call the non-steps and reissue, if and when they are able to get cheaper financing at any point beyond the initial call date. Hence, the issuer will not be immediately locked into a fixed funding cost for a period of 5 or 10 years as they would be in a scenario where they called the non-steps and issued a replacement non-step.

30




We do not disagree with the any of the above, and in fact our approach to the valuation of the non-steps incorporates some of the rationales which are advanced in order to substantiate the view that these will not be called.

Fundamentally, we believe that the valuation framework that we have proposed for the non-steps is a transparent and objective method to value these instruments, where we probabilistically determine a wide range of scenarios in order to derive a final valuation. We think that this approach is a more satisfactory alternative to the bi-polar views that these instruments should be priced either as 10-year or perpetual instruments.

Having seen how we look at Non-Steps, we finish our valuation analysis by looking at some results from the framework over time.

31




7. Historical Analysis The valuation framework which we have outlined provides fair values for benchmark Tier I, both in Euro and Sterling, based on the current valuations and assumptions that we have previously outlined. In terms of extending this analysis further, we have made use of available historical Tier I spread data, in order to test the robustness of the valuation framework we have developed8. In essence, this back-testing of the model involves generating a time series of the 'fair values' as determined by the model and comparing these to the actual historical spread data for the specific security. Given that issuer senior ratings are important drivers of the fair values produced by the valuation framework, we will highlight two issuers that have experienced experienced a certain degree of ratings volatility over the time-frame analysed.

BAVB 6.988% €11 Banca Intesa experienced negative ratings action in 2002 due to credit losses stemming from exposure to large corporate defaults and some LATAM exposure. Following the implementation of a successful restructuring, Banca Intesa has more recently recovered its Aa3 rating from Moody’s.

Figure 19: BAVB 6.988% €11 spread versus ‘fair value’ Bp

0

50100

150

200

250

300350

400

Jan-02 Sep-02 May-03 Jan-04 Sep-04 May-05 Jan-06 Sep-06

A1 A2 A3 Market Spread

0

50100

150

200

250

300350

400


A1 A2 A3 Market Spread

0

50100

150

200

250

300350

400


A1 A2 A3 Market Spread Source: JPMorgan.

Figure 19 plots how the 'fair values' would have evolved through time as the rating changed and compares this with the actual historical values for the BAVB 6.988% €11. Both data sets follow a similar overall pattern, however the actual market spreads are more reactive than the 'fair values' which have been generated by our valuation framework. This is unsurprising and is the result of the market 'pricing in' more quickly any negative or positive newsflow on an issuer, whereas the ratings agencies (given their modus operandi) will generally be less reactive to newsflow. Hence, given that the valuation framework is driven by the senior ratings, the time-

8 We hope to perform a more complete back-test at a later stage, to show how our valuation framework performs as a trading signal.

32




lag in taking ratings action will result in a deviation of market value and ‘fair value’. What the model is particularly useful for, is for seeing the valuation bounds on the instrument, i.e. where we think it should move to if the ratings are changed. In practice, we would expect to be using a different senior rating well before the rating agency actually announces a ratings change.

In all of the valuations that we have derived using our valuation framework, we have used the senior ratings for the issuer, specifically using the lower of the Moody’s and S&P rating. However, in the event that an investor may have a view on the rating of an issuer which may not be aligned with that of the ratings agencies, then this can applied to the valuation framework in terms of deriving a fair value for the Tier I.

CRDIT 5.396% £15 Following its acquisition of HVB, CRDIT was subject to expected ratings downgrades due to the scale of the acquisition and the weaker credit fundamentals of its target. Figure 20plots how the 'fair values' would have evolved through time as the rating changed and compares this with the actual historical values for the CRDIT 5.396% £15. This CRDIT £ Tier I is a good example of how the valuation model can provide insight into the valuation of the asset class, particularly when the investor has a view on the ratings of the issuer different to that of the agencies. Figure 20shows the market spread as already ‘pricing-in’ a two-notch downgrade before November 3, 2005, when Moody's downgraded CRDIT from Aa2 to A1. This ratings action was entirely expected by the market in the wake of the HVB acquisition, hence an investor using the expected A1 rating prior to the rating agency downgrade would have been able to derive 'fair value' which would have been more commensurate with the issuer's financial condition and more in line with prevailing market spreads.

Such a view would have told you that spreads should be unchanged on an actual ratings downgrade, as this was already priced-in to current spreads. This is exactly what happened, with no negative spread reaction once the downgrade was announced – the Tier I bond was already trading close to the model spread for a two-notch downgrade. This examples underscores the usefulness of the valuation framework in that it allows the investor to determine the valuation bounds on the instrument, i.e. where we think spreads should move to if the ratings are changed.

Figure 20: CRDIT 5.396% £15 spread versus ‘fair value’ Bp

405060708090

100110

120130140

Oct-05 Dec-05 Feb-06 Apr-06 Jun-06 Aug-06

Aa2 A1 Market Spread

405060708090

100110

120130140


Aa2 A1 Market Spread

405060708090

100110

120130140


Aa2 A1 Market Spread Source: JPMorgan.

33




Appendix A – Full Results of the Valuation Framework Table 4: Fair Value Pricing for Benchmark € Tier I Bond Step-up Spread (bp) Rating Current Spread (bp) Rock-bottom Spread (bp) Excess Spread

(bp) Relative Value

CDEE 4.625% €15 153 Aa2 78 23 56 Cheap CDEE 5.25% €14 184 Aa2 68 19 49 Cheap DEXGRP 4.3% €15 173 Aa2 72 24 48 Cheap LLOYDS 4.385% €17 168 Aa2 76 29 47 Cheap FORTIS 5.125% €16 200 Aa3 89 47 42 Cheap UBS 4.28% €15 158 Aa2 63 22 41 Cheap BBVASM 3.798% €15 165 Aa3 84 44 41 Cheap RZB 5.169% €16 195 A1 118 81 38 Cheap LLOYDS 6.35% €13 250 Aa2 50 13 36 Cheap BNP 5.868% €13 248 Aa2 49 13 36 Cheap LLOYDS 7.375% €12 233 Aa2 45 10 36 Cheap FORTIS 4.625% €14 170 Aa3 74 40 34 Cheap BNP 6.342% €12 233 Aa2 41 9 31 Cheap RBS 4.243% €16 169 Aa3 70 39 31 Cheap HBOS 4.939% €16 173 Aa3 78 47 31 Cheap SOCGEN 4.196% €15 153 Aa3 71 41 30 Cheap ACAFP 4.13% €15 165 Aa3 72 45 28 Cheap NYKRE 4.901% €14 170 Aa3 67 39 28 Cheap DB 5.33% €13 199 Aa3 61 33 28 Cheap HSBC 5.3687% €14 200 Aa3 62 36 26 Cheap SVSKHB 4.194% €15 168 Aa3 71 45 26 Cheap BACR 7.25% €10 295 Aa2 31 6 25 Cheap HSBC 5.13% €16 190 Aa3 70 46 23 Cheap BNP 6.625% €11 260 Aa2 32 8 23 Cheap IBSANP 8.126% €10 350 Aa3 37 15 22 Cheap SOCGEN 5.419% €13 195 Aa3 56 34 22 Cheap HSBC 8.03% €12 365 Aa3 46 26 20 Cheap ACAFP 7.047% €12 261 Aa3 44 25 19 Cheap KNFP 8.32% €10 350 Aa3 31 13 19 Cheap HBOS 7.627% €11 288 Aa3 41 23 19 Cheap SOCGEN 7.875% €10 295 Aa3 29 11 18 Cheap RBS 6.467% €12 210 Aa3 45 27 18 Cheap FORTIS 6.5% €11 237 Aa3 39 21 18 Cheap CRDIT 4.028% €15 176 A1 91 78 13 Cheap LEH 3.875% €11 160 A1 58 50 8 Cheap AAB 4.31% €16 166 A1 86 80 6 Cheap AIB 5.142% €16 198 A1 87 81 6 Cheap BFCM 4.471% €15 185 A1 83 78 5 Cheap DRSDNR 5.79% €09 50 A1 33 34 -1 Dear HSHN 7.4075% €14 215 A2 120 126 -6 Dear BAVB 6.988% €11 260 A1 43 53 -10 Dear CRDIT 8.048% €10 325 A1 36 46 -10 Dear AIB 7.5% €11 333 A1 37 50 -13 Dear CS 7.974% €10 320 A1 30 43 -14 Dear BKIR 7.4% €11 326 A1 35 50 -15 Dear CS 6.905% €11 320 A1 35 55 -20 Dear BCPN 5.543% €14 107 A2 90 113 -23 Dear BCPN 4.239% €15 195 A2 102 128 -25 Dear HVB 6% €08 0 A2 28 65 -37 Dear KRCP 8.22% €09 390 A2 35 78 -43 Dear STANLN 8.16% €10 380 A2 35 84 -50 Dear MONTE 7.99% €11 390 A2 46 96 -51 Dear HVB 7.055% €12 257 A2 55 110 -55 Dear CMZB 5.012% €16 215 A3 117 206 -90 Dear MTFG 4.85% €16 205 A3 112 205 -93 Dear ESPSAN 5.58% €14 265 A3 97 206 -109 Dear EURHYP 6.445% €13 367 A3 87 204 -117 Dear Source: JPMorgan, Priced at COB 7th Sep 2006. Source for ratings: S&P & Moody’s.

34




Table 5: Fair Value Pricing for Benchmark £ Tier I Bond Step-up Spread (bp) Rating Current Spread (bp) Rock-bottom Spread (bp) Excess Spread

(bp) Relative Value

BNP 5.945% £16 113 Aa2 167 65 102 Cheap BNP 5.945% £16 113 Aa2 167 65 102 Cheap HBOS 8.117% £10 385 Aa3 125 50 75 Cheap RABOBK 5.5656% £19 146 Aaa 118 48 69 Cheap HBOS 6.461% £18 285 Aa3 162 93 69 Cheap SANTAN 6.984% £18 186 Aa3 157 93 64 Cheap CS 8.514% £15 440 A1 167 116 51 Cheap LLOYDS 7.834% £15 350 Aa2 111 60 51 Cheap NRBS 8.399% £15 472.5 A1 165 116 49 Cheap HSBC 5.862% £20 185 Aa2 118 78 40 Cheap HBOS 7.286% £16 364.5 Aa3 115 83 32 Cheap ACAFP 5.136% £16 157.5 Aa3 113 85 28 Cheap DANBNK 5.563% £17 144 Aa3 111 90 21 Cheap NAB 5.62% £18 193 Aa3 115 96 19 Cheap AAB 5% £16 148 Aa3 104 85 19 Cheap RBS 7.387% £10 138 Aa3 70 55 15 Cheap SVSKHB 5% £10 176 Aa3 68 55 14 Cheap NRBS 7.053% £27 183.5 A1 153 144 9 Cheap SANTAN 7.037% £26 375 Aa3 119 112 7 Cheap CRDIT 5.396% £15 176 A1 122 118 4 Cheap ALLNCE 5.83% £16 213 A1 117 119 -3 Dear STANLN 8.103% £16 427.5 A2 167 171 -4 Dear CMZB 5.905% £18 216 A2 156 175 -19 Dear NWIDE 5.769% £26 208 A1 120 140 -21 Dear KBC 6.202% £19 193 A1 111 132 -21 Dear BKIR 6.25% £23 170 A1 111 139 -27 Dear BRITNA 5.5555% £15 205 A2 136 164 -28 Dear MQB 6.177% £20 235 A2 147 178 -32 Dear FSPAA 5.75% £16 192 A2 129 165 -37 Dear ANGIRI 7.625% £27 240 A2 132 185 -53 Dear Source: JPMorgan, Priced at COB 7th Sep 2006. Source for ratings: S&P & Moody’s.

35




Appendix B – Detailed Workings of Our Hybrids Valuation Framework This Appendix presents a detailed work-through of the steps we take to value our financial hybrids, with some simplifications to aid explanation. The aim is that a reader with enough time and a spreadsheet should be able to build their own rock-bottom spread calculator to value corporate hybrids.

A Recap on Rock-Bottom Spreads9 The rock-bottom spread methodology starts from the assumption that the extra return paid to corporate bond holders in spread terms should compensate them for the risk of default. The methodology seeks to find the minimum or ‘rock-bottom’ spread that an investor should be willing to accept for taking exposure to this credit risk.

For example, if we want to find the rock-bottom price (or spread) of a simple 1-year A-rated bullet bond, we take the current rating of the bond and use the ratings transition matrix to calculate the probability of being in each rating state at the maturity of the bond in 1 year’s time (as shown in Figure 21).

Using the values in each future rating state and the probability of being in each state (with a volatility adjustment) we can calculate the rock-bottom price and spread of a bond. Our valuation of financial hybrids will work in this way also – what is new is our extension of the possible future states as coupon deferral and call states also need to be considered.

Figure 21: Rock-Bottom Price Calculation for an A-Rated Bond with 1 Year to Maturity

Source: JPMorgan

The rock bottom spread first calculates the average price across scenarios to calculate a “breakeven” spread, but then also adjusts for price volatility across scenarios10 (where each rating state is a ‘scenario’), using the formula in Equation 1. 9 For a full description of the methodology, see ‘Valuing Credit Fundamentals: Rock-Bottom Spreads’, P Rappoport, October 2001

AAA

AA

A

BBB

BB

B

CCC D

Rating State Value in State

105

105

105

105

105

105

105 40

0.04%

2.48%

90.87%

5.56%

0.72%

0.21%

0.01% 0.11%

Prob of State

A-rated 104.93 (Ave Price Across Scenarios)

36




Equation 1: Rock-Bottom Price Calculation

Return) Free Risk

coreDiversityS ScenariosAcross Vol Price*Ratio nInformatio Target ScenariosAcross Price Ave

Price Bottom-Rock+

−

=1(

For example, for the 1-year A-rated bond (illustrated above), paying a coupon of 5% and with a recovery rate of 40%, where the risk free rate is 4.00%:

Extending the Rock-Bottom Framework: A Simplified Example A simple illustration of how this works for financial hybrids will serve to round-off our explanation of the intuition behind our valuation framework (a more comprehensive step-by-step explanation is given later on).

If we consider a corporate hybrid in year 9, i.e. with 1 year to go to the first call date, we will want to find the probability of each rating state and value in each rating state. The extension from our rock-bottom example before is that we now have more than just two ‘value in states’ (Default / No Default) to consider. The states we need to consider for year 10 are:

Call: At a certain rating state the bond will be called.

Perpetual: If the bond is not called, then if it is not deferring its coupons it will be worth the value of the perpetual and will pay a coupon.

Perpetual With No Coupon: If it is not called and the rating is sufficiently low, it will be worth the value of the perpetual and will not pay a coupon (it will defer).

Default: If the company defaults, it pays the recovery rate.

Figure 22 illustrates how these states work out:

10 For Financials hybrids we are not using a risk-neutral framework and include a risk (volatility of scenarios) adjustment in the calculation. This is most appropriate for an instrument that is established and is held within a portfolio of instruments subject to credit risk. Calibration to market pricing shows that the market (as we would expect) demands compensation for the risks of Tier I bonds and is not pricing these in a risk-neutral way.

.100.77%)00.41(

7015.2*5.093104

=+

−

=− Price BottomRock

37




Figure 22: Fair Value Price for a Hybrid with 1 Year to First Call Date, Senior Rating = A

Source: JPMorgan

Using this method we are able to value all the features of a financial hybrid in a single, intuitive framework. What we will arrive at is the ‘fair value’ price of a financial hybrid given its cashflow triggers. This valuation framework is transparent and allows investors to input their own assumptions in order to price hybrids.

We now turn to a full step-by-step explanation of the stages to value financial hybrids. The explanation is intended to be comprehensive, so readers not wanting this level of detail should stop here – you’ve done well enough by getting to the middle of Appendix B.

Steps to Value Financial Hybrids11 We value the financial hybrids in two steps: valuing the perpetual and valuing the 10 year bond.

A) Valuing the Perpetual To value the perpetual, we need to start at the last year and work our way backwards through each state calculating the value in and probability of each state. This would seem problematic with a perpetual, since by its nature it has no final state.

The way we model the perpetual is by valuing a 50-year bond starting in 10 years. The reason for this is that the effect of call triggers and defaults creates absorbing states at each end of the transition matrix. For example, if the bond defaults it does not come out of the default state (it is absorbed) and if it reaches Baa2 it is called and also does not come out of that state. Looking at the state of any bond after 60 years (10 + 50 years) with a transition matrix of this kind gives close to zero probability of being anything other than Called or Defaulted state at year 60. See Figure 23 for an illustration of this for an Baa3 rated Tier I hybrid.

11 For a good understanding of the mechanics of rock-bottom spread calculations it is worth reading the Rock-bottom spread mechanics, P Rappoport, Oct 2001.

AAA

AA

A

BBB

BB

B

CCC D

Call

Call

Call

Perp + Cpn

Perpetual

Perpetual

Perpetual Default

Rating State Bond State Value

Par + Coupon

Par + Coupon

Par + Coupon

Value of BBB starting Perp + Cpn

Value of BB starting Perp

Value of B starting Perp

Value of CCC starting Perp Recovery Rate

Prob of State

Fair Value Price

0.04%

2.48%

90.87%

5.56%

0.72%

0.21%

0.01% 0.11%

38




Figure 23: Probability of Being in Each Rating State at Year 60, Using a Transition Matrix With Absorbing States for Default and For Calls

0% 0% 1% 3% 3%

17%

4%

24%

39%

0% 0% 0% 0% 0% 0% 0% 0%9%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%


Source: JPMorgan

So, we can value the Perpetual as a risky 50 year floating rate bond, with deferrable coupons at certain rating states and a Call at certain rating states12.

Year 50 States to Value Year 49 We start in the final year of our “50 year” perpetual and look at the value in each rating state. We will use the BPIIM 6.742% €15 to give the characteristics in this work-through, using an 8-state transition matrix. We will use a senior rating of BBB (as if BPIIM senior rating was BBB) in order to illustrate the work-back process. The base case inputs are therefore as follows:

Table 6: Base Case Assumptions for Worked Calculations Base Case Assumptions Recovery Rate 0% Coupon Deferral BB Call Trigger A Information Ratio 0.5 Diversity Score 70 Coupon (to year 10) 6.742 Step-up Spread (bp) 525 Flat Risk Free Rate 4.00% Senior Rating BBB Source: JPMorgan

For illustration purposes, we will show the workings using annual payments and with the annual 8-state historical transition matrix (in reality the bond will pay quarterly payments, with quarterly call options). We will use a standard transition matrix (which includes corporates) in these examples – in reality we use a Financials-only transition matrix for Financial hybrids. Non-annual steps introduce several practical complications which we ignore for the purpose of explanation13.

Figure 24 shows the values in each final state and the probabilities of each state. We use the rock-bottom price calculation in Equation 1 to find the (clean) rock-bottom price of a (senior rated) BBB hybrid of 100.44. We calculate this breakeven price in year 49 for each rating in the same way as we have shown for a BBB-rated hybrid.

13 The most challenging of these is the use of non-annual transition matrices. In our internal modelling we adopt the “generator matrix” approach of Israel, Rosenthal and Wei (Mathematical Finance, Vol.11, No.2, April 2001) to create non-annual transition matrices.

39




Figure 24: Year 50 States and Pricing an BBB-Rated Hybrid in Year 49

Source: JPMorgan

Using Equation 1 from before, we get 100.44 with the following inputs: Average price across scenarios = 104.73 Information Ratio = 0.5 Volatility of Prices = 4.63 Diversity Score = 70 Discount Rate = 4.00%

In Year 50, the Call / No Call trigger does not come into play as the bond will pay Par + Libor + Spread whether it is called or not, as it is maturing. We now move onto year 49 where this call trigger will come into play.

Year 49 States to Value Year 48 In Year 49, the ‘values in state’ are the clean Rock-bottom prices we calculated looking forward to Year 50, plus coupons where these are paid. We can now use these Year 49 values to calculate the values for year 48. E.g. the BBB 'value in state' = 100.44 + 5.25 = 105.62. Again we use the transition matrix to determine the probability of being in each state. Our Call trigger now comes into play, as the hybrid will be called at ‘Par + Libor + Spread’ if it is above the call rating trigger. Figure 25 shows the future year 49 states and values.

Figure 25: Year 49 States and Valuing a BBB Rated Bond in Year 48

Source: JPMorgan

AAA

AA

A

BBB

BB

B

CCC D

Mature

Mature

Mature

Mature

Deferral

Deferral

Deferral Default

Rating State Bond State Value in State

105.25

105.25

105.25

100 0

Prob of State

105.25

100

100

0.04%

0.29%

6.24%

86.96%

5.15%

1.09%

0.05% 0.18%

BBB

100.44

AAA

AA

A

BBB

BB

B

CCC D

Call

Call

Call

No Call + Cpn

Deferred coupon

Deferred coupon Deferred coupon Default


105.25

105.25

105.25

65.51 0

Prob of State

105.69

94.30

87.51

0.04%

0.29%

6.24%

86.96%

5.15%

1.09%

0.05% 0.18%

BBB 88.53

40




Having worked back from year 50 to year 49, we can work this back all way to Year 0 of the perpetual which is Year 10 of our hybrid. This will give us the value of the perpetual at Year 10 of the hybrid. We use these values to tell us the value of our hybrid in Year 10 if it is not called but turns into a perpetual. In other words, we now know the value of all the Year 10 final states.

B) Valuing the 10 Year Hybrid We can now work back through the years of our 10 year hybrid using the regular Rock-Bottom Spread methodology with two modifications. First, our final state values will depend on rating states to assign whether it is called or becomes a perpetual (as shown above). Second, we factor-in coupon deferral at our coupon deferral trigger.

Again, we start at year 10 and work this back to our current time to calculate the ‘fair value’ price of the bond and then turn this into a ‘fair value’ spread. This fair value spread is the minimum spread an investor should be willing to accept given the probability-weighted future value and volatility of the hybrid.

Figure 26 shows the valuation of our year 9 bond using the final states in year 10 and the probabilities of being in each given by the transition matrix.

The fact that hybrids will turn into a perpetual in year 10 means we have to assign the value of the perpetual as the future value in state for each of the ‘non-call’ states, i.e. BBB-CCC in our example. We also factor in our coupon deferral by making no coupon payment in the rating states where there will be coupon deferral – states BB-CCC in our example. Putting this together we calculate a fair value price = 77.42 for our BBB-rated hybrid for year 914.

Figure 26: Year 10 States and Calculating the Value of a BBB-Rated Hybrid in Year 9

Source: JPMorgan

14 These rock-bottom prices are on the low side due to the fact we are using a corporate transition matrix and adjusting for ‘corporate’ volatility. The economics of a Banks Tier I piece don't compensate you for corporate downgrades and defaults, as we would expect. We are forced to use a corporate transition matrix for the examples as we only have an 18-state Banks-only transition matrix. For the 8-state examples we borrow the corporates matrix.

AAA

AA

A

BBB

BB

B

CCC D

Call

Call

Call

Perp + Coupon

Perp No Coupon

Perp No Coupon Perp No Coupon Default


100 + 6.742

100 + 6.742

100 + 6.742

0

0.04%

0.29 %

6.24%

86.96 %

5.15 %

1.09 %

0.05% 0.18 %

Prob of State State

3.55

76.13 + 6.742

22.69

8.45

BBB 77.42

41




We now simply work back using the Rock-Bottom Spread methodology, calculating the Rock-bottom price in each year back to the current time (for a newly issued hybrid this would be year 0). Each year of the fixed rate bond we need to remember to factor in our coupon deferral states where a coupon will not be paid.

From Theory to Practice The ‘fair value’ prices in the main section of this paper have a few additional complexities to the simplified examples presented in this Appendix. These additions are intended to make the valuation as real-world as possible and the framework is otherwise identical.

• 18-State Transition Matrices: We use the 18-State transition matrices to give us greater granularity in pricing.

• Modified Transition Matrix: We make slight modifications to the historical transition matrix in order to make the default probability monotonically increase for lower ratings. Due to some historical anomalies the historical-implied probability of default for A-rated companies is higher than AA-rated.

• Full Yield Curves: We use the full upward sloping risk-free curve for discounting.

• Quarterly Perpetual Payments, Calls and Transition Matrices: When the hybrid becomes a perpetual it pays quarterly and has quarterly call options. We model this using quarterly periods and use quarterly transition matrices to accurately calculate the forward probabilities.

42




Analyst Certification: The research analyst(s) denoted by an “AC” on the cover of this report certifies (or, where multiple research analysts are primarily responsible for this report, the research analyst denoted by an “AC” on the cover or within the document individually certifies, with respect to each security or issuer that the research analyst covers in this research) that: (1) all of the views expressed in this report accurately reflect his or her personal views about any and all of the subject securities or issuers; and (2) no part of any of the research analyst’s compensation was, is, or will be directly or indirectly related to the specific recommendations or views expressed by the research analyst(s) in this report.

Important Disclosures

Explanation of Credit Research Ratings: Ratings System: JPMorgan uses the following sector/issuer portfolio weightings: Overweight (over the next three months, the recommended risk position is expected to outperform the relevant index, sector, or benchmark), Neutral (over the next three months, the recommended risk position is expected to perform in line with the relevant index, sector, or benchmark), and Underweight (over the next three months, the recommended risk position is expected to underperform the relevant index, sector, or benchmark). JPMorgan’s Emerging Market research uses a rating of Marketweight, which is equivalent to a Neutral rating.

Valuation & Methodology: In JPMorgan’s credit research, we assign a rating to each issuer (Overweight, Underweight or Neutral) based on our credit view of the issuer and the relative value of its securities, taking into account the ratings assigned to the issuer by credit rating agencies and the market prices for the issuer’s securities. Our credit view of an issuer is based upon our opinion as to whether the issuer will be able service its debt obligations when they become due and payable. We assess this by analyzing, among other things, the issuer’s credit position using standard credit ratios such as cash flow to debt and fixed charge coverage (including and excluding capital investment). We also analyze the issuer’s ability to generate cash flow by reviewing standard operational measures for comparable companies in the sector, such as revenue and earnings growth rates, margins, and the composition of the issuer’s balance sheet relative to the operational leverage in its business.

JPMorgan Credit Research Ratings Distribution, as of July 3, 2006

Overweight Neutral Underweight

EMEA Credit Research Universe 27% 49% 25% IB clients* 57% 62% 62%

Represents Ratings on the most liquid bond or 5-year CDS for all companies under coverage. *Percentage of investment banking clients in each rating category.

Analysts’ Compensation: The research analysts responsible for the preparation of this report receive compensation based upon various factors, including the quality and accuracy of research, client feedback, competitive factors and overall firm revenues. The firm’s overall revenues include revenues from its investment banking and fixed income business units.

Other Disclosures

Options related research: If the information contained herein regards options related research, such information is available only to persons who have received the proper option risk disclosure documents. For a copy of the Option Clearing Corporation’s Characteristics and Risks of Standardized Options, please contact your JPMorgan Representative or visit the OCC’s website at http://www.optionsclearing.com/publications/risks/riskstoc.pdf.

Legal Entities Disclosures U.S.: JPMSI is a member of NYSE, NASD and SIPC. J.P. Morgan Futures Inc. is a member of the NFA. J.P. Morgan Chase Bank, N.A. is a member of FDIC and is authorized and regulated in the UK by the Financial Services Authority. U.K.: J.P. Morgan Securities Ltd. (JPMSL) is a member of the London Stock Exchange and is authorised and regulated by the Financial Services Authority. South Africa: J.P. Morgan Equities Limited is a member of the Johannesburg Securities Exchange and is regulated by the FSB. Hong Kong: J.P. Morgan Securities (Asia Pacific) Limited (CE number AAJ321) is regulated by the Hong Kong Monetary Authority and the Securities and Futures Commission in Hong Kong. Korea: J.P. Morgan Securities (Far East) Ltd, Seoul branch, is regulated by the Korea Financial Supervisory Service. Australia: J.P. Morgan Australia Limited (ABN 52 002 888 011/AFS Licence No: 238188, regulated by ASIC) and J.P. Morgan Securities Australia Limited (ABN 61 003 245 234/AFS Licence No: 238066, a Market Participant with the ASX) (JPMSAL) are licensed securities dealers. New Zealand: J.P. Morgan Securities New Zealand Limited is a New Zealand Exchange Limited Market Participant. Taiwan: J.P.Morgan Securities (Taiwan) Limited is a participant of the Taiwan Stock Exchange (company-type) and regulated by the Taiwan Securities and Futures Commission. India: J.P. Morgan India Private Limited is a member of the National Stock Exchange of India Limited and The Stock Exchange, Mumbai and is regulated by the Securities and Exchange Board of India. Thailand: JPMorgan Securities (Thailand) Limited is a member of the Stock Exchange of Thailand and is regulated by the Ministry of Finance and the Securities and Exchange Commission. Indonesia: PT J.P. Morgan Securities Indonesia is a

43




member of the Jakarta Stock Exchange and Surabaya Stock Exchange and is regulated by the BAPEPAM. Philippines: This report is distributed in the Philippines by J.P. Morgan Securities Philippines, Inc. Brazil: Banco J.P. Morgan S.A. is regulated by the Comissao de Valores Mobiliarios (CVM) and by the Central Bank of Brazil. Japan: This material is distributed in Japan by JPMorgan Securities Japan Co., Ltd., which is regulated by the Japan Financial Services Agency (FSA). Singapore: This material is issued and distributed in Singapore by J.P. Morgan Securities Singapore Private Limited (JPMSS) [mica (p) 069/09/2006 and Co. Reg. No.: 199405335R] which is a member of the Singapore Exchange Securities Trading Limited and is regulated by the Monetary Authority of Singapore (MAS) and/or JPMorgan Chase Bank, N.A., Singapore branch (JPMCB Singapore) which is regulated by the MAS. Malaysia: This material is issued and distributed in Malaysia by JPMorgan Securities (Malaysia) Sdn Bhd (18146-x) (formerly known as J.P. Morgan Malaysia Sdn Bhd) which is a Participating Organization of Bursa Malaysia Securities Bhd and is licensed as a dealer by the Securities Commission in Malaysia

Country and Region Specific Disclosures U.K. and European Economic Area (EEA): Issued and approved for distribution in the U.K. and the EEA by JPMSL. Investment research issued by JPMSL has been prepared in accordance with JPMSL’s Policies for Managing Conflicts of Interest in Connection with Investment Research which can be found at http://www.jpmorgan.com/pdfdoc/research/ConflictManagementPolicy.pdf. This report has been issued in the U.K. only to persons of a kind described in Article 19 (5), 38, 47 and 49 of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2001 (all such persons being referred to as "relevant persons"). This document must not be acted on or relied on by persons who are not relevant persons. Any investment or investment activity to which this document relates is only available to relevant persons and will be engaged in only with relevant persons. In other EEA countries, the report has been issued to persons regarded as professional investors (or equivalent) in their home jurisdiction Germany: This material is distributed in Germany by J.P. Morgan Securities Ltd. Frankfurt Branch and JPMorgan Chase Bank, N.A., Frankfurt Branch who are regulated by the Bundesanstalt für Finanzdienstleistungsaufsicht. Australia: This material is issued and distributed by JPMSAL in Australia to “wholesale clients” only. JPMSAL does not issue or distribute this material to “retail clients.” The recipient of this material must not distribute it to any third party or outside Australia without the prior written consent of JPMSAL. For the purposes of this paragraph the terms “wholesale client” and “retail client” have the meanings given to them in section 761G of the Corporations Act 2001. Hong Kong: The 1% ownership disclosure as of the previous month end satisfies the requirements under Paragraph 16.5(a) of the Hong Kong Code of Conduct for persons licensed by or registered with the Securities and Futures Commission. (For research published within the first ten days of the month, the disclosure may be based on the month end data from two months’ prior.) J.P. Morgan Broking (Hong Kong) Limited is the liquidity provider for derivative warrants issued by J.P. Morgan International Derivatives Ltd and listed on The Stock Exchange of Hong Kong Limited. An updated list can be found on HKEx website: http://www.hkex.com.hk/prod/dw/Lp.htm. Korea: This report may have been edited or contributed to from time to time by affiliates of J.P. Morgan Securities (Far East) Ltd, Seoul branch. Singapore: JPMSI and/or its affiliates may have a holding in any of the securities discussed in this report; for securities where the holding is 1% or greater, the specific holding is disclosed in the Legal Disclosures section above. India: For private circulation only not for sale.

General: Additional information is available upon request. Information has been obtained from sources believed to be reliable but JPMorgan Chase & Co. or its affiliates and/or subsidiaries (collectively JPMorgan) do not warrant its completeness or accuracy except with respect to any disclosures relative to JPMSI and/or its affiliates and the analyst’s involvement with the issuer that is the subject of the research. All pricing is as of the close of market for the securities discussed, unless otherwise stated. Opinions and estimates constitute our judgment as of the date of this material and are subject to change without notice. Past performance is not indicative of future results. This material is not intended as an offer or solicitation for the purchase or sale of any financial instrument. The opinions and recommendations herein do not take into account individual client circumstances, objectives, or needs and are not intended as recommendations of particular securities, financial instruments or strategies to particular clients. The recipient of this report must make its own independent decisions regarding any securities or financial instruments mentioned herein. JPMSI distributes in the U.S. research published by non-U.S. affiliates and accepts responsibility for its contents. Periodic updates may be provided on companies/industries based on company specific developments or announcements, market conditions or any other publicly available information. Clients should contact analysts and execute transactions through a JPMorgan subsidiary or affiliate in their home jurisdiction unless governing law permits otherwise.

Revised September 14, 2006.

Copyright 2006 JPMorgan Chase & Co. All rights reserved.




Documents

A Framework for Valuing Financial Hybrids - …quantlabs.net/academy/download/free_quant_instituitional_books_/[JP... · A Framework for Valuing Financial Hybrids Introducing and