Investment Grade credit financials versus industrials, as seen by European insurers

Transcription

1 Investment Grade credit financials versus industrials, as seen by European insurers Introduction To secure competitive rates of return on their assets, insurers have invested heavily in investment-grade bonds in recent years. These investors are not solely concerned about mark-to-market fluctuations; they are also responding to two special factors. First, Solvency II will influence their assessment of assets attractiveness, and second, they continue to make decisions largely with buy-and-hold in mind. Under a traditional approach to financial markets, an asset s expected return is equal to the risk-free interest rate plus a risk premium, which increases with the amount of risk taken. The Solvency II approach, which will regulate insurers capital requirements starting in January 213, provides a way to calculate a benchmark risk premium below which investments become unattractive. 1 Solvency II and measuring the attraction of credit bonds Applying the same rationale to credit bonds, a breakeven spread can be computed and compared with actual spreads provided the probable loss given default is also factored in. Under a traditional approach to financial markets, an asset s expected return is equal to the risk-free interest rate plus a risk premium, which expands with the amount of risk taken. The Solvency II approach, which will regulate insurers capital requirements starting in January 213, provides a way to calculate a benchmark risk premium below which investments are not attractive. Under the earlier system, insurers regulatory capital was calculated on a flat-rate basis, independently of asset allocation. A new feature of Solvency II is that it bases the capital requirement (known as the solvency capital requirement, or SCR) on a detailed map of the risks effectively assumed (i.e. one-year VaR of 99.5%). The imperfect correlation of risks is taken into account so that the aggregate capital requirement is less than the sum of its parts. For one of the risks, market risk, the regulators stipulate that the SCR must be calculated as the result of a specific market stress scenario for each asset class. Thus, the SCR is closely tied to asset allocation. Moreover, the insurer may calculate a required capital contribution under Solvency II and a cost of capital for each portfolio asset and each potential investment. This represents the breakeven point for comparison of the asset s expected risk premium. For credit bonds, it is tempting to compare the annual cost of capital under Solvency II with the spread, which under insurers typical buy-and-hold approach is an important factor in credit bonds expected returns. However, the relationship is not a direct one, because the spread also contains a component to offset probable losses linked to defaults. Thus we define the breakeven spread for a given bond as the sum of two terms:. Breakeven spread = probable loss given default + Solvency II cost of capital Under this approach, industrial credit bonds hold little appeal, while spreads on financial bonds are well above breakeven spreads, notably in the banking sector and regardless of the level of subordination. On the risk side and from a buy and hold perspective, insurers mainly suffer when ratings deteriorate, thus pushing up their capital requirements under Solvency II and, naturally, in the event of a default or a downgrade to high yield. Countries ability and political will to prop up their banks wilted in the aftermath of the crisis, and Basel 3 has led to changes in banks liability structure, with issuance of covered bonds and increases in capital. Both factors have a negative impact on senior debt but are less harmful to subordinated debt. The deleveraging phase is largely over for industrials, but not for banks. For the latter, a medium-term investment in credit still outperforms an equity investment, provided the issuers are carefully selected on the basis of their business model and assets.

2 The first term depends on the probability of default based on the bonds rating and the loss given default rate. Break down of Solvency II breakeven spread along ratings (5-yr bonds) The second term is the product of the expected ROE and the credit SCR attached to the bond, which depends on the bond s duration and the effect of the rating-specific stress scenario on the spread, as stipulated by the regulator. In theory, it should be adjusted by a factor specific to the insurer to account for the lack of correlation with other risks and its target solvency ratio. We carried out a calculation using default probability assumptions compatible with a moderate recovery economic scenario. It is shown in the table below. We also show the regulator s stress spread scenarios under the standard formula. spread (bp) Compensation for capital charge Losses due to expected Stress Spread QIS5 Default probability AAA.9%.% AA 1.1%.5% A 1.4%.4% BBB 2.5% 1.% BBB 4.5% 1.3% B 7.5% 2.2% 5 AAA AA A BBB BB B Comparison between breakeven spread and spreads quoted on a sample of bonds 4-6 yr The loss given default rate is set at 6% and the required ROE at 1%. The insurer s adjustment factor is here set at 1. Under these assumptions, the 1 st graph shows how the loss effects linked to defaults and capital returns are distributed in the breakeven spread of a five-year bond. The portion of the spread that makes up for default losses is negligible only for AAA and AA ratings. It increases as the rating deteriorates, but the portion of return on equity remains preponderant. spreads en bp median spread industrials median spread financials breakeven spread 5 years These breakeven spreads can then be compared to market spreads, both to assess the bond market s costliness and to help select credit bonds with better-than-breakeven spreads. For quoted spreads, we used the swap curve as a reference yield curve. Corporate credit bonds (main indices) were then divided into groups by ratings and maturity ranges. The 2 nd graph shows the comparison with breakeven spreads. We focused on the four-to-six-year maturity range and separated financials and industrials. We found that for high-yield and financial bonds of all ratings, the median quoted spread was higher than the breakeven spread. For industrials, the median spread was below breakeven for A and BBB bonds. These ratings include the largest number of bonds. For financials, quoted spreads were even higher than breakeven spreads for a rating one notch lower. In fact, a closer look at financials shows very wide variation amongst sectors and degrees of subordination, as the 3 rd graph shows for the subset of financial bonds rated A. We note in particular the high spread levels on subordinated bonds. 1 AAA AA A BBB BB B Comparison between breakeven and quoted spreads A rated 3-7yr bonds, detailed by financial subsectors IND FIN Bk senior Spec fin Insur senior median - quantile 75% median - quantile 25% Bk T2 breakeven spread A 5y breakeven spread BBB 5y Bk T1 & UT2 Insur sub 2

3 2 Reasoning is still largely buy and hold, thus risk perceptions is broader than mark-to-market Under Solvency II assets are valued at mark to market; price fluctuations thus influence the value of assets and consequently the insurer s situation with respect to capital requirements (the solvency ratio). But insurers also continue to perform certain calculations that are based on a buy-and-hold approach. For example, they calculate earnings distributable to the policy holders for euro-denominated life insurance contracts. In this case, the insurer is not sensitive to price fluctuations to the exclusion of other factors. But the bond yield and the credit spread when the bond is bought are very important, and bonds with attractive spreads relative to breakeven are attractive, even if spreads do not come down after the bond has been bought. If the insurer has new cash flows to reinvest, it will even see spread compression as something of a drawback, because it causes investment opportunities to disappear (conversely a limited rise in spreads may even be experienced as an opportunity). In terms of adverse events, of course, the insurer will obviously loose money if the bond defaults. The insurer also suffers if rating deteriorates, because under Solvency II the capital requirement is then increased. Moreover, if a bond is downgraded to high-yield, the insurer may be obliged by internal rules to sell the bond usually at a loss. Finally, if callable subordinated bonds are not redeemed on the call date, the lengthened duration simultaneously reduces profitability and increases capital consumption. In contrast, a call works in the insurer s favour. We have examined the current situation for investment-grade bonds relative to their breakeven spreads, viewing risk from this highly specialised perspective Impact of sovereign crisis on financials spreads 12-7 Source: Amundi Strategy Itraxx Senior Finls spreads (in bp, l.h.s.) itraxx Sovereign spread DJ Stoxx Banks 3m implied vol. (%, r.h.s.) For investment-grade industrials, many spreads have dropped below breakeven, and unless a rating upgrade is expected, these bonds attractiveness has become more limited. For financials, the spreads are attractive, but (for banks in particular) two factors may influence ratings: the increase in sovereign risk and changes in banking regulations. The 28 crisis and the eurozone sovereign debt crisis reduced governments ability to prop up troubled banks. Agencies are factoring this point into their analysis of banks, noting that this situation may lead to a slight downgrading in ratings independently of the degree of subordination. This is all the more true for banks in countries in difficulty. The markets have factored in this point, and while financials spreads had largely normalised before the sovereign crisis, they went back up once the crisis broke. This created a decoupling from implied equity volatility, as shown in the graph opposite. Insurers whose exposure to high spread sovereigns is significant, should pay close attention to the correlation that has appeared between banks and sovereigns spreads. because even though the exposure to Eurozone sovereigns has no impact on the capital requirement, if it is compounded by credit exposure to banks in the same country, it may lead to volatility in the assets and the solvency ratio. Even for very highly rated countries, the political will to prop up banks may flag. Germany has passed a law to push more of the responsibility for losses onto the bondholders (the haircut is applicable to LT2 assets if a bank has significant losses) rather than taxpayers. These measures led Moody s to downgrade subordinated LT2 debt of 24 banks, citing an implied decrease in government support for these institutions. The gradual integration of Basel III is leading to changes in banks capital structure. Some subordinated debt does not have enough loss absorption capacity under the new regulations, and issuers should be redeeming such bonds even before replacing them because their coupon rates are too high. The effect of this shift would at once be positive for prices (for bonds trading below par) and duration. The stronger capital position, with decreased leverage, would have a protective effect on senior bonds and seasoned subordinated debt that is not eligible for inclusion in capital. The result may be improved ratings in the medium term. However, we also note a surge in covered bond issues, which have become banks preferred financing instrument to the detriment of senior debt issues. By giving senior bondholders less protection than covered bondholders, this substitution is unfavourable to senior bonds. However, it has no impact on subordinated debt. 3

4 3 Consequences for bets on financials versus industrials and equities versus credit Performance of financials versus non financials, comparison between credit and equities Since 29, financials behaviour versus industrials has been very different in credit and in equities (cf. graph). The start of the year brought some improvement, but trends are opposite. Globally, financial equities are underperforming industrials, whilst the trend in the spread differential is fairly flat or slightly rising. The deleveraging phase, which is favourable to credit, is largely completed for industrials, but not for financials, where regulatory changes are prolonging this phase in which credit traditionally leads equities. The process is far from completed in Europe, especially in the countries that fell prey to bubbles (property in Spain and Ireland). And even when it is finally finished, there is no assurance that the divergence will narrow. In fact, banks ROE will not return to its pre-crisis level, nor will the scale of banks asset base, because they are also deleveraging by reducing assets. Relative to pre-crisis levels, spreads may normalise, but equity prices will not. The situation recalls that of the telecoms sector, which also underwent a lengthy deleveraging process after the internet bubble burst in 2. -1,95-2,85-3,75-4,65-5,55-6,45-7, spread A IND-FIN (l.h.s.) MSCI Eur Banks / MSCI Euro ex Banks Source : Amundi ( h Quant ) Research For industrials, the less attractive valuation of spreads and the advanced stage of deleveraging should normally favour equities over (investment-grade) credit. However, for financials, with spreads trading above breakeven even in cases of one-notch ratings cuts, the potential gain on credit is attractive and well protected from a medium-term buy-and-hold standpoint. Given financials ongoing deleveraging, we now view them as a more attractive credit than equity play. Careful choice of issuers, with attention to their systemic risk and the degree of risk in their activities, is vital to success. At equivalent ratings, we recommend subordinated bonds of trusted issuers rather than senior debt of issuers who, on the whole, are less highly rated. 4

5 Contributor Editor Philippe Ithurbide Head of Research, Analysis and Strategy Paris Contributors Sylvie de Laguiche Head of Quant research Paris Sergio Bertoncini Credit Strategy Milan DISCLAIMER Chief editor: Pascal Blanqué Editor: Philippe Ithurbide This document may not be reproduced, fully or partly, or communicated to third parties without our authorisation. Published by Amundi a joint stock company (société anonyme) with a registered capital of euros. An investment management company approved by the French Securities Authority (Autorité des Marchés Financiers - AMF ) under No.GP436. Registered office: 9, boulevard Pasteur 7515 Paris - France RCS Paris. The information contained in this document is not intended for distribution or use by any person or entity in a country or jurisdiction where such distribution or use would be contrary to legal and regulatory provisions, or which would require Amundi and its affiliated companies to comply with the registration procedures of said countries. All of the products and services may not be registered or authorized in all countries or available to all clients. The data and information in this document is provided solely for information purposes. None of the information contained in this document constitutes an offer or appeal by any member of the Group Amundi to provide investment advice or services or to buy and sell financial instruments. The information contained in this document is based on sources that we deem to be reliable, but we cannot guarantee that it is exact, complete, valid or relevant, nor should it be considered as such for any purpose whatsoever.