- This year has so far been the busiest in terms of supply since 2011
- But we have revised down our supply outlook for 2016 as a whole
- Overall, euro benchmark market is likely to shrink this year
- Negative net supply and CBPP3 to result in further spread compression
- No material change seen in CBPP3, despite likelihood of more QE
- Negative yields spreading, now already 76% of iBoxx index
- H2 should bring more news on harmonisation of covered bonds
DISCLAIMER: This report has not been prepared in accordance with the legal requirements designed to promote the independence of investment research, and that it is not subject to any prohibition on dealing ahead. This report is marketing communication and not investment research and is intended for professional and eligible clients only.ABN-AMRO-Covered-Bond-Watch-Mid-term-review-less-supply-in-H2.pdf (523 KB)
A busy first half of 2016
The first half of this year was rather busy in terms of new supply of euro benchmark covered bonds. New issuance increased by 50% compared to last year, amounting to EUR 91bn. Having said that, the start of the second half of the year has been relatively slow, with EUR 3bn of euro benchmarks having been issued at the time of writing, whilst almost EUR 18bn was issued in July last year. Nevertheless, this year has so far been the busiest in terms of supply since 2011.
A breakdown by region shows that 62% of this year’s supply has come from issuers located in the euro area, 23% from other European countries, and some 15% from outside Europe. The increase compared to last year is mainly stemming from France, Norway, Germany, the Netherlands, Sweden, and the UK.
Euro area market shrinking on a net basis
However, on a net basis, the euro area covered bond market keeps on shrinking, as redemptions outpaced gross supply (mainly due to the Spanish market). It is the non-European market that is currently expanding the most, reflecting that covered bonds are increasingly becoming a global product. Canadian banks have for instance continued to keep a solid pace of new issuance, while they do not face significant amounts of redemptions yet. We have also seen inaugural euro benchmarks from Turkey and Singapore.
The Eurosystem’s purchases of covered bonds under CBPP3 are further reducing the net supply figure, as the central bank is regarded as a buy-and-hold investor. The Eurosystem bought around EUR 17bn of covered bonds in the primary market in 2016H1. This implies that the euro area market actually shrank by some EUR 23bn in the January-June period, while the total net supply number also turns negative when taking into account the primary purchases under CBPP3.
Supply outlook in 2016H2
We have revised down our outlook for this year’s total issuance of euro benchmark covered bonds by EUR 10bn to EUR 141bn. This would imply that some EUR 50bn of euro benchmarks will be issued in the second half of the year, clearly marking a slowdown in the pace of issuance compared to the first half of the year. What is more, it would imply that net supply will be EUR 5bn negative this year, shrinking for the third time in four years.
Strong issuance in 2016H1, the possibility to take up money for nothing in the TLTRO II for banks in the eurozone, the need to meet capital requirements, as well as reduced liquidity needs more generally, are the main factors driving the downward revision of our forecast. Another impediment could be that yields on covered bonds are currently firmly negative up to around 5 to 8yrs for most countries (see below), which would force issuers to issue at the longer end of the curve if they want to offer investors a positive yield. On the other hand, issuers could lock in cheap funding for longer by doing so.
TLTRO II to reduce issuance in periphery
The biggest adjustment in our new issuance projections is in Italy and Spain. Not only has issuance been rather low compared to our initial expectations (especially in Italy), but we expect that banks in these countries will also make maximum use of the free money offered in the TLTRO II operations by the ECB. Indeed, Bloomberg reported that Spanish and Italian banks increased their net borrowing by some EUR 25bn in the TLTRO II, with all the major banks involved.
Lower forecast for German Pfandbriefe due to taps
We have also lowered our forecast for issuance of euro benchmark covered bonds of German banks by some EUR 4bn. This is also based on the impact of the TLTRO II, albeit much less than in the periphery. More importantly, German issuers have recently increased the use of taps to meet funding needs, which is a more flexible and easier way to raise funding than issuing a new benchmark. It also increases liquidity of a bond. So looking forward, we expect more German issuers to tap the market rather than issue new benchmark deals. That most of German Pfandbriefe up to 8yrs are trading at negative yields will also be a barrier to come to the market, despite the fact that investors do not seem this to be a large issue.
UK issuers frontloaded in run-up to referendum
UK issuers seem to have frontloaded issuance this year, likely in anticipation of the referendum. We expect supply from UK issuers to remain limited during the remainder of this year due to the referendum result, which could make investors more cautious to buy UK paper. In any case, most UK issuers have already entered the market, with only some building societies (e.g. COVBS and YBS) still wanting to enter the market. See also our note on what the Brexit vote would mean for UK covered bonds.
More to come from Canada
Canadian banks have continued to be among the most active issuers so fa this year, and we expect them to continue to be so in the second half of the year. Some of the issuers have not yet shown their face to the market this year (e.g. NACN and CCDJ), while others have a record of issuing two or more deals per year. Most Canadian banks have still sufficient room for issuance before reaching the cap set by authorities.
A little more from the Netherlands
Looking at the Dutch market, all regular issuers have already sold benchmark deals this year. Still, we expect some issuance in the second half of the year. SNS Bank might be the most likely candidate, as the bank has completed its restructuring after the nationalisation in 2012, while the covered bonds have regained their triple-A status. The bank is now allowed to re-enter the market, which would be the first time since 2012. ING Bank could also sell a public benchmark deal, as it has almost fully used its smaller EUR 5bn covered bond programme that it has used for private placements.
Spreads at tight levels but still offering value versus government bonds
Covered bond spreads have reached relatively tight levels again, following the widening that set in in July 2015. The tightening was largely driven by improvement in risk sentiment on the back of the ECB announcement of more stimulus measures in March this year (increase in monthly amount of asset purchases, TLTRO II, and CSPP). As expected, this benefitted peripheral covered bonds by the most. Meanwhile, non-CBPP3-eligible covered bonds also did relatively well (Australia, Canada, Norway, Sweden), likely reflecting that they offer an attractive spread pickup versus covered bond that the Eurosystem buys.
In contrast, UK covered bonds have underperformed so far this year, which happened mainly in the run-up to the Brexit referendum. Furthermore, spreads of German Pfandbriefe have not yet reached levels at the start of this year, although they have tightened recently. Looking forward, we think that another year of negative net supply and ongoing central bank purchases should continue to put downward pressure on spreads.
Covered bonds are relative attractive versus government bonds
Covered bonds have become relatively attractive versus government bonds, which have seen their spreads tighten sharply on the PSPP purchases. This holds in particular for core countries. Indeed, the spread of for instance ING Bank’s 7y covered bond is trading at a spread pickup of some 22bps (z-basis) versus 7y Dutch government bonds. In Germany, most Pfandbriefe are trading at a pickup of some 30bps versus Bunds.
UK covered bonds look attractive
Spreads of UK covered bonds widened sharply in the run up to the Brexit referendum. They widened further directly after the UK vote to leave the EU, but have started to tighten ever since, while remaining at relatively high levels. Still, UK spreads are at similar levels as that of Ireland, while trading slightly below the larger Italian banks. The spread of UK covered bonds have also widened sharply versus core covered bonds, reaching around 20-30bps in the belly of the curve.
Overall, we think that UK covered bonds look relatively attractive, given that the impact of the UK vote to leave the EU on the quality of UK covered bonds should remain limited. Indeed, cover pools should withstand a weakening in the UK housing market, while rating buffers seem sufficient to keep a triple A rating in adverse scenarios (see a more detailed note here). Therefore, a tightening in spreads seems warranted, while an aggressive spread widening is unlikely in our view. However, volatility will probably remain high.
Meanwhile, we stick to the view that it is attractive to be overweight non-CBPP3 eligible paper, as it is still trading at an interesting spread pickup versus paper issued from euro area banks. Furthermore, we think that this would provide a hedge in case the Eurosystem would reduce CBPP3 purchases, as this would in our view result in an underperformance of CBPP3- eligible paper. Although, we do not expect the central bank to change its current behaviour in terms of CBPP3, the spread pickup will generate a running yield in the meantime.
No material changes to CBPP3
The total amount of CBPP3 currently stands at EUR 186bn, implying that the Eurosystem has already bought around EUR 43bn of covered bonds issued by euro area issuers so far this year. Actually, gross purchases are some EUR 4bn higher, reflecting redemptions and valuation adjustments. This implies that the central bank currently owns one third of the outstanding euro benchmarks in the iBoxx index that are issued by issuers located in the euro area, while that is some 25% of the total index.
The Eurosystem has bought on average EUR 7bn of covered bonds per month, 41% of which in the primary market and 59% in the secondary market. In recent months, the monthly purchases have been a bit below the average, largely due to less activity in the primary market. Indeed, average purchases in the secondary market have fluctuated around the EUR 200mn a day level so far.
Interesting to note is that the EUR 20bn increase in QE from April onwards has not led to an increase in CBPP3 purchases, with most of the increase being absorbed by the PSPP, and also subsequently the CSPP since June. This is another reason why we do not expect that an increase in the monthly amount of asset purchases later this year will result in a significant increase in the Eurosystem’s covered bond purchases. Actually, we expect no material changes to the central bank’s policy regarding CBPP3 in H2.
We expect the central bank to continue to buy around 25-30% of primary deals, with its share depending on overall appetite for new deals. Furthermore, we pencil in that the Eurosystem will continue to buy around EUR 1bn a week in the secondary market, although it may become harder to find paper because the central bank already owns a relatively large share of available covered bonds. This will become the main impediment going forward in servicing the CBPP3 needs. Looking further out, it is therefore more likely that CBPP3 purchases will slowdown rather than increase.
Negative yields spreading
The share of euro benchmark covered bonds in the iBoxx that are trading at a negative yield has increased from around 11% at the start of the year to some 76% at the time of writing. This amounts to EUR 575bn of euro benchmark covered bonds in the iBoxx index. The average negative yield is around -0.15%. As expected, the covered bonds of core countries trade most negatively. Meanwhile, the share of the amount of bonds that is trading at a negative yield, as well as the average negative yield, is less in the peripheral countries, the UK, and Australia. Most covered bonds up to 5y to 8y are trading at negative yields.
More news on harmonisation
The European Commission (EC) is expected to publish its final view on harmonisation of covered bonds some time soon. This follows the consultation that started in September last year. A public hearing was held in February, at which the EC was cautioned not to hurt a well-functioning market. Indeed, the general message from participants at the conference was that an integrated EU framework for covered bonds would not result in a material change in pricing or increase in the investor base. Therefore, a pan-EU framework should be flexible and principle-based, if any.
We wonder whether it is really necessary to move to an integrated EU covered bond framework. In our view, the EBA best practices already provide a sufficient incentive for convergence of covered bond frameworks, while also harmonising reporting standards. The EBA has already made clear that this is necessary in order to keep warranted the preferential risk weight treatment of covered bonds. If the implementation of the EBA best practices will be set as a necessary condition to keep the preferential risk weight, we expect that this will be a sufficient incentive for countries to make changes to covered bond frameworks. This, in turn, will lessen the need for additional EU wide regulation.
 Please note that the chosen bonds are just used as an illustration and should not be regarded as an investment recommendation.