This year's fashion of long dated bonds has certainly caught the eye, with France's 50 year turn particularly striking. But what is not so clear is the level of supply that we can expect to follow in the coming months and the types of investors that will buy the paper. Neil Day reports.
When Agence France Trésor on February 3 announced that it was launching an exploratory study into the possibility of issuing a 50 year OAT, it said that the results should be handed in before the end of the month.
But by the end of February the market knew a lot more than it had expected to as an overwhelmingly enthusiastic response from investors meant that the survey had been concluded, the deal launched and history made in the following three weeks.
"What was most surprising about the deal was the speed at which it could be executed," says Benoit Coeuré, deputy chief executive officer of AFT. "We had envisaged various scenarios regarding the timing and the reality was the quickest of any that we could have imagined."
The survey was carried out by the eventual bookrunners of the transaction — BNP Paribas, Barclays Capital, Deutsche Bank and HSBC CCF the week after the project was announced. AFT then embarked on a roadshow in the week starting February 14.
"The week after we did the survey of investors and using its results, we met with a range of clients in various countries with a view to releasing the results of this consultation on the Friday," says Coeuré. "But it turned out that the reaction of those investors was so positive that we were able to launch the transaction immediately after announcing the results of the survey and roadshow.
"When we had initially made public our intention to embark on this project, we had thought that our discussions with the market would take longer as we would have to convince investors to get involved, but it turned out they did not need to be convinced."
The execution of the trade underlined the market's enthusiasm for the new point on the yield curve.
AFT achieved a size of Eu6bn for the 4% 25 April 2055 OAT after building a book of more than Eu19bn, having attracted orders from more than 200 accounts in 22 countries. This also enabled pricing of just 3bp over the 2035 OAT, at the tight end of the 3bp-7bp range, since Eu15bn of the book held at that level.
The success of the project excited market participants. "A banner transaction," said one syndicate member at the time. "It expands the boundaries of the euro market and highlights the huge bid that exists for duration. Not too many months ago 30 years was a push, but the dynamics have changed dramatically."
Longer trends emerge
Such sentiments this year have not, however, been based solely on the execution of the French landmark — a series of successful trades in the 30 year part of the curve have also added to the excitement at the long end.
The Kingdom of Spain, for example, had produced the euro zone government bond market's previous longest dated issue, a Eu6bn 32 year deal in the first half of January via Barclays, Banco Bilbao Vizcaya Argentaria, Calyon, Deutsche, Dresdner Kleinwort Wasserstein and Santander Central Hispano.
The deal attracted Eu17bn of demand in just over 24 hours and set the tone for the long dated activity that followed. It was priced at the tight end of its 4bp-7bp over the 2034 Bund guidance, providing Spain with its narrowest ever margin over the German curve on a syndicated deal.
The trends driving 30 year issuance were also becoming clear.
"There are many factors that can explain this dynamic activity," says José María Fernández, head of the public debt department at the Spanish Treasury in Madrid. "From the investor's point of view, a growing structural demand could be developing, fuelled by the changes in the social security systems and by the aging population issue. At the same time, the existing low yields could be encouraging some investors to buy longer dated paper that trades above a certain yield floor.
"From the issuer's standpoint, satisfying demand together with the aim of locking in low yields are the most probable drivers."
Such factors encouraged the Hellenic Republic to launch its first 30 year benchmark at the beginning of March. BNP Paribas, Citigroup, Deutsche Bank, Morgan Stanley and the National Bank of Greece were leads on the September 2037 transaction.
A book of Eu9bn based on price guidance of 25bp-27bp over Bunds enabled Greece to increase the deal from the minimum Eu3bn size to Eu5bn. Pricing of 26bp over enabled the sovereign to price the trade at 2.8bp over the Italian curve, the same margin as in 10 years.
Spyros Papanicolau, director general of the Greek public debt management agency in Athens, was pleased with the result. "We maintained a differential of around 3bp to Italy," he told EuroWeek, "which is quite extraordinary given that we are single-A and Italy is double-A in terms of rating."
However, some market participants raised questions over the solidness of the transaction and the sector as a whole, with the spread widening slightly on the Greek deal and yields having backed up. "There has been a huge amount of supply in 30 and 50 years," said one. "A lot of people who bought long dated paper have lost money and are getting a little more cautious."
Such fears that the activity at the very long end of the market would prove short lived were swept away when the Netherlands sold its first 30 year government bond since 1998.
Sold via the Dutch State Treasury Agency's Dutch direct auction method, the issue attracted Eu17bn of orders in just two hours, enabling the DSTA to allocate Eu5.228bn of the bond at 4bp over Bunds, the tight end of the 4bp-7bp guidance.
"The success of the transaction shows that investors are keen to buy long dated debt and are looking for diversification in the triple-A band," Ron Bruggink, head of money and capital markets at the DSTA in Amsterdam said at the time of the deal. "And the fact that the new bond is still trading at plus 4bp to Bunds in the secondary market shows the pricing was fair, neither too cheap nor too expensive."
And with the European Investment Bank having added a supranational angle to the 30 year sector with a Eu5bn EARN via Barclays, BNP Paribas, HSBC and Morgan Stanley in mid-May that was twice covered, the very long end of the market appears in great shape.
"The deals that have come have generally been extremely successful," says Hans den Hoedt, head of public sector origination at ABN Amro in London. "Spain, France and Holland have gone very well, as did the 30 year Bund auction and the EIB trade."
However, den Hoedt says that it is important to recognise that this year's activity is the culmination of a trend that has been ongoing for several years, rather than the result of any seismic shift on January 1, 2005.
"It is a natural development of what has been building up over the last few years," he says. "Two to three years ago there was duration lengthening more into the 15 year sector, and last year and this year it has started to happen into 30 years. But 30 years is nothing new for France or Germany or Italy — they have been issuing in the maturity for some time and their auctions have usually been very successful."
One of the reasons why such activity has captured the market's attention this year has been the increasing use of syndication to launch 30 year lines and its use on France's 50 year.
That said, syndication is merely being used to better exploit the opportunities available.
A greater use of the sales method rather than auctions was predicted when the Republic of Italy became the first sovereign to use syndication to sell a 30 year issue, in September 2003 when it launched a Eu7bn BTP via Citigroup, Deutsche, ING, JP Morgan and UniCredit Banca Mobiliare.
"The 30 year maturity has been almost untouched by European sovereigns through syndicated issues," said a Deutsche Bank official at the time. "But the phenomenal success of this Italy transaction will probably make others look closely at syndicated 30 year deals."
The demand that these sovereigns are so keen to tap into is for the most part driven by two types of demand.
"Firstly," says Bruggink at the DSTA, "there are a lot of long term liabilities in the world for which long term assets are a good match. You see this trend in the world of insurance companies and in the world of pension funds. That is the real demand that is driving issuance.
"Secondly, you have the trading community making that even bigger because they are anticipating future demand. And when you combine the two you have the trade of the year."
The old debate
To what extent activity this year has been driven by the first factor or the second has been a matter of heated debate.
This is partly because in explaining the growing bid at the long end, several parties highlighted changes needed to pension systems as a result of ageing populations that would lead to an inexorable increase in demand for long dated assets.
EIB shines in sovereign class
The supranational's Eu5bn 4% October 2037 EARN will face stiff competition for this year's deal of the year — not least from France's 50 year OAT — but it was undoubtedly another landmark in the development of the SSA sector in much the same way as last year's long dated issue.
Like the French issue, orders were swiftly placed for the benchmark. "The books opened on Monday morning London time and closed 36 hours later," Barbara Bargagli-Petrucci, head of the EIB's funding department in Luxembourg, told EuroWeek.
"After reaching close to Eu8bn on day one, the total order book closed on Tuesday afternoon at over Eu11bn, more than twice oversubscribed relative to the final deal size. This is the largest order book ever for an EIB EARN."
Led by Barclays Capital, BNP Paribas, HSBC and Morgan Stanley, the benchmark was priced in mid-May at 2bp over mid-swaps. This was the tight end of 2bp-3bp guidance — about 2bp back from where the 15 year EARN was trading, but flat to where the 2020 was priced last July.
One of the goals of the transaction that was achieved was for the EIB to reaffirm its credentials as what it likes to call a "sovereign-class" credit as an aggregate of its EU government owners.
With the curves of triple-A France, Germany, the Netherlands and Spain, and also Aa1/AA+/AA Belgium inverted beyond 10 years, the supranational marketed itself as offering value, while its triple-A ratings and stability argued for pricing inside the upward-sloping curves of Italy and also Greece.
Long end short of candidates
Before the EIB's benchmark, the largest outstanding 30 year non-government bond in euros was Réseau Ferré de France's (RFF's) Eu2bn 5% October 2033 deal. This amount was reached in January when Credit Suisse First Boston and Goldman Sachs tapped the issue for Eu750m.
The issuer has been one of the biggest issuers of long dated as well as inflation-linked assets, but few other borrowers have the kind of assets of RFF that match such long dated liabilities. And bankers say that with the credit curve likely to prove steeper than the EIB's for any lesser credit, issuance beyond 15 years is likely to prove prohibitively expensive.
The only other class of issuer to extend beyond that part of the curve this year have been covered bond issuers. Led by Caja Madrid with a Eu2bn 20 year cédulas hipotecarias via Barclays, Caja Madrid HSBC and Ixis CIB in late January, Spanish issuers have built on successful 15 year trades in 2004 to open up this new maturity.
AFT highlighted Europe-wide accounting changes as well as reforms in Denmark, the Netherlands and the UK as examples of the structural changes that would boost demand for instruments such as 50 year bonds. The Dutch reforms in particular received much attention.
For instance, one fixed income analyst of Dutch pension funds activity said: "This has been a major focus for the last two years, primarily because of the sheer size of the funds (close to Eu475bn is the latest estimate) and level of the duration gap between assets and liabilities. The proximity of legal changes (January 2006) and published concerns by the DNB (Dutch National Bank) have brought this issue back to centre stage.
"Those funds with insufficient asset cover will be forced by the PVK (Dutch pension fund regulator) to match liabilities within one year. The DNB stated that the average duration of assets is approximately five years and that liabilities are approximately 15 years."
That said, pension funds were only cited as making up 8% of the distribution of the 50 year OAT, which appeared at first glance to be a relatively modest amount. "Was that what they were really hoping for?" asked one banker at the time. "I don't think so."
Bankers both involved in the deal and away from it, however, say that such a figure should not be so surprising.
"There were legitimate reasons to have high expectations for the participation of pension funds," says Christopher Marks, head of sovereigns, agencies and supranationals at BNP Paribas in London, "but I think there was also a fair amount of over-enthusiasm and maybe some self-deception at the beginning of the year that the immediate buying of long duration securities was the obvious and sole solution for those investors facing reforms.
"The more educated view is that such investors, including the Dutch, have a whole range of options available to cover the requirement and have put in place a sustainable asset structure to cover their liabilities. They can do swaps, they can do equities, they can do bonds — there are a lot of ways to skin the cat. And frankly making a big decision to move a lot of money into the bond market right now when yields are so low is a hard decision to make."
Pension funds in disguise
At the same time, says Marks, the numbers attributed to pension funds for their participation in the OAT and 30 year issues may well have understated their involvement. "My supposition is that behind a lot of the asset manager accounts that represented the largest category of buying in the 50 year and the EIB 2037, and at least a third of the Greece 30 year was insurance and pension money," he says.
And France was also apparently able to achieve a high hit rate on those pension funds surveyed. The 8% participation in the bond of pension funds compared to 10% participation in the survey and an 11% share of those investors that expressed an interest in a 50 year bond.
Nevertheless, Coeuré at AFT acknowledges that interest in the bond did not pan out in exactly the way he had initially expected.
"We found a mix of investors buying the bond as a result of reforms of insurance and pension regulation and in anticipation of reforms, but also a completely different set of investors that were not buying the bond for anything related to ageing populations at all, but because of the technical characteristics of the bond," he says.
"We thought that this would be a pension fund product but it turned out that it could be used in various ways, not just to match long term liabilities, but just for the sake of its convexity, and we found key fund managers who wanted to buy the bond just to hedge against interest rate risk."
The reason why the convexity of the bond rather than its duration was especially attractive is highlighted by a comparison between the 50 year OAT and shorter dated bonds. Whereas the extra duration investors gain by moving from a 2035 OAT to the 2055 increases by only 34%, rising from 16.33 years to 22 years, the convexity increases by around 80%, from 4.04 to more than seven.
Old fashioned demand
In addition to the insurance companies and pension funds that bought the bonds for risk management or asset liability management reasons — and the hedge funds and other trading accounts that bought in anticipation of a growth in such demand — there has also been demand for what Marks at BNP Paribas refers to as "old-fashioned" reasons.
"Frankly the market is exceptionally high and directionally very strong, and there are very practical reasons for people to buy," he says. "They need spread, they need duration and you can't get yield anywhere further down the curve. No-one has any illusions about the direction of European fundamentals and rates are not going up, so you have to make the most of what the market is giving you."
PJ Bye, director, public sector syndicate at HSBC in London, holds similar views.
"When you have 10 year Bund yields at all-time lows of below 3.5%, there is a natural incentive to switch longer," he says, "and even if going 20 years further to get a 4% coupon might not sound that great, in relative terms there is value in going down the curve when your alternative is buying something at the richest level it has ever been.
"Furthermore 30 year yields are not at record lows and a lot of people believe that the 10-30 curve will go flat."
Bye points to the UK experience as evidence of how structural reforms will help flatten the euro zone yield curve. This comparison was highlighted by HSBC analysts in a research piece entitled Total Flattening published at the beginning of February. "For a considerable period the UK has experienced demand for ultra-long bonds as pension funds and life insurance companies reallocated out of equities into bonds," said the analysis.
Addressing the dilemma
But it is not just investors that are benefiting from the new supply of assets at the long end. A 50 year liability at such a low fixed rate is an attractive liability for issuers as well, say bankers.
"Let's be honest," says one, "issuers are doing so much at the long end because rates are attractive. We know that France has promoted the 50 year, quite correctly, as a policy instrument to satisfy a requirement born of the public finance dilemma of the next 50 years, but the fact is: it's great funding."
But others say that this side of the equation should not be overplayed. "It has not been overly opportunistic," says one syndicate manager. "Most of the governments that have issued have already done 30 year bonds so they have just pushed a little bit further down the curve."
In Amsterdam, Bruggink says that the DSTA is committed to the 30 year sector. "People had been asking us to issue a 30 year for some time (the last 30 year Dutch State Loan was issued in 1998), but we were not sure that we had the room for it and we wanted to make sure that when we came back to the sector it would not be a one-off," he says.
"Now we have come back and we are strongly committed to being active in this sector for the next few years. We see room to issue over the next three years and that is why we issued a 32 year. And even if rates are 5% next year, we will issue."
AFT, meanwhile, will be increasing the size of the 2055 OAT over the coming year. "It is now included in the regular auction programme," says Coeuré. "When we will tap it, it will be tapped via auction, and it is very likely that it will be tapped before the end of the year because we are committed to bringing liquidity to it.
"I cannot say what volume the bond will have by the end of the year, but in the medium term it will be much bigger than it is today so that it will meet the normal standards of our benchmarks. One could imagine that it will reach Eu10bn-Eu15bn, but the pace at which that will be achieved will depend on demand."
In the meantime, Coeuré would be pleased to see France's peers adding liquidity at the ultra-long end of the curve. "It is not for us to comment on anyone else's issuance policies and everybody has their own priorities and constraints in terms of portfolio management," he says. "But for the sake of the liquidity and the depth of this segment of the market, we would certainly welcome any other issuer. And we would also welcome non-sovereign supply."
After AFT had successfully executed its 50 year, several candidates for the ultra-long maturity were cited, but to date no euro zone sovereigns have made a commitment to follow the French example.
Italy, which has like France been quick to take advantage of new opportunities and which has in the last couple of years been most active at the long end of the market, was said to be considering extending its curve out to 50 years. However, the republic recently declared that it would be launching a new 30 year BTP.
Germany is also seen as a potential entrant into 50 years, although not until after it has launched its first post-war foreign currency bond. Also, the Federal Republic last year gave itself the option to issue inflation-linked bonds and its debut in that market could also take priority over a 50 year.
But some market participants feel that it is only a matter of time before the sovereign goes ultra-long.
"It is not in Germany's plans for this year," says one DCM official, "but it is too efficient a funding option for them to ignore. It serves a number of other purposes, such as diversification and profile. And frankly, given Germany's place in the market it is logical that they push out that far simply for reasons of standing.
"They also need all the money that they can get. The public finance situation in Germany is not getting any better, so where they can take some pressure off their issuance by spreading it a little thinner, it makes sense for them to do so."
While there are only one or two candidates for potential 50 year issues, the number of sovereigns active in 30 years or longer is not expected to grow. Apart from Greece, whose 30 year was its first, all the other governments active in this part of the curve had already been present in the past.
"You are only really looking at a handful of sovereign issuers that want to do 30 years," says Nick Dent, head of frequent borrower syndicate at CSFB in London. "If you look at some of the smaller sovereigns, at present they have no desire to extend out to 30 years, so I don't think we are going to see a rush of new names coming to the very long end of the market."
In a way, some sovereigns may almost suffer from the fact that they have less funding to do. "It is a bit of a dilemma for them," says one DCM official. "I'm sure that in some ways it is with some regret that they don't have sufficiently large requirements to take advantage of these new opportunities, particularly when they offer attractive funding.
"An issuer like Finland, where the public finances are in good shape, has faithfully followed the policy of trying to bunch together its issuance into large benchmarks to get on the electronic systems and to maintain focus on their securities in a world of bigger issuers, and as a result they simply can't exploit new opportunities in the same way."
However, some bankers believe that smaller sovereigns might occasionally find it interesting to move beyond 10 years, if not all the way out to 30.
"Maybe after the long dated activity of this year people will say to them that it will not be a problem if they aren't present in, say, 10 years and instead do something longer," says one frequent issuer origination banker. "If you look at the EIB, they haven't done a 10 year for some time, and have 2013, 2020 and now 2037 points on their curve.
"Under the old way of thinking, you should do a five year and a 10 year because that's where the money is, and if you are trying to maintain profile then you have to be in those maturities. But frankly the frontiers of the market have been altered in the past year. The EIB has in fact retained profile by moving into the new zones of interest and the fact that they don't have a fresh 10 year doesn't matter."
Irrespective of whether more sovereigns join the activity at the 30 and 50 year points on the curve, bankers are confident that after this year's boom the very long end of the curve will be a permanent fixture.
"This year we have seen a virtuous circle of supply and demand for the long end and that long dated issuance is here to stay," says Anneke de Boer, head of frequent borrower origination at Morgan Stanley in London. "The regulatory and accounting changes affecting the pension and insurance industries mean that depth of demand is certainly here for the foreseeable future."