Ask 10 investors, bond issuers and issuing houses which benchmark yield spread they look at to judge the relative value of a bond, and you may get 10 different answers. That is the joy of the new euro landscape, in which at least four governments are competing to have their bonds represent the benchmark rate on at least one point of the yield curve.
The natural benchmark should of course be the German government curve, since Germany is regarded as the best credit in euroland. But a glance at the French and German government yield curves shows that there are anomalies and strange spread differentials, widest at seven years (where the Bund is cheaper than the OAT by 14 basis points) and 30 years (where the OAT is 19bp cheaper, see chart).
Seven-year glitch
“There’s no natural spread between France and Germany,” says Luca Jellinek, bond analyst at Paribas in London, “it varies from bond to bond.” To illustrate his point he mentions the rumour that the old 10-year OATs of 2008 have been “put away by French insurance companies”, so they’re expensive and difficult to short. Their French customers were benefiting from tax advantages on an eight-year insurance contract, so last year they snapped up the eight-year bonds. Hence the spread against Bunds that have seven years to run. Another factor is the higher-than-market coupon, which makes the bonds pricier than those with coupons closer to the market.
At the shorter end of the curve the French OATs and BTNs are more liquid than the Bunds. That is because of the greater efficiency of the French repo market. Bunds, although the undisputed benchmark at 10 years, have nevertheless shown that they are vulnerable to a squeeze, such as that experienced last September. That was due to the enormous volume of open interest built up in the Swiss/German futures exchange Eurex’s 10-year Bund contract, dwarfing the cash positions in the underlying bonds. Speculation in those bonds cheapest-to-deliver into the Bund contract forced those with open positions to pay up for the next cheapest to deliver. “We were affected in terms of our basis trades [cash against futures],” says a Spanish institutional investor. “Obviously there are ways of profiting from such Eurex mistakes.”
There are times, argue the French, when it would be convenient if OATs and Bunds were deliverable into the same futures contract – reducing the risk of a squeeze in the underlying cheapest-to-deliver (CTD) bond.
With that in mind French futures exchange Matif announced in January that it had made Bunds deliverable into its euro notionnel (10-year) and euro five-year contracts. It also launched a two-year Franco-German E-note contract. Last year it had already announced that German, Austrian, Spanish and Italian bonds would be deliverable into the 30-year E-bond contract and a new 10-year multi-issuer (“also-ran”) contract. None of these has any liquidity, yet, in the months where Bunds are deliverable. But the French are not dismayed.
“Technically the dual-issuer 10-year contract is almost perfect for market makers,” says Stéphane Levy, head of government bond trading at Banque Nationale de Paris. Levy was on the committee of experts that helped Matif design the contract.
The potential flaw in such bi-issuer or multi-issuer contracts is that the lower-perceived credit will become the CTD. Since France is generally regarded as a worse credit than Germany (Moody’s on February 8 threatened to downgrade France from Aaa), a Franco-German contract would retain the characteristics of the French CTD bond. Véronique Bourcier, head of marketing at Matif, argues that there will be convergence of French and German rates, and the dual-issuer contracts will encourage that convergence. “If the contract is successful there’ll be more convergence between Bunds and OATs,” Bourcier says. “The contract will contribute to convergence because of demand from investors.”
Dual-index doubts
But, apart from these enthusiasts, many market participants are more sceptical about the dual-issuer concept. There will always be a credit differential between France and Germany. Why have a composite future when German, French, Italian, and Spanish government bonds can be hedged separately, on various European exchanges, with no basis risk? (Some dealers are prepared to take basis risk in favour of liquidity and will hedge, say, Italian BTPs with Eurex Bund futures. The BTP future traded on the London futures exchange Liffe is much less liquid than the Bund future, but other dealers are happy to use it for basis trades. “Liffe’s BTP future is basically dying,” says one Italian trader. A similar contract on the Italian exchange MIF trades only around 1,000 lots a day). “If you’re going to have a multi-issuer contract, it would make more sense to offer a cash-settled multi-issuer index,” suggests Doug Huggins, relative-value analyst at Deutsche Bank. “With an index, if Spain widens against Germany, you might get a better hedge.”
An example of the dual-issuer futures contract was seen on Eurex, when the bonds of the Treuhandanstalt (the privatization agency for east Germany) were made deliverable into the Bund and Bobl (five-year) futures contracts. “The Treuhand was Bund-guaranteed,” says Michael Klein, head of government bonds and repo at DG Bank, “but the Treuhand bonds traded at a slight spread (one or two basis points) because some clients didn’t accept them.” The Treuhand bonds became the CTD, turning the Bund future into a virtual Treuhand future. The Treuhand deliverables are being phased out and no more Treuhand bonds are being issued.
Given the uncertain nature of multi-issuer bonds, isn’t it better to accept that a plethora of benchmarks will survive until, perhaps, a single, standard yield curve develops?
In the new euro zone, primary bond issuers and investors are learning to live with the wide choice of benchmarks against which to price their issues or the performance of their corporate bond portfolios. For some, such as the European Investment Bank, it is unimportant, because they look at absolute yields or cost of funds. But, take a triple-A issuer of seven-year Eurobonds. Should its issue be priced at 12bp over OATs or 1bp over Bunds? For most French issuers the choice is clear, likewise for the Germans, such as government-owned KfW. Landesbank Hessen-Thüringen, which offered €1 billion ($1.1 billion) of seven-year mortgage bonds on January 19, priced them over Bunds. “For the time being we price everything over Bunds,” says Jörg Kranz, head of new issues at Helaba. “But over time we’ll let the market decide.”
But there’s no doubt that issuers and investors are confused. “Many issuing clients look at their own government bonds,” says one syndicate chief. “The Spanish look at OLOs.” He sees no clear indication that this will change over time. “Investors will have to get used to it.” In the meantime, there are plenty of arbitrage opportunities for those prepared to put money on where they think the true price should be.
Do as the clever boys do
About two-thirds of Eurobond issuers are pricing over Bunds, and one-third over OATs, says Brian Mooyaart of bond-analytics firm Mooyaart Consult. “The clever boys are quoting over both.” The danger, he says, is that an issue will be seriously mispriced, leaving too much on the table. The greatest anomaly, Mooyaart says, is in seven years. “If the German government issued in its own name in seven years,” says Mooyaart, “the yield would be substantially below where Bunds are trading in the secondary market.”
But the finance ministry in Bonn isn’t that bothered. “We like to focus on several parts of the yield curve,” says a debt department official, “and we think that two, five, 10 and 30 years is enough. At the moment we don’t see the need to tap seven-year liquidity, we have to spot the parts of the curve where we get the largest relative advantage.” The Bund has two bonds of 2006 maturity, the official says, with a volume of roughly Dm40 billion ($22.7 billion): “From the sheer size point of view we see enough liquidity.” The non-current coupon may be a hindrance to trading, he admits, “but the French seven years have a 7.25% coupon.”
Eurobond issuing houses have to steer a careful political path. “We use the benchmark the borrower is used to being traded against and also often indicate spreads over other benchmarks, depending on where investor demand comes from,” says Guillaume Redaud, syndicate official at Paribas. But in the land of a dozen benchmarks the issuing houses themselves tend to use the swap curve, as the only homogenous rate from one to 10 years. “Most traders are looking at the swaps to price the bonds,” says Redaud, “it’s simply more practical.”
At Banca Commerciale Italiana, head of fixed income Luciano Steve concurs: “We try internally to price bonds against the swap curve for the time being.”
For hedging a Eurobond issue, the swap is the most obvious instrument, but, says Paolo Dallera, head of fixed income derivatives trading at San Paolo-IMI, “that may not always be the best solution”. In today’s confused euroland it “may be better to exploit the spread with the government bond market. In some cases I may rather buy Bunds than receive [fixed rate] on the swap. It’s nice there are so many variables.”
Borrowers are tempted to use the benchmark that gives them the narrowest spread, for optical effect and comparison with their peer group, although for selling to investors they do better to use the benchmark that shows the fattest spread.
Kingdom of Sweden issued e2 billion of seven-year bonds on January 28, priced at 12bp over Bunds (it would have been 20bp over OATs). “The Bund was more relevant,” says Eric Thedeen, head of funding at the Kingdom of Sweden debt office. Sweden swapped the funds. Before this year its after-swap funding cost against dollar Libor would have been the benchmark. Now it’s more inclined to measure itself against the bond yields of non-core eurozone issuers such as Finland, Belgium, Portugal and Italy, says Thedeen.
The Council of Europe Social Development Fund insists it has “no preference” for a French or German benchmark. “Our issues are quoted against Bunds and OATs,” says Jacques Mirante, the Council of Europe’s head of funding. Rumour has it that the fund, based in Paris, came under heavy pressure from the French treasury, to use the OAT benchmark.
Cades, the French public Caisse d’Amortissement de la Dette Sociale, unashamedly uses the French curve. “The only benchmarks we use are the BTN and the OAT,” says Christophe Frankel, head of capital markets at Cades. “The French curve is really accurate, with maximum liquidity and regular auctions.” Cades launched two five-year deals in February, priced over BTNs, and happily sold 20% of one and 30% of the other into Germany, Frankel says.
Frankel believes Matif’s dual-issuer futures contract is a good idea “because the larger the basket [of deliverable bonds] the better it is. It’s a way to open the structure to a wider range of investors and banks.”
The dual-issuer notionnel contract may have a sporting chance. Bourcier at Matif brushes aside intellectual objections, such as those in a research paper from Paribas (it argues that uncertainty about the CTD bond will introduce unwelcome optionality into the contract). “If you observe a historical simulation of the bi-issuer CTD,” she says, “you may reach that conclusion. But the divergence between the OAT and the Bund CTD will decrease. We’re not creating a contract for the past, but for the future.”
On February 22, volume in the June dual-issuer notionnel was 1,508 lots, with open interest of 6,153. Volume in the two-year E-note March contract was 80 lots, with open interest of 118.
Electronic candidates
The battle of the benchmarks also involves efforts to enhance trading efficiency and liquidity. To most people, that means, eventually, electronic trading of cash bonds and repos. Many rival systems are already out there, or about to come into play. The hot favourite is EuroMTS, based on the highly successful Italian MTS system. MTS allows electronic trading and settlement of Italian government bonds. About e20 billion of repos are also traded over the system daily, although the credit checks are done by telephone.
Greece has a similar system for trading domestic bonds, and the Netherlands is developing one. The EuroMTS system is being developed by 24 major bond-trading houses initially to trade the benchmark French, German and Italian government bonds. The aim is liquidity and efficiency. “Nobody got rich trading on the MTS system,” says an Italian fixed-income chief. EuroMTS is expected to start trading at the end of March.
A German contender is Xetra, owned by the Deutsche Börse, which trades cash bonds and other securities. An agreement has just been signed with the Vienna Stock Exchange to trade its cash products too by the end of the year.
Then there is Coredeal, a solution being developed by the International Securities Market Association (Isma), to trade Eurobonds, warrants and global depositary receipts. It was due to start operating in early March. In a second phase it plans to introduce anonymous repo trading.
Its chief rival in this respect will be a system being developed by the Swiss Exchange (SWX) to trade repos electronically. Its target is the repo of French, German, Italian and UK government bonds, but it also has corporate bonds in mind. Its launch is planned for June 18.
Bond broker Cantor Fitzgerald is preparing to launch E-Speed, an electronic wholesale trading system for European government bonds, as it has already done in the US. A further threat to the competition is a bond futures exchange in Europe, modeled on the one Cantor has opened in New York. Last month the UK treasury recognized the Cantor Financial Futures Exchange in New York, meaning that it can place remote terminals in London.
Euroland is set for an almighty clash of systems as these electronic platforms compete for liquidity in the same instruments. The major houses are likely to invest in each of them, until perhaps one or two become the standard. There will be pressure to look for other bonds to trade besides the government benchmarks. One source close to EuroMTS says it will be looking at corporate bonds too. As a foretaste, EIB last September made a couple of euro-fungible issues worth over €5 billion tradable on the Italian MTS system Others are bound to follow.
None of these candidates for pan-European trading is linked to a single clearing house, apart from Xetra. But surely the ultimate goal is to have a single platform that offers trading of cash, repos and futures, cross-margined at the same clearing house. “That would only be helpful if everyone is connected to one system,” says Ricardo Matteucci, head of repo at San Paolo Bank. There are also obstacles to same-day delivery between the various systems and clearing houses, he says. Clearnet, an initiative by the French stock exchange, in theory will do all these things, allowing cross-margining between over-the-counter products, cash bonds and repos – including German government securities – and Matif products.
While the outcome of the competition above is undecided, banks and brokers are thriving on a certain lack of transparency, as investors and borrowers come to terms with the new realities of euroland. For over-the-counter brokers the writing is on the wall, unless they can come up with products that add value. Banks run the risk that they will be disintermediated, as customers gain direct access to electronic trading systems. Those systems might then be superseded by trading facilities on the internet.
