The hunt for yield has seen higher institutional bond holdings. Hugo Cox explains how this has slowed down bond trading and led to information risks when large trades do occur. But can a new breed of e-platforms help?
The evidence for the tightening of corporate bond markets has been compelling for years. Figures collected by the International Monetary Fund (IMF) in its semi-annual Global Financial Stability Report show that inventory held by investment banks has decreased steadily since the financial crisis.
More recently, since May 2015 and until early October, US primary dealer corporate bond inventories declined by around $12 billion (€10.6 billion), to $43 billion. The drop has been driven mainly by high-yield corporate bonds, which saw their inventory fall by $11 billion.
This is noticeable especially in the US credit market. New York Federal Reserve data collected by the International Institute of Finance (IIF), a Washington-based industry association, reveals that US primary dealer corporate bond inventories decreased from a high of $250 billion in 2007 to $50-60 billion in 2014.
Trading has plummeted, too. US corporate bond secondary market trading volumes dropped from $32 million traded per day in February 2012, to $8.3 million per day in January 2014, according to a 2014 Fed study.
Emerging market bond trading volumes have held steady. Emerging Markets Trading Association figures collected by IIF show that on secondary markets there was $5.9 trillion worth of emerging market bonds traded in 1997, similar to the $5.6 trillion traded in 2013.
However, the boom in emerging market debt issuance since the crisis has meant trading as a share of the outstanding universe dropped sharply.
The reason typically given for bank withdrawal of inventory is the higher cost of trading imposed by regulations.
On the one hand, US firms can no longer house inventories in their proprietary trading books (an important source of liquidity from which to fill client orders), because the ‘Volcker rule’ has banned proprietary trading. On the other hand, in Europe, Basel II capital requirements have made it more expensive for banks to hold inventory as part of market-making operations.
But the decline in dealer inventories started well before the Dodd-Frank Act, which contains the Volker rule, and Basel II were fully fleshed out, let alone implemented. Dodd-Frank was signed in July 2010 and Basel II was drawn up in 2010-11 – yet the drop-off in inventories, as the IMF data shows, was well underway by 2009.
Whatever the reasons for banks withdrawing, their retrenchment is only half the story about the structural shift in fixed income ownership. The other half has been the increased participation of institutional investors and the asset managers that represent them in their hunt for yield, which has increased their holdings.
The arrival of institutional investors has reduced turnover. They have longer investment horizons than market makers, so selling is less frequent, and the chase for yield has led these investors to own longer duration bonds, which they are likely to hold for longer.
“Where institutions own assets to meet liabilities, very few need daily liquidity on those assets,” says Nick Gartside, international chief investment officer for financial institutions at JP Morgan Asset Management. Instead, what these investors want is visible cash flows, he says.
Precluding wide-scale defaults, these remain on offer even when prices fall. During the financial crisis, he says, while the bond market moved sharply, the cash flows from corporate bonds remained stable. So don’t expect liquidity to come back from this segment.
Philippe Lespinard, co-head of fixed income at Schroders in London, points to the difficulty created by all this for redemptions. When investors have to exit positions within a certain timeframe, he says, the challenges can be considerable. Dwindling liquidity affects the price that can be achieved, especially for larger orders, or
in cases where a market for a bond is thin.
Electronic trading platforms are wrestling with this problem. Stu Taylor, CEO of one such platform, Algomi, says: “Just an enquiry about a decent-size trade in an illiquid market could be enough to move the price against you.” In part this is because brokers will often contact clients to find out who can take the other side of the illiquid trade; even this small leak of interest starts to move the market against you.
One way to avoid moving the price against you is to deal in smaller chunks. Dealers are reluctant to distribute data about trade sizes but the growth of electronic platforms like Algomi means there is more efficiency processing the large number of small deals around so as not to move prices.
Trades larger than $100,000 were 10% lower in 2014 than in 2008, according to data from MarketAxess, the largest of the new platforms.
In 2013, information presented to the US Congress suggested that the average size of larger trades – so-called ‘block’ trades – had declined from $28 million in 2005, to $14 million.
The upshot of these developments is that bond liquidity is highly fragmented across a wider number of participants. “This is particularly the case in corporate bonds, where it is often heard that liquidity is a mile wide and an inch deep,” says Liz Callaghan, market practice and regulatory policy director at the International Capital Markets Association (ICMA).
The new breed of technology firms working to uncover and join up the disparate sources of liquidity falls broadly into two categories: those that improve information sharing pre-trade, and those that support execution.
In the first category is Algomi’s Honeycomb, a virtual anonymous trading platform that effectively embeds listening posts for buy-side firms on bond trading floors, while keeping the identity of the banks hidden until they express an interest to trade.
It means the investor (assuming the dealers have cleared it for access) can identify the counterparty best placed to trade, removing the need to contact a large number of dealers with details of the order and risking information leakage.
Honeycomb displays pertinent details – such the history of trading in a security – for the list of dealer firms without identifying the firms who are not aware of the level of investor scrutiny at that stage. Only when the client decides to open a conversation with the chosen dealer are the identities of each party revealed to each other.
By focusing on virtual rather than real inventory, the system benefits the banks, too, says Taylor. Ambitiously, he reckons by itself it can “bring the required velocity back to corporate bond markets”.
Algomi currently has 14 banks, including Deutsche Bank, Credit Suisse, HSBC and Commerzbank and 140 buy-side firms, with more than 100 buy-side firms in legal negotiations, according to Taylor.
A different solution to the same problem is provided by Project Neptune. Here the approach is to shine light on the pre-trade process by using the FIX messaging format to standardise messages used to record bonds held by participants. By replacing a number of disparate protocols for a single one for exchanging this information, the initiative aims to reduce the likelihood that a participant may miss the fact that a counterparty has bonds it wants to trade.
Unlike Algomi, Neptune is user-led. It is run by Etrading Software, but was initiated by a group of lenders. When it went live in August, it boasted 42 users: 15 banks, including BNP Paribas and Credit Suisse, and 27 European money managers, including Axa and Aviva.
These non-execution platforms join a host of existing solutions that focus on execution itself, such as Bloomberg, MarketAxess and Tradeweb – the three incumbents that command the majority of current flows.
But the number is growing. The ICMA’s ICMA Electronic Trading Platform (ETP) Mapping Study counts 22 platforms in European cash bonds. Eighteen are execution venues – such as Liquidnet, MarketAxess and Tradeweb – and two are owned by Euronext.
GreySpark Partners, a London-based consulting firm, recently estimated there were 32 venues, with nine more due to launch.
Rick McVey, chief executive of MarketAxess, is quick to point out the low level of uptake among new platforms. Perhaps there is something perverse in the existence of potentially more than 40 technology platforms trying to solve the problem of distributed liquidity.
For users it’s not only about ensuring the chosen platform has enough participants to provide adequate liquidity. Consortium-owned initiatives, such as Neptune, have to prove to users that owner-users won’t receive preferential treatment in terms of pricing or visibility.
“They are also looking for owners with sturdy governance structure, deep commitment and deep pockets,” says the ICMA’s Callaghan. Connecting to platforms involves expensive and time-consuming IT projects. Users are wary of going to the time and expense of installing a new system only to see it fail, leaving them to start all over again.
They were starkly reminded of the dangers in July, when Bondcube, a platform backed by Deutsche Börse that had been trading since June, ran out of money and filed for liquidation. The Börse said the venture failed to provide “sufficient business prospects”.
Expect some natural selection from here, says Callaghan – more failures and some mergers and acquisitions. The principle of matching buyers and sellers with the speed and anonymity that technology can provide is, after all, quite a simple one. Looking down the ICMA’s platform list, a number of ‘unique’ selling points seem to be remarkably similar.
©2015 funds europe